DEV Community: Prabhash Jha

AI Is Heading Toward Autonomy: What ‘Dark Factories’ Tell Us About What's Next

Prabhash Jha — Mon, 17 Aug 2026 10:52:26 +0000

A "dark factory" is a production facility that runs with the lights off. Nobody inside needs to see. That's the image everyone reaches for when they talk about where AI is heading. Every few months a new one gets announced and the coverage repeats itself. No workers. One unit per second. The future has arrived.

Honestly, that framing gets the interesting part backwards. The dark factory isn't proof that machines finally got good enough. Machines were good enough for a lot of this a decade ago. What changed is everything around the machines. And that surrounding list is the actual lesson, because it's also the checklist that tells you whether any process you run could survive being left alone.

I care about this for an unglamorous reason. I run automation that operates unattended, including the pipeline that publishes this site. The constraint has never once been whether the software could do the task. It's always been whether anything downstream could tell the difference between a good result and a confident, well-formatted, empty one.

What a dark factory actually is, and what the headline leaves out

A dark factory, or "lights-out" facility, is one where machines and software run production end to end with minimal human presence on site. Robots handle assembly, movement and packing. Scheduling, inspection and logistics run in software. These exist. They're mostly in electronics and high-volume component manufacturing. And yes, they are genuinely impressive.

Three things the announcement coverage consistently omits.

There are still people. The staff moved rather than disappeared. Out of the line and into supervision, maintenance, tooling, exception handling, and the changeovers that happen when the product changes. "Lights-out" describes the assembly floor during a run. It doesn't describe the building's headcount. Every real facility I've read about in detail has humans somewhere in the loop. Usually doing the hardest and least automatable work in the plant.

Headline production rates are peaks. "One unit per second" is a line-rate figure. Measured under ideal conditions on a running line. Sustained annual output is always materially lower, because real facilities have maintenance windows, changeovers, defect batches, supply gaps, and demand that isn't flat. That's true of every factory that has ever existed and it isn't a scandal. But a peak rate reported as a throughput rate is how a lot of automation gets oversold. The same trick shows up in software demos.

Nothing here generalises to messy work. These plants automate the most automatable process humans have ever designed. A fixed product. Identical components. A controlled environment. An immediately measurable pass/fail on each unit. That's not a stepping stone to autonomous everything. It's the extreme easy end of a spectrum, and the distance from there to "runs my business" is much larger than the coverage implies.

If you want the state of industrial automation from the people who actually count it rather than from the announcement cycle, the International Federation of Robotics publishes the primary numbers in its World Robotics reports. Its press releases and free downloads are the source most of the secondary coverage is loosely paraphrasing.

The five conditions that had to be true before the lights went off

This is the part worth stealing. Before anyone was willing to leave a factory unattended, five things had to be true. None of them is about how capable the robots are.

1. The input is constrained and identical every time. Components arrive in known orientations, in known quantities, within known tolerances. The system never has to decide what it's looking at. Where inputs vary, humans are still there.

2. The output is measured automatically, immediately, and by something other than the thing that produced it. Every unit gets checked by a separate system. The machine that built it doesn't get to grade its own work.

3. Failure is detectable by the process, not by a person noticing later. A jam, a tolerance drift, a missing part. Each produces a signal at the moment it happens. Nothing depends on somebody eventually spotting that the output looks wrong.

4. There is a stop condition. When something goes outside bounds, the line halts. It doesn't carry on producing defective units efficiently. The default on uncertainty is stop, not continue.

5. Somebody owns it. A named person is responsible for the line, gets woken up when it stops, and is accountable for what it produced. Autonomy didn't remove the owner. It removed the operator.

Look at that list again and notice what isn't on it. Intelligence. Not one of the five conditions is "the AI got smart enough". Four are about verification and one is about accountability. That's the actual precondition for autonomy, in a factory or anywhere else.

Score your own process against those five

Take any process you were thinking of handing over to AI and run it down the list. The result is usually clarifying and slightly deflating.

Condition	Factory floor	Typical knowledge work
Constrained input	Yes, identical parts	Rarely. Every brief, email and dataset is a bit different
Independent measurement	Yes, separate inspection	Almost never. The output is judged by reading it
Failure detectable by the process	Yes, immediately	No. Wrong output looks exactly like right output
Stop condition	Yes, the line halts	Usually absent. It just continues
Named owner	Yes	Often nobody, once it is "automated"

Most knowledge work fails on rows two and three. And it fails there regardless of how good the model is. That's the whole point. If nothing downstream can distinguish a correct output from a plausible one, then improving the model doesn't get you to autonomy. It gets you to failures that are harder to spot.

This isn't a theoretical concern. It's the most common way automation quietly stops working. The job runs, reports success, and produces nothing or nonsense, for weeks, because a green tick isn't a measurement. I've shipped that bug more than once, and the automation failure nobody catches is the long version of what it costs and how to build against it.

Autonomy is a supervision level, not a capability level

The usual three-stage model, assisted, automated, autonomous, is useful. But it's normally described as if it were a ladder of machine ability. It's more accurate, and far more actionable, to read it as a description of where the human sits.

Assisted. The system suggests. A human decides and is responsible for every output. Drafting, summarising, code suggestions. Errors are caught because a person reads everything before it matters.

Automated. The system executes inside tight rules. A human sets the rules and reviews in aggregate rather than per item. Sorting, classification, routine responses. Errors are caught by sampling and by exception reports. Which means the exception reports have to actually exist.

Autonomous. The system decides and acts within a goal. A human owns the outcome and is notified by exception only. Errors are caught by the verification you built in advance, or they aren't caught at all.

Moving a process up this ladder isn't a matter of the model becoming capable enough. It's a matter of you building the checking apparatus that each level requires. Skipping that construction doesn't leave you at the previous level. It puts you at the top level with none of the safeguards, which is strictly worse than staying at the bottom.

The framing that has stayed useful for me: you don't get autonomy by trusting the system more. You get it by needing to trust it less. The factory doesn't trust its robots. It inspects every unit.

Where this actually lands for a small team

The pattern repeats in software. And this is the part that matters if you aren't in manufacturing. Agents that plan and carry out multi-step tasks are the dark factory of knowledge work. Automation moving from a single task to a whole workflow.

The opportunity is real and it's asymmetric in favour of small teams. A person with well-built systems can now cover work that used to need a department. The constraint is no longer headcount or budget. It's how clearly you can specify what "done correctly" means. That's a skill, not a purchase.

But the bottleneck moves rather than disappearing. And it moves somewhere uncomfortable. When execution is cheap, the scarce inputs become deciding what should be built, and being able to tell whether the result is any good. Both are judgement, and neither is automatable by the thing whose output you are judging. That's the argument in the skills that actually matter in the age of AI. The factory version of it is that the humans who remained on site are the ones doing exception handling. The work that only shows up when something unexpected happens.

There's also a failure mode specific to putting a language model in the loop. It's the exact inverse of the factory's. A jammed robot stops. A model that has been given nothing useful produces a fluent, confident, well-structured answer about nothing, and passes every check that only asks "did it return something?". Understanding why it does that isn't optional if you're going to run one unattended. Why ChatGPT gives wrong answers covers the mechanism, and the practical rule that falls out of it is worth stating plainly:

Let the model do the judgement. Let deterministic code do the checking. Never let the model grade its own work.

That's condition two from the factory list, translated. It's also the single design decision that has saved me the most trouble.

What to actually do about it

Four things, roughly in order of how much they pay.

1. Automate your own repetitive work first, and start with the verification. Pick something you repeat weekly. Before you build the automation, write down how you'll know next month whether it's still working. If you can't answer that, you aren't ready to automate it. You're ready to do it manually and pay attention. How to automate your work with AI is the practical starting point, including the part where the maintenance cost, not the build cost, decides whether a workflow was ever worth having.

2. Learn to direct rather than to compete. The valuable skill is briefing, reviewing and combining. Not producing faster. How to actually use ChatGPT and Claude for real work is what that looks like day to day, and the reviewing half is the half people skip.

3. Build the stop condition before you build the capability. Decide in advance what output means "halt and tell a human". A workflow with no defined halt state will run through every edge case it meets, confidently, forever.

4. Put your name on it. Autonomy removes the operator, not the owner. If nobody can answer "who finds out when this stops working?", the honest answer is nobody, and you've built a liability with a dashboard. Formal risk frameworks say the same thing at greater length. NIST's AI Risk Management Framework is organised around exactly this. Mapping and measuring a system's behaviour and assigning accountability, rather than around how capable the model is.

Does this take jobs?

It moves them. And it moves them in a direction that is legible from the factory example rather than from speculation.

The line workers didn't become unnecessary and simultaneously stay in the same job. The roles that grew were the ones adjacent to the machines. Maintenance. Tooling. Systems integration. Quality engineering. Exception handling. The roles that shrank were the ones consisting of a repeated physical action with a measurable output. Which is to say, precisely the roles that satisfied the five conditions.

The knowledge-work version follows the same rule, and you can apply it yourself: the more your work resembles a constrained input producing an immediately checkable output, the sooner it changes. The more it involves ambiguous inputs, contested definitions of "correct", relationships, or accountability for a decision, the longer it holds. Not because it's impossible to automate, but because nobody can build condition two around it.

That's a more useful thing to plan against than a headline about robots. Will AI take my job works through what to do with the answer. The historically consistent pattern is that automation waves reward the people who move toward the new tools early and penalise those who wait to be told. There's nothing in this one that suggests otherwise.

The mistakes I see most

Reading a demo as a throughput figure. The peak rate under ideal conditions isn't the sustained rate, in a factory or a product demo. Ask what it does over a month, on real inputs, including the bad ones.

Automating capability without automating verification. The most expensive version of this error. You now produce ten times as much output and have exactly the same ability to check it. Which is to say you have industrialised whatever mistake you were already making.

Assuming a green status means the work happened. The tool reports that it executed, not that the outcome is correct. Those are different claims and the gap between them is where unattended systems fail.

Treating "autonomous" as a purchase. It's a property of the system you build around a tool, not a feature of the tool. Nobody sells it to you.

Removing the owner along with the operator. Unowned automation is where silent failure lives, because everybody assumes somebody else would have noticed.

Concluding that judgement no longer matters. Cheap execution raises the value of deciding what to execute. The factory didn't eliminate its engineers. It eliminated the repetition around them.

FAQs

Are fully autonomous factories real, or is it marketing?

Both, in different proportions than the headlines suggest. Lights-out production runs are genuinely real in electronics and high-volume component manufacturing, and they are a real engineering achievement. What's marketing is the implication of zero humans and the use of peak line rates as though they were sustained output. The accurate version, where machines run production while a smaller specialist staff handles supervision, maintenance and exceptions, is less dramatic and considerably more useful to reason from.

What is the difference between automation and autonomy?

Automation executes a defined procedure. Autonomy pursues a goal and decides the procedure itself. The practical difference isn't the software, it's where the human sits. Automation needs someone to have specified every step. Autonomy needs someone to have specified what a correct outcome looks like and what should happen when it isn't achieved. That second specification is the harder one, and it's the one people skip.

Can I make my own work "lights-out"?

Parts of it. The five conditions will tell you which parts far more reliably than intuition will. Anything with a stable input, a checkable output and an obvious failure signal is a candidate today. Anything where you would only notice a bad result by reading it carefully isn't. Regardless of how impressive the tool is. Not until you've built something that does the reading.

Isn't the AI good enough now that verification matters less?

It's the reverse. As output quality rises, errors stop looking like errors. A bad result now arrives well-formatted, plausible and internally consistent, so the cheap human check (does this look wrong?) stops working precisely when you start relying on it. Higher capability raises the cost of each undetected mistake and lowers your odds of detecting it by glancing. Verification gets more important as the model improves, not less.

What should a small business actually do this quarter?

One process, end to end, with the checking built first. Pick something repetitive whose failure you would notice within a week. Write down how you'll verify it monthly. Automate it. Put a name against it. That's worth more than a strategy covering everything you might automate later, and it teaches you the thing you actually need. Which isn't how to use the tool, but how to tell whether it worked. Marketing automation is a reasonable place to find the first candidate if nothing obvious comes to mind.

Key takeaways

Dark factories are real, and they're the easy end of the spectrum: fixed product, identical inputs, immediate pass/fail on every unit.
The staff moved rather than vanished, into supervision, maintenance and exception handling. "Lights-out" describes the line, not the payroll.
Headline rates are peak line rates. Sustained output is always lower, in factories and in demos.
Five conditions had to be true before the lights went off, and four of them are about verification, not intelligence.
Most knowledge work fails on "measured independently" and "failure detectable by the process". A better model doesn't fix either.
Autonomy is a supervision level, not a capability level. You reach it by needing to trust the system less, not more.
Let the model do the judgement and deterministic code do the checking. Never let it grade its own work.
Build the stop condition before the capability. A workflow with no halt state runs through every edge case forever.
Automation removes the operator, never the owner. Put a name on anything that runs unattended.
The closer your work is to a constrained input with a checkable output, the sooner it changes. That's a better planning tool than any headline.

Affiliate Tracking Breaks Quietly, Catch It Before Payout Day

Prabhash Jha — Mon, 17 Aug 2026 10:52:25 +0000

Most tracking problems don't announce themselves. Nothing errors. No alert fires. The campaign runs, the dashboard fills with clicks, and everything looks normal. Right up until the end of the month, when the network's conversion count and yours disagree and somebody has to decide who's right.

By then you're negotiating about the past. That's the worst position to be in. Neither side can prove anything, and the money's already been spent.

The fix isn't a better tracker. It's checking one thing before launch instead of after payout.

Everything hinges on one value surviving the journey

Strip away the vocabulary (postback, S2S, pixel, macro, subid) and affiliate tracking is a single idea. If none of it is familiar yet, how to track affiliate sales as a complete beginner is the version before this one.

You attach a unique ID to a click. Later, the advertiser hands that same ID back to you attached to a conversion. Match the two, and you know which click earned the money.

That's the whole mechanism. Which means there's really only one failure. The ID does not survive the round trip. Everything else is a variation on that.

When the ID is missing, the conversion still happens. The customer still buys. The advertiser still books the revenue. The only thing that disappears is your ability to prove the sale was yours. And unprovable sales don't get paid. That makes broken tracking one of the mistakes that quietly kill beginners, because nothing on the dashboard looks wrong.

What the chain actually looks like

Before the failure modes make sense, it helps to see the whole path written down. Most people have only ever seen their own end of it.

Your link. Someone clicks a URL that points at your tracker, not at the advertiser.
Your tracker mints an ID. A unique string for this click, stored on your side with the source, the creative and the timestamp.
A redirect to the offer. Your tracker sends the browser onward to the advertiser's landing page, appending that ID to the URL as a parameter.
The advertiser captures it. Their page reads the parameter and stores it. In a cookie, in session storage, in a hidden form field, or against a user record if the visitor signs in.
A conversion happens. Possibly minutes later, possibly on another device, possibly after a support call.
Their server fires a postback. A server-to-server HTTP request back to your endpoint, carrying the ID, an order value, and ideally a transaction ID.
Your tracker matches. ID found, conversion recorded, attributed to the original click.

Seven steps. At least three parties. No single owner. The ID has to be written correctly at step 3, read correctly at step 4, survive whatever happens between steps 4 and 5, and be echoed back accurately at step 6. Any of those can fail without anything else appearing wrong.

Which is honestly the whole problem in one sentence. A chain of handoffs where nobody sees the full picture, and where a silent failure at any step ends the same way. You don't get paid.

And nobody in that chain is incentivised to look upstream on your behalf. Your tracker is looking at what arrived. The advertiser is looking at what left their box. The gap in the middle is yours to inspect.

The six places the ID actually gets lost

In roughly the order I see them.

1. It was never in the offer URL. The click ID exists in your tracker, but nobody appended it to the destination URL. Nothing to hand back. This is the most common one. Also the most embarrassing, because it's entirely preventable in thirty seconds.

2. The macro didn't resolve. You put {clickid} in the URL and the platform expected {click_id}. Or the reverse. The unresolved token gets passed through as literal text. Your tracker then tries to match a conversion to a click called {clickid}, which does not exist.

3. The parameter names don't agree. You send subid. They store aff_sub. They post back aff_sub. Your tracker is listening for subid. Both sides are working perfectly. Neither can hear the other.

4. A redirect in the chain strips it. Every hop between the click and the landing page is a chance to lose the query string. Long chains, shorteners, geo-redirects and consent gateways are all common culprits. The more intermediate hops, the more likely one of them rebuilds the URL from scratch and drops what it doesn't recognise.

5. The session broke. The ID reached the landing page but the conversion happened somewhere else. A different subdomain. An app handoff. A checkout on a separate host. A purchase completed days later on another device. Whatever was holding the ID didn't make it across.

6. The postback fired but nobody was listening. The advertiser sent it. Your endpoint returned an error, timed out, or sat behind a firewall rule. Many platforms fire once and never retry.

Notice something. Only one of these is really about the postback URL. The rest happen upstream, before anyone gets near the part most guides spend their time on.

Failure two deserves its own paragraph

The macro mismatch is worth singling out because it's the one that looks correct in every screenshot.

There is no standard for what these placeholders are called. Different platforms use different token syntax and different parameter names for the same concept, and both sides usually assume theirs is the obvious one. You will see the same value called click_id, clickid, cid, transaction_id, subid, sub1, aff_sub and s1 depending on whose documentation you're reading. And the token wrapping it may be {...}, {{...}}, [...] or ##...##.

That means "I put the macro in" is not a verified state. The only verified state is I clicked the link and saw a real value in the address bar. An unresolved macro is visually almost identical to a resolved one at a glance. {clickid} and a7f3c9d2 both look like something is there. Read the actual characters. Not the shape of the string. The characters.

The postback itself: four things worth getting right

The postback is the shortest part of the chain. It's also the part most likely to be set up once and never inspected.

Send a transaction ID, not just a click ID. The click ID says which click. A transaction ID (your order number, your payment reference, anything already unique in the advertiser's own system) says which sale. Without it, you have no way to tell a genuine second purchase from the same postback arriving twice. And platforms do resend. Reuse an existing order ID rather than inventing a new identifier space. It's already unique and it's already in the advertiser's database when you need to look something up.

Test what happens on a duplicate. Fire the same postback twice, deliberately, with identical click ID and transaction ID. One payable conversion should be recorded. The second should be logged as a duplicate. If two appear, you have a system that will inflate your own numbers and destroy your credibility in the first dispute you raise.

Log the raw request, not just the parsed result. When a discrepancy appears three weeks later, "our tracker shows no conversion" is a weak position. "Here is the raw inbound request we received at 14:22:07, with these parameters and this response code" is a strong one. Keep the raw log for at least as long as your longest payment cycle.

Check the response code. Your endpoint should return 200. A redirect, a 403 from a firewall rule, or a timeout under load all look like silence from the advertiser's side. And most platforms fire once and move on. If your tracker offers a delivery log for inbound postbacks, read it during the first week. It's the cheapest early warning available.

Why the browser is now the weakest link

Server-to-server isn't a preference. It's a response to browsers deliberately becoming worse at exactly the thing affiliate tracking used to rely on.

Safari's WebKit announced full third-party cookie blocking in March 2020, "cookies for cross-site resources are now blocked by default across the board", alongside "deleting all of a website's script-writable storage after seven days of Safari use without user interaction on the site", which covers LocalStorage, IndexedDB, SessionStorage and Service Worker registrations (WebKit: Full Third-Party Cookie Blocking and More). Other browsers have moved in the same direction at their own pace.

What that means in practice, for anyone running offers:

A conversion that happens more than a few days after the click is the fragile case, not the exception. Long consideration cycles (insurance, lending, travel, anything with a comparison step) are exactly where browser-held identifiers expire before the sale lands.
Anything depending on a third-party cookie is already gone on a meaningful share of traffic. If your setup still leans on one, your reported conversion rate is not a measure of your offer.
The click ID must be captured server-side, or into first-party storage under the advertiser's own domain, at the moment of arrival. Whatever holds it has to belong to the site the user is actually on.

Which is also why pixel-versus-postback isn't really a debate any more. Use a pixel only where server-to-server is genuinely not offered. Treat those campaigns as measured with a known undercount rather than measured accurately.

The ten-minute test that prevents almost all of it

Before a campaign takes real budget, run one conversion through the entire system yourself.

Click your own tracking link. Not the offer URL. Your link, the way a real user would arrive.
Look at the address bar on the landing page. Is your click ID actually there, in full? Copy it somewhere. If it's missing at this step, stop. Nothing downstream can save you.
Complete the action. Buy, register, submit. Whatever counts as a conversion. Use a test mode if the advertiser has one.
Check that the conversion appears in your tracker, and that it carries the exact ID you copied in step 2.
Check the advertiser's side too. Both systems should show one conversion with the same identifier.

If all five line up, your tracking works. If step 2 fails, it's a link problem. If step 4 fails but the advertiser recorded the sale, it's a postback problem. That single distinction tells you which half of the system to look at. Which is most of the diagnosis, actually.

Ten minutes before launch. Compare that to a month of disputed conversions. This is the affiliate-specific version of the more general "verify tracking before you run anything" discipline that sits at day 14-30 of the first 90 days of an agency account, in the order they actually matter. Same rule, sharper edges, higher cost of getting it wrong because the reconciliation is with a third party who won't pay you if the numbers don't line up.

Two additions worth making once you've done this a few times. Run it on a phone as well as a desktop. App handoffs and in-app browsers break sessions in ways desktop testing will never reveal. And run it from the geography the traffic will come from if the offer has geo-redirects. A redirect that only fires for certain countries is a redirect you will never see from your own desk.

Download the pre-launch tracking check. The five steps above, the six failure points, and the discrepancy table, on one sheet. Free, no signup. Works in Excel, Google Sheets and Numbers.

When the numbers disagree anyway

They will, sometimes. Some gap is normal. Different attribution windows, different timezones, cancelled and returned orders, deduplication rules. A small, stable discrepancy is a fact of the channel, not a bug.

Worth separating the two things that can be wrong here, because they look identical from a dashboard. Everything in this post is about the technical chain failing. The other possibility is that the chain is intact and the rule is the problem, which is a negotiation rather than a bug — attribution windows are a commercial term, not a technical setting covers the case where a partner asks to widen the window and what it actually costs to agree.

What matters is the shape of it.

Shape of the gap	Most likely cause	Where to look first
Small and steady, month after month	Definitional (windows, timezones, refund handling)	Write both definitions down once, then stop relitigating
Sudden, from a specific date	Something shipped	Release notes, consent banner changes, new redirect
One traffic source only	That source's redirect chain	The hops between that source and the landing page
Total (you have nothing)	The ID never arrives	Back to the ten-minute test
Yours is higher than theirs	Duplicates on your side	Your dedupe rule and your transaction IDs

That last row is the one people forget. And it's the most dangerous to your position. If you raise a dispute using numbers inflated by double-counted postbacks, you lose the argument and some standing with it. Check your own side first, every time, before writing the email.

All of these diagnoses assume you have a number of your own to compare against. If you don't (if every figure in the conversation came from the network) then there's no shape to read and no discrepancy to investigate. Only a payment to accept. That's the strongest practical argument for owning a click layer before you need one. It's also why starting affiliate marketing without a website is fine right up until the first month the numbers look wrong.

Untrusted numbers are also unusable numbers. CAC, ROAS and payback all read off this data, which is why the metrics guide is worth nothing until the tracking is proven.

The useful instinct is to check whether the pattern changed, not whether the numbers match exactly. Exact matching isn't the goal and never was.

How to raise a discrepancy so it actually gets fixed

When you do have to escalate, the difference between a resolved ticket and a three-week thread is entirely in what you attach.

Bring four things. The date range. Your count against theirs. A handful of specific click IDs you can prove converted. And the raw log line showing what arrived at your endpoint and when. Specific IDs are what turn the conversation from a disagreement about totals into a lookup somebody can perform. A partner manager can search a transaction ID in thirty seconds. They cannot do anything at all with "we think we're being underpaid by around eight per cent".

Ask one narrow question rather than making an accusation. Can you confirm whether a conversion was recorded against click ID X on this date, and whether a postback was attempted? You want the answer, not the argument. And whatever comes back (including "no postback was attempted, that is on us") write down the definition you both settled on. Half of all recurring discrepancies are the same misunderstanding being rediscovered every month by different people.

The habit that actually matters

Test tracking as a launch step, not a troubleshooting step. That's really the whole rule.

Every campaign. Every new offer. Every time the advertiser redesigns their checkout. Every time someone adds a redirect "temporarily". Not because tracking is fragile in some mysterious way. Because it depends on a value surviving a chain of systems that nobody owns end to end. And any one of them can change without telling you.

There's a second habit worth adding. Re-run the ten-minute test on a schedule for anything that carries real budget, not only at launch. Nothing tells you when the advertiser's checkout changed. The first signal is a month of missing conversions, and by then the evidence has aged past the point where anyone can reconstruct it.

This is the unglamorous half of the performance marketing playbook. You cannot optimise toward a profit metric you cannot measure. The people who lose money on affiliate tracking are rarely the ones who don't understand postbacks. They're the ones who found out in week four what they could have found out in minute ten.

FAQs

Why do my numbers never exactly match the network's?

Because you're counting slightly different things. Attribution windows, timezones, refund and cancellation handling, and deduplication rules all differ between platforms. A small consistent gap is normal. Agree on the definitions once, document them, and only investigate when the pattern changes. Chasing an exact match is a fight nobody wins.

Should I use a pixel or a server-to-server postback?

Server-to-server, wherever the advertiser supports it. Pixels run in the browser, so ad blockers, privacy settings, tracking prevention and users closing the tab early all cost you conversions. A postback is a server talking to a server and isn't affected by any of that. Use a pixel only when S2S isn't offered, and treat those campaigns as carrying a known undercount.

What is the single most common tracking mistake?

The click ID never making it into the offer URL in the first place. It's trivially avoidable and it silently invalidates the entire chain. Every downstream setting can be perfect and you'll still get paid for nothing.

What is the difference between a click ID and a transaction ID?

The click ID identifies which click. It's minted by your tracker when someone clicks your link, and it's what attribution matches on. The transaction ID identifies which sale, usually the advertiser's own order or payment reference. You need both. The click ID to attribute, and the transaction ID to deduplicate. Without a transaction ID you can't tell a genuine repeat purchase from the same postback arriving twice.

How long should a click ID stay valid?

As long as your agreed attribution window, and no longer. The important part is that both sides use the same number and the same timezone. A seven-day window measured in IST and a seven-day window measured in UTC will disagree on every conversion that lands near midnight. Which is a steady, unexplained gap that survives every investigation until somebody finally compares clocks.

Do browser privacy changes break affiliate tracking?

They break the browser-dependent parts of it. Safari's WebKit has blocked cross-site cookies by default since March 2020 and clears script-writable storage after seven days without interaction on the site, and other browsers have moved similarly. Server-to-server postbacks are unaffected, which is why long-consideration offers in particular should be measured that way rather than by anything held in the browser.

How do I test tracking without spending money?

Ask the advertiser for a test mode, a coupon that zeroes the order, or a staging endpoint. If none exists, make one real minimum-value purchase yourself and refund it afterwards. And tell the partner manager you're doing it, so the refund doesn't look like fraud. The cost of one test order is always smaller than one disputed month.

Brand Building: Five Lessons From Working Inside SMEs

Prabhash Jha — Sun, 16 Aug 2026 09:57:21 +0000

Most writing on brand building assumes you already have a marketing team. This isn't that. These are five things I picked up working inside small and mid-sized businesses. The kind where the founder is still doing sales, and "brand" sounds like something you get to later.

Each one below is stated plainly first. Then what it actually means in practice. Because the lesson on its own is easy to nod at and hard to act on.

1. Brand building has to start on day one, because later never comes

The lesson: brand building should be carried out from day one, because it gets difficult once the actual firefighting begins.

Every founder means to get to brand work. Almost none do. And the reason isn't laziness.

It's that the point when you have time for it, early, when you're small, is also the point when it feels least urgent. By the time it feels urgent, you're firefighting. A client is unhappy. Payroll is close. A competitor undercut you. Nobody works on positioning in that week.

So the window is at the start. And it closes quietly.

Good news is that day-one brand work isn't expensive or elaborate. It's four decisions and they take an afternoon:

Who you are for, and who you are not for. The second half is what people skip. It's also the half that does the work. A brand that serves everyone is legible to no one.

What you refuse to do. Constraints define a brand faster than aspirations. "We don't take retainer clients under X" or "we don't do work we can't measure" tells people more than any mission statement.

How you sound. Not a style guide. Just: are you formal or direct? Do you explain or assert? Get this wrong and every piece of communication for the next five years fights itself.

What you want to be known for, specifically. Not "quality" or "trust". Something a customer could repeat to someone else in a sentence.

Write those down. That's it. That's day-one brand work. And having it written is what makes the next hundred decisions consistent instead of ad hoc, which is all brand consistency actually is.

2. Don't hire names early. Start lean.

The lesson: at the start, don't try to surround yourself with people hired for their names. Either the people joining the freshly floated ship are equally responsible and dedicated, or the leader should start lean. It sounds good early, but over a longer run, one step at a time is what helps.

This is the one that costs people the most. It's also the one nobody talks about, because it feels counterintuitive. A recognisable name early looks like validation. Investors like it. It shortens conversations.

Here's why it usually goes wrong.

Senior hires from established companies bring a system, not a business. They're excellent at operating a machine that already exists. Process, budget, team, brand recognition. A young company has none of those. The skill that made them valuable doesn't transfer to a context where the job is to invent the machine.

Their incentives aren't yours. A founder's downside is total. A hired name's downside is a line on a CV. That asymmetry shows up in every risk decision. And it shows up hardest in exactly the moments that decide whether the company survives.

They raise the cost base before revenue supports it. Which shortens your runway. Which forces decisions you'd otherwise not make.

The test is right there in the lesson: are they equally responsible and dedicated? Not "are they impressive". Are they carrying comparable risk and comparable ownership? If yes, seniority is an asset. If no, you've bought a signal rather than a contributor.

The alternative is genuinely staying lean. Fewer people. More ownership each. Hire when the pain of not hiring is specific and repeated rather than anticipated. Slower, honestly, and far more of these businesses are still trading in year five.

3. People leave a brand when they stop believing in it

The lesson: a person leaves a brand when he stops believing in it.

True of customers. True of employees. The mechanism is the same, which is why treating internal and external brand as separate exercises is a mistake.

Belief goes in a specific order. Rarely one dramatic event:

A gap appears between what the brand says and what it does. Late payments after "we look after our people". A rushed delivery after "quality first".
The gap is explained away rather than fixed. This is the decisive step. One inconsistency is a mistake. A defended inconsistency is a signal about what's actually true.
The person stops giving the benefit of the doubt. Ambiguous things now read as confirmation.
They disengage well before they leave. By the time someone resigns or churns, the decision is usually months old.

Which gives you the practical version: the fixable moment is step 2, and it's fixable by fixing the thing rather than explaining it. The most brand-protective habit a small company has is naming its own failures before anyone else does. The cheapest way to find the gaps before someone else names them is the audit in your brand is whatever your worst touchpoint is. A Tuesday afternoon walking your own funnel produces a ranked list of exactly the say-do gaps that erode belief before anyone resigns or churns.

Which also explains why brand can't be delegated to marketing. Marketing controls what the brand says. Operations, finance and management control what it does. Belief is destroyed by the gap between them. And only one of those departments gets blamed for it.

4. Role, responsibility and repercussion collapse when trust goes

The lesson: role, responsibility and repercussion are mostly not considered by the employee or the employer once either fails to establish and sustain trust.

Once trust breaks, the formal structure stops functioning. Not because anyone tears it up. Because it stops being what people act on.

Role stops being a description of what you own and becomes a boundary you defend. "Not my job" is a trust symptom, not an attitude problem.
Responsibility stops being taken and starts being assigned. People document to protect themselves rather than to inform.
Repercussion loses its meaning. Consequences applied without trust read as arbitrary, so they change behaviour in the direction of concealment rather than performance.

Why does this matter for brand? An organisation in this state can't deliver a consistent customer experience, whatever the marketing says. Customers perceive internal dysfunction long before they could describe it. As inconsistency. As slow answers. As people who won't commit to anything.

Rebuilding trust is slow and mostly boring. Do what you said, at the size you said, on the date you said, repeatedly, including when it costs you. No faster route exists. Every attempt to find one, a values workshop, a rebrand, an offsite, is read correctly as an attempt to skip the work.

5. Company credibility scales you. Personal credibility saves you.

The lesson: in a competitive market a brand's credibility helps scale the business, and during tough times an employee's credibility in the market helps sustain and generate business relations. Keep a good balance between the two to do well in all situations.

This is the most useful of the five. It names two different assets that most people conflate.

Company credibility is what lets you scale. Shortens sales cycles. Supports higher prices. Attracts people who came to you. Survives any individual leaving. Compounds slowly. Belongs to the entity.

Personal credibility is what survives a downturn. When budgets are cut, buyers stop taking meetings with companies and keep taking calls from people they trust. When a company fails, personal credibility is what people leave with. And it's what lets them start again.

Neither substitutes for the other:

All company, no personal. The business works while it's working. Everyone is stranded when it isn't. Common in larger organisations. And the reason capable people find themselves unemployable at 45.
All personal, no company. The business can't grow beyond the founder's own relationships. Can't be sold. Stops the moment they do.

The balance in practice: let your people be visible under their own names. Small companies often resist this, fearing that someone who builds a reputation will leave. Some will. But a company where everyone is anonymous has no credibility to draw on in a bad year. And it's far less attractive to the people you want to hire.

Publish under individual bylines. Let people speak at events as themselves. Encourage them to answer questions publicly in their field. Costs almost nothing. Builds the second asset alongside the first.

How these five connect

Read separately they're five observations. Read together they describe one mechanism, which is more useful.

Brand is the accumulated evidence of how you behave. That's the thread. Lesson 1 says start recording that evidence deliberately from day one. Lesson 2 says don't let cost structure force behaviour that contradicts it. Lesson 3 explains how belief is lost when the evidence stops matching the claim. Lesson 4 describes what an organisation looks like once that has happened internally. Lesson 5 says the evidence attaches to two entities at once, the company and each person in it, and both are worth building.

Which means brand can't be fixed by any of the things people reach for when they notice it's weak. A new logo changes no evidence. A values workshop changes no evidence. A campaign changes what's claimed, which widens the gap rather than closing it. In a company where trust has already gone, the campaign is read internally as proof that leadership is unserious.

The only thing that works is behaving consistently for long enough that the evidence accumulates. Slow. Unglamorous. And the reason most small companies never get there despite genuinely wanting to.

The compensation is that it compounds. Ten years of consistent behaviour can't be bought by a better-funded competitor. It's the one advantage a small business holds that money doesn't erase.

What to do with all five

If you take one action from this, take the day-one exercise in lesson 1. The four written decisions. It takes an afternoon and it makes the other four lessons enforceable rather than merely agreeable.

Then the sequence that follows from the rest. Stay lean until the pain of not hiring is specific. Close the say-do gap the moment it opens, publicly. Protect trust as infrastructure rather than culture. Build company and personal credibility deliberately and at the same time.

None of this is fast. Brand is the accumulated evidence of how you behave. There's no way to accumulate evidence quickly.

FAQs

When should a small business start brand building?

Day one. Because the window closes. Early on you have time and it feels unimportant. Later it feels urgent and you're firefighting. Day-one work is just four written decisions: who you serve, who you don't, what you refuse to do, and what you want to be known for specifically.

Should a startup hire senior people from big companies early?

Usually not. Senior hires from established firms are skilled at operating systems that already exist, and a young company has none. The test is whether they carry comparable risk and ownership. If they do, seniority is an asset. If not, you've bought a signal rather than a contributor.

Why do employees leave even when pay is fine?

Because belief goes before people do. A gap opens between what the brand says and what it does. The gap gets explained away rather than fixed. The person disengages months before resigning. The fixable moment is when the gap is first defended instead of closed. Edelman's 2026 Trust Barometer puts numbers on the say-do stakes: 42% of employees said they'd rather leave a job than report to a manager with very different values, and another 34% said they'd reduce effort for one. Belief mismatch isn't soft. It's a majority-behaviour outcome. And it's the one that shows up in your P&L as unexplained turnover.

Is brand building the same as marketing?

No. Marketing controls what the brand says. Operations, finance and management control what it does. Belief is destroyed by the gap between the two, which is why brand can't be delegated to a marketing team alone.

Should employees build personal brands?

Yes, and companies should encourage it. Company credibility lets a business scale. Personal credibility is what sustains relationships during a downturn and what people carry if the business fails. A company where everyone is anonymous has nothing to draw on in a bad year.

Key takeaways

Do the day-one brand work now. Four written decisions, one afternoon.
Stay lean. Hire for equal responsibility, not for a recognisable name.
People stop believing before they leave. The decisive moment is when a say-do gap gets defended rather than fixed.
When trust breaks, role and responsibility stop functioning regardless of the org chart.
Build company credibility to scale and personal credibility to survive. You need both.

Affiliate Marketing Without a Website: How I'd Start From Zero

Prabhash Jha — Sun, 16 Aug 2026 09:57:20 +0000

Yes, you can do affiliate marketing without a website. Almost every page answering this question says so. Then it hands you the same list: YouTube, Instagram, a newsletter, Quora, Medium.

That list is correct. It isn't the useful part. The useful part is what changes when you don't own a page in the middle. Three specific things break. All of them invisible on day one. All of them expensive in month four. Nobody selling you a tracking tool or a link-in-bio product is going to lead with that.

I've spent eight years on the performance side of this. Including the part nobody writes about, which is reconciling what a network says it recorded against what you believe actually happened. That work is where you find out exactly how much a website was doing for you. Not in the "traffic" sense. In the boring plumbing sense. The record-keeping sense. The "I can prove what happened" sense.

That's the frame this whole piece runs on. Everything below is written from the point where the sale either lands in your dashboard or quietly doesn't.

The short answer

You don't need a website to earn a commission. You need permission to promote, a place to publish, and a way to know what happened.

Without a website you keep the second one. You have a harder time with the first. And you effectively give up the third.

Whether that trade is fine depends entirely on which stage you're at:

	Starting out	Once it is working
What decides your income	Whether anyone trusts you	Which offers convert and why
What you need to see	Very little. Did anything sell?	Where the drop-off is, per offer
Cost of having no site	Near zero	High, and it compounds
Right call	Start without one	Build one

So the honest recommendation: start without a website, and treat that as a decision with an expiry date rather than a permanent position. The people who get hurt aren't the ones who start without a site. They're the ones still without one two years later, having built their entire income on a channel that can restrict them without explanation.

Obstacle one: getting approved is harder than getting started

The first thing that actually blocks people isn't traffic. It's the application form. And this is where the "you don't need a website!" articles quietly mislead.

Most affiliate programmes and networks ask where you intend to promote. Some accept a social profile or a channel. Some require a live site with real content. Some accept you provisionally and then apply a performance condition. That last category is the one to understand, honestly, because it has a clock on it.

Amazon's Associates programme is the clearest example. You can apply with a channel rather than a site. But the account only becomes permanent once you've referred three qualifying sales, completed, shipped, paid-for orders from real referrals, not your own purchases, within 180 days of applying. Miss that window and the application is withdrawn. You start over.

A beginner who signs up on day one, spends four months learning to make videos, and starts promoting in month five has failed a test they didn't know they were sitting.

I've watched this happen more than once. It always feels like bad luck to the person it happens to. It isn't. It's a policy written on the sign-up page that nobody read because the sign-up page felt like a formality.

What to do about it, in order:

Apply to the programme after you have published, not before. Ten honest pieces of content on a channel with a pulse turns a rejection into an approval far more often than an empty profile and a promise does.
Read the acceptance criteria before you pick the offer, not after. Which is half of what choosing the right affiliate program is really about. Approval odds and payout terms decide more of your outcome than commission rate does.
Start the clock deliberately. If a programme applies a time-boxed condition, apply when you're ready to promote, not when you're ready to learn.
Have a second programme approved in the same niche. Single-programme dependency is the most common way a beginner's income goes to zero overnight. It costs nothing to avoid.

Obstacle two: without a site, you are blind in the middle

This is the real cost. It's the one that gets discovered late.

A website isn't primarily a place to publish. It's a measurement point in the middle of the funnel. Remove it and your entire view of what happened collapses into two numbers you didn't generate.

With a site, the chain looks like this. Your content, your page, the merchant, the sale. You own the middle step. So you can see how many people arrived, how many clicked through, and therefore whether a weak result was a traffic problem or a persuasion problem.

Without a site: your content, the merchant, the sale. The only numbers you get are the ones the network reports. They describe the part of the journey you have the least control over.

Practically, here is what you lose:

You cannot tell a traffic problem from a conversion problem. Two hundred clicks and no sales means either the wrong people clicked or the offer page didn't persuade them. Those have opposite fixes. Without a middle step you're guessing which one you have.
You cannot retarget. No page, no pixel, no audience of people who showed interest. For paid strategies this isn't a limitation, it's a disqualification.
You cannot test anything except the content itself. Every variable that lives between the click and the merchant is now someone else's.
You have no independent record. This is the one that eventually costs money, and it deserves its own section.

Obstacle three: the payout dispute you cannot win

Sooner or later a network will report fewer conversions than you believe you generated. And the entire question becomes whether you have any evidence of your own.

If you have your own click data, this is a conversation. Here is my click count for that window, here is yours, the gap is X, let us find out where it went. Discrepancies are normal. Cookie blocking, tracking that broke on one browser, a landing page that started redirecting, an attribution window that expired. They're also usually fixable once someone can see them. Affiliate tracking breaks quietly covers what the common causes look like from the inside.

If you have no data of your own, it isn't a conversation. It's you accepting a number. Because the only party with a record is the party paying you.

You can't even tell the difference between "nobody bought" and "the tracking stopped working in March". They look identical from where you are standing. A zero is both a valid answer and a broken pipe.

And you find out which one you had, if ever, only after you've stopped promoting because you assumed the offer was dead. By then whatever was working has stopped, because you stopped feeding it.

So build the one piece of infrastructure that gives you a record without building a website: a link layer you control.

The cheapest version is a link shortener or redirect service where you own the account and can export click data. Your content points at your link. Your link redirects to the affiliate link. It costs nothing meaningful. Takes an afternoon. And buys you the single most valuable thing a website was giving you, which is a click count that the network didn't produce.

Then, monthly, do the reconciliation:

Export your click count per link for the period.
Pull the network's reported clicks for the same period.
Compare. A gap of a few percent is normal. A gap of thirty percent is a broken redirect, a stripped parameter or a platform rewriting your URL.
Write the numbers down somewhere permanent. "It looked about right in June" is worth nothing in October. A logged number is a baseline you can argue from.

That routine is the whole of tracking affiliate sales as a complete beginner compressed into four steps. And it's the thing that separates people who get paid correctly from people who assume they did.

Where to actually publish, and what each place costs you

Pick the platform on buyer intent and durability, not on reach. Reach is the metric platforms advertise because it's the one they control.

Platform	What it gives you	What it costs you
YouTube	The highest purchase intent available without a site. "Best X for Y" is a buying query, and the description holds real links	Slow to produce; the archive is the asset, so results take months
Newsletter	The only audience you actually own; links work normally; deliverability is yours to manage	You must build the list from another channel first
Short video	Fastest reach, cheapest to test	Link handling is restricted; intent is low; almost nothing is archived or searchable
Communities (Reddit, Quora, forums)	Extremely high intent. People arrive already asking the buying question	Most ban or heavily restrict affiliate links; breaking the rule ends the account
Medium / LinkedIn articles	Searchable, long-form, holds a proper honest review	You are building an asset on someone else's domain, and the reach is theirs to withdraw

Two things that table makes obvious once you see it laid out.

Search-shaped platforms beat feed-shaped ones for this specific job. A feed post is consumed once and gone. A video answering "best budget X for Y" is found by someone who has already decided to buy, for years. Affiliate income is a function of intent far more than of audience size. Which is why a small channel on buying queries routinely out-earns a large one on entertainment.

Community platforms are the trap. They have the best intent on the list and the strictest rules. The failure mode isn't a warning. It's a ban that also erases the reputation you spent months building. If you use them, use them to be genuinely useful and to grow the newsletter. Not to drop links.

Which is the thread running through the whole table. Whatever you choose, run an email list alongside from the first month. Every row above is rented. The list is the one asset that survives a platform decision. It's also the only place where you control the link, the timing and the record.

Disclosure is stricter without a website, not looser

You must disclose the commercial relationship. Doing it without a site takes more care rather than less. Because the places people hide disclosures are exactly the places that don't count.

The FTC's guidance for creators is direct about this. A disclosure has to be hard to miss, in the content itself, and understandable at the moment someone encounters the recommendation. Its Disclosures 101 for Social Media Influencers and the longer Endorsement Guides FAQ spell out the failure modes. A disclosure only in a bio. Only in a description nobody expands. Buried behind a "more" link. Or spoken so fast in a video that it doesn't register.

Programmes layer their own requirements on top. Amazon's Operating Agreement requires associates to state clearly and prominently that they earn from qualifying purchases.

The practical version for someone with no website:

In video: say it out loud, early, and put it on screen. Not only in the description.
In a post or caption: put it before the fold, in the first line or two, not after three paragraphs and a "…more".
In a newsletter: one plain line above the recommendation.
Everywhere: in plain words. "I earn a commission if you buy through this link" beats any abbreviation. It costs you nothing. Readers who trust you don't mind, and the ones who would have minded were never going to buy.

Treat this as reputational infrastructure rather than compliance. Disclosure is one of the few signals a stranger has that you aren't hiding anything. Undisclosed links are the fastest way to lose the trust the whole model runs on.

The first ninety days, in order

This is the sequence I would follow starting from zero today. The ordering matters more than any individual step.

Days 1–14: pick the niche and the platform. One of each. Choose a niche where you can honestly answer buying questions and a platform you'll still be using in a year. Don't join any programme yet.

Days 15–45: publish ten pieces with no affiliate links at all. This feels backwards. It's the highest-leverage thing on the list. It builds the body of work that gets your applications approved. It tells you whether you can sustain the format. And it means your first recommendation arrives to people who already found you useful rather than to strangers.

Days 46–50: set up the link layer and apply. Redirect links you control first. Then the two or three programmes. Apply with the ten published pieces attached, and start any performance clock deliberately.

Days 51–90: publish buying-intent content and reconcile monthly. "Best X for Y". Honest reviews. "What I actually use and what I stopped using". Then run the four-step reconciliation above at the end of each month, and write the numbers down.

What this sequence deliberately avoids is the standard beginner order. Join programmes, get links, post links, wonder why nothing sells. Trust is the input and the links are the output. Done in the other order, the links are just spam with your name attached. That inversion is the first of the seven mistakes that quietly kill beginners.

When to stop being website-less

Build the site when one of these three becomes true. And not before. A site built too early is a distraction with a hosting bill.

You can name the two offers that produce most of your commission. That's the signal you have something worth optimising. Optimising is exactly what you can't do without a middle step you own.

You have had your first payout discrepancy. Once real money has gone missing once, the cost of not owning your own data has stopped being theoretical.

You want organic search traffic. This is the hard boundary. Search is the most durable acquisition channel in affiliate marketing. It compounds. It doesn't depend on you posting today. And it's the one channel genuinely unavailable without a site of your own. SEO basics only start applying at the point where you have a page to rank.

The migration is much easier than starting there. By then you know which content converts and you're rebuilding around evidence rather than guesses. Which is why starting without a website is good advice and staying without one is not.

The mistakes I see most

Treating the network's dashboard as the truth. It's one party's record of a transaction you also participated in. Keep your own count. Or accept whatever you are told forever.

Promoting before publishing. Links added to an account with no body of work behind them convert at approximately nothing. And they burn the trust you were about to build.

Depending on one programme. Terms change. Categories get recategorised. Commission rates get cut. Accounts get closed for reasons nobody explains. Two approved programmes in the same niche is cheap insurance.

Chasing commission rate over conversion rate. A high rate on something nobody buys pays less than a modest rate on the obvious choice in the category. Beginners consistently pick the wrong one of these. How much you can really make is mostly a story about this.

Recommending things you have not used. It's detectable. It's the fastest route to a reputation you can't repair. And the alternative, being honest about what you actually use and what you stopped using, outperforms it anyway.

Never checking whether tracking still works. Links rot. Platforms rewrite URLs. Programmes migrate networks. Redirects break silently. A link that stopped tracking in March looks exactly like an audience that stopped buying in March.

FAQs

Can I really earn without a website, or is that just marketing?

You can genuinely earn without one. Commission is paid on referred sales, and nothing about that requires a domain of your own. What's oversold is the idea that it costs you nothing. It costs you attribution visibility, retargeting, testing, and any independent record at payout time. Those are affordable at the start and expensive later. Which is why the honest framing is "start without one" rather than "you never need one".

Which platform is best if I have no website?

YouTube if you can sustain video. A newsletter if you can write. Realistically both, because they solve different problems. One brings strangers with buying intent, the other is the only audience you own. If you can only do one thing, pick the one you will still be doing in twelve months. Consistency over a year beats platform choice by a wide margin. The archive is the asset.

Do I need money to start?

No, other than time. Content-based affiliate marketing on video, social or a free newsletter has no meaningful cash cost. A link redirect layer is free or close to it. Paid traffic is a different discipline entirely and it's the wrong place to start. It requires knowing your conversion rate and payout timing before you spend, and you can't know either on day one.

How long until the first commission?

Plan on three to six months of publishing before anything regular. Treat an early sale as encouragement rather than a trend. The dominant variable is buying intent, not volume. Ten pieces answering questions people ask right before purchasing will outperform a hundred pieces of general content. If nothing has sold after six months of intent-shaped content, the problem is usually the offer or the audience match, not the effort.

What happens if my account gets restricted?

Whatever you built on that platform stops earning immediately. You generally will not get a detailed explanation or a fast appeal. This is the structural risk of having no site. The only real mitigation is that email list, an audience you can contact without a platform's permission. Everything else is a rental agreement that can be terminated by the landlord.

Is affiliate marketing without a website worth doing at all?

Yes, for the reason most people miss. It strips the work down to the only skill that actually matters, which is making a recommendation a stranger believes. A website adds measurement, durability and search. It adds all of them to whatever persuasion ability you already have. If that ability is zero, a site multiplies zero. Start where the skill is, then build the infrastructure around it. The longer view is in the practical affiliate guide.

Key takeaways

You can start without a website. What breaks isn't earning. It's approval, attribution and evidence.
Programme approval is the first real obstacle. Publish ten pieces before you apply, and know whether the programme puts a clock on your application.
Without a site you lose the middle of the funnel. So you can't tell a traffic problem from a conversion problem.
Build a redirect link layer you control. It's the cheapest way to own a click count the network didn't produce.
Reconcile your clicks against the network's monthly, and write the numbers down. A logged number is the only thing you can argue from.
Choose platforms on buyer intent, not reach. Search-shaped beats feed-shaped for affiliate income.
Community platforms have the best intent and the strictest rules. Be useful there. Don't drop links.
Disclose in the content itself, in plain words. Bio-only and description-only disclosures don't count.
Run an email list from month one. Every other channel is rented and can be withdrawn without explanation.
Build the site when you know which two offers convert, when a payout goes missing, or when you want search traffic.

Affiliate Marketing: The Practical Guide I Wish I Had on Day One

Prabhash Jha — Sun, 16 Aug 2026 04:55:33 +0000

Summary

Affiliate marketing is getting paid a commission for driving a sale you didn't have to fulfil. You send a qualified buyer to someone else's product through a trackable link; when they buy, you earn.

The people who make real money at it don't “post links.” They build an audience or a traffic system, earn trust, recommend things they'd stake their name on, and treat it like a business with numbers — not a side hustle with hope.

This guide covers how it actually works, the exact model I use to pick offers, a step-by-step way to start with zero audience, the mistakes that cost me money, and a checklist you can act on today.

The problem

Most people who try affiliate marketing quit within 90 days. Not because it doesn't work — it's a large, fast-growing industry for a reason — but because they were sold the wrong version of it.

The wrong version says: grab a link, spray it across WhatsApp and Instagram, wait for passive income. That version fails because it skips the only two things that actually produce commissions — trust and traffic.

The right version is boring and it compounds: pick a niche, build one channel that reliably reaches buyers, recommend products you genuinely understand, and measure everything. This guide is the boring version. It's the one that pays.

Background: how affiliate marketing actually works

There are four players in every affiliate transaction:

The merchant — the company that owns the product (an e-commerce brand, a SaaS tool, a course creator).
The affiliate — you. You recommend the product.
The network or platform — the tech layer that tracks clicks and sales and handles payouts (Amazon Associates, Impact, CJ, Cuelinks or vCommission in India, or a brand's in-house program). Which one you join matters more than it looks — how to choose an affiliate program is the checklist I use.
The customer — the person who clicks your link and buys.

The mechanics:

You get a unique tracking link with your affiliate ID.
A customer clicks it. A cookie is dropped on their device with an attribution window (for example, 24 hours on Amazon, 30–90 days on many SaaS programs).
If they buy within that window, the sale is attributed to you.
You earn a commission — a percentage of the sale (physical goods are often 1–10%) or a flat bounty (SaaS and finance can be much higher per sale).

Commission models you'll encounter:

CPS (cost per sale) — paid when a purchase happens. Common in e-commerce and courses.
CPL (cost per lead) — paid for a qualified signup or form. Common in finance, insurance and B2B.
CPI (cost per install) — paid for an app install. Common in mobile apps.
Recurring — paid every month the customer stays. Common in SaaS and subscriptions.

The single most important concept: recurring and high-ticket beats high-volume-low-value for almost everyone. One recurring SaaS sale can out-earn hundreds of tiny physical-product commissions — and take less traffic to get there.

The framework: the Offer × Traffic × Trust model

Every affiliate income stream is three variables multiplied together. Weak in any one, weak overall.

Income ≈ Offer quality × Traffic quality × Trust

1. Offer — would I recommend this to my own family? Pick offers where the product is genuinely good, the commission justifies the effort, the attribution window is fair, and the payout is reliable. Score every offer on relevance, payout size, recurrence, and the conversion rate of the merchant's page. A great product with a terrible landing page still won't convert — you're borrowing their checkout.

2. Traffic — do I reach people at the moment they're deciding? The best affiliate traffic has buying intent. From highest to lowest intent:

Search (“best CRM for small business”) — highest intent, best ROI, compounds.
YouTube reviews and comparisons — high intent, high trust.
Email list — you own it, converts best.
Social content — volume, lower intent.
Paid ads — fast, but you're buying margin; only works with strong economics.

3. Trust — do they believe me? Trust is why the same link converts far better from a respected reviewer than from a stranger. You build it by being useful before you sell, disclosing affiliate relationships openly, and only recommending what you've used or vetted. Trust is the moat competitors can't copy.

How to use the model: don't start by picking a product. Start by picking a niche where you can build trust and reach intent-driven traffic — then find the best offer in it.

Examples: three models that actually work

I've watched these three models play out again and again. None of them rely on a secret tactic — they're about picking the right game.

Model 1 — Content + SEO (the compounding play). A single well-researched “best [X] for [use case]” article that ranks in search can keep earning commissions for years after you publish it, with almost no ongoing cost. This is the model I trust most, because the work compounds — last year's article still pays this year. If you're starting with more time than money, this is where I'd tell you to begin.

Model 2 — High-ticket / recurring (the leverage play). Instead of chasing hundreds of tiny e-commerce commissions, you focus on a few products that pay well — especially SaaS tools with recurring commissions. Here's the shape of why it works: refer a handful of customers to a tool that pays you a recurring cut, and that income repeats every month and grows as you add more — while a physical-product affiliate has to earn every sale again from scratch. Same effort, very different ceiling. (That's the logic, not a promise — real results depend entirely on your niche and traffic.)

Model 3 — Paid media (the fast-but-tight play). You buy traffic and send it to a review or bridge page for an offer. It's the fastest to scale and the fastest to lose money on. This is where a performance-marketing instinct matters most: it only works when the unit economics are proven first — what a click costs you versus what a conversion pays back. Beginners almost always run this before they're ready, and the market quietly takes their money. Earn the right to run paid traffic by proving the funnel converts for free first.

Mistakes I learned from

These are the lessons that shaped how I work now. I'm sharing the principle, not the paperwork — but each one cost me something to learn.

Promoting for the commission, not the customer. The most tempting trap is pushing whatever pays the most, regardless of whether it's actually good. It works once. Then the refunds, the silence, and the lost trust show up — and trust is the entire asset. My filter now is simple: would I recommend this to a friend who trusted me? If not, I pass, no matter the payout.
Chasing volume over intent. Early on I optimised for clicks and traffic numbers because they felt like progress. They're not. Clicks don't pay; qualified buyers do. A hundred people actively deciding what to buy are worth more than ten thousand who are just curious.
Ignoring the merchant's conversion rate. You're borrowing someone else's checkout. I've sent genuinely good traffic to a weak landing page, watched it die, and blamed the traffic. Now I check that the merchant's page actually converts before I send it real volume.
Not owning my audience. Anything you build on a platform, the platform can take away the day it changes its rules. The lesson I wish I'd acted on sooner: build an email list from day one. It's the one audience no algorithm can switch off.
No tracking. If you can't see which piece of content, link, or source drives an actual sale, you're guessing. Set up tracking before you scale anything — especially before you spend.
Diversifying too early. Five half-built channels earn less than one finished one. I spread myself thin chasing every shiny format before finishing the first. Win one channel completely, then expand.

Step-by-step: how to start with zero audience

Pick a niche you can be useful in. The intersection of what you know, where there are buyers, and where products pay commissions.
Pick one traffic channel. Best for beginners with no budget: SEO content (a simple blog) or YouTube. Both build trust and compound. Choose the format you'll actually keep doing for 12 months.
Find 3–5 offers before you write a word. Join relevant programs. Check commission, recurrence, cookie window and payout threshold — and, critically, whether the product is genuinely good. Prioritise recurring and high-ticket.
Create buyer-intent content. The formats that convert: “best [X] for [use case]”, “[product] review”, “[product A] vs [product B]”, “how to [outcome] with [product]”. Be honest — include downsides. Honesty converts better than hype.
Add your links naturally and disclose. Place links where the reader is deciding. Add a clear disclosure — it's legally required in many regions and it builds trust rather than reducing it.
Build an email list in parallel. Offer a simple lead magnet such as a checklist or template. Every visitor you capture is an asset you own and can recommend to repeatedly.
Track, then double down. Use sub-IDs and analytics to see what drives sales. Kill what doesn't; do more of what does.
Only then consider paid traffic. Once you have one proven converting page and know your numbers, paid can pour fuel on the fire. Not before.

Checklist

Copy this before you launch anything.

Niche chosen at the intersection of knowledge + buyers + commissions.
One traffic channel chosen — and I'll commit to it for 12 months.
3–5 affiliate offers vetted for quality, payout, recurrence, cookie window and payout threshold.
Verified the merchant's landing page actually converts.
First three buyer-intent pieces planned (“best”, “review”, “vs”).
Affiliate disclosure written and placed.
Email capture + lead magnet live.
Tracking set up (sub-IDs + analytics) before any spend.
One number I'm optimising for (for example, commission per 1,000 visitors).

Resources

Networks to know: Amazon Associates, Impact, CJ Affiliate, PartnerStack (SaaS); and in India, Cuelinks, vCommission and EarnKaro.
Tools worth knowing: Google Analytics for tracking, a link-management or cloaking tool, and a simple offer-scoring spreadsheet (there is a free one in the next bullet). Keep the stack minimal until volume justifies more. If you have never set tracking up, start with how to track affiliate sales as a complete beginner — and know the ways tracking breaks quietly before payout day.
Download the offer scoring sheet — free, no signup. It scores three offers side by side on the eight things that decide whether an offer is worth building content for, and then estimates what the winner is actually worth against the traffic you can realistically send. Works in Excel, Google Sheets and Numbers.

FAQs

Is affiliate marketing still worth it?

Yes — but the easy version is dead. Trust-based, intent-driven affiliate marketing is bigger than ever because buyers research before they buy. Link-spamming is what's dead.

How much can a beginner realistically earn?

For the first few months, often near zero — you're building the asset before it pays. How much you can really make with affiliate marketing goes into the shape of it properly. A realistic shape is: your first small commissions somewhere around month three to six, and meaningful, repeatable income by roughly month nine to twelve if you stay focused on one channel. Anyone promising fast passive income is selling you the dream, not the reality.

Do I need a website?

No, but it helps. YouTube, email, or even a strong social presence can work — here is how I'd start affiliate marketing without a website. A website (SEO) is the most durable because it compounds and you own it.

How much money do I need to start?

You can start with almost nothing using content, SEO or YouTube — just time. Paid-traffic affiliate marketing needs a budget and proven unit economics first.

Do I have to disclose affiliate links?

Yes. You're generally required to disclose affiliate relationships (the FTC's Endorsement Guides in the US, ASCI guidance in India). Disclosure also builds trust — always do it.

Affiliate marketing vs dropshipping — which is better?

Affiliate marketing means no inventory, no fulfilment and no support — you trade higher control for lower operational headache. Dropshipping gives more control and margin but you own fulfilment and support. Affiliate is the lower-risk entry point.

Key takeaways

Affiliate marketing = a commission for driving a sale you don't have to fulfil.
Income = Offer quality × Traffic quality × Trust. Weak in one, weak overall.
Recurring and high-ticket beat high-volume-low-value for almost everyone.
Buyer intent is the highest-leverage traffic — SEO and reviews compound.
Own your audience (email). Track everything. Win one channel before expanding.
The boring, trust-first version is the one that actually pays.

Related reading: SEO basics for the traffic that compounds, how to build an email list from zero for the audience you actually own, and performance marketing: the practical playbook for the paid side of the same economics.

7 Affiliate Marketing Mistakes That Quietly Kill Beginners

Prabhash Jha — Sun, 16 Aug 2026 04:55:32 +0000

Affiliate marketing does not fail the way beginners expect it to. There is rarely a moment where something snaps and you know it. The traffic keeps arriving, the clicks keep registering, the article keeps getting read — and the commissions just do not turn up. Nothing is visibly on fire. That is precisely the problem.

Paid media tells you when you are wrong. Money leaves the account, conversions do not come back, and the loss sits on your screen in a colour you cannot ignore. Affiliate is the opposite. You are paid on someone else's tracking, through someone else's checkout, on someone else's payout schedule. When any one of those three quietly stops working, your own dashboard looks exactly like a slow month.

So the useful question for a beginner is not "what is the best affiliate strategy". It is: what is broken right now that I cannot see? Almost every first-year failure is an answer to that question, not a strategy problem.

Here are the seven I see most — with the cause, the symptom you will actually notice, and the fix.

Mistake 1: Promoting for the payout instead of the fit

Pushing whatever pays most, regardless of quality, works exactly once — then the refunds and the lost trust arrive. That is the version everyone repeats. The mechanics underneath it are worth spelling out, because they are what actually costs you money.

The cause. Affiliate networks rank offers by payout, so the first thing you see is the biggest number. Nothing on that screen tells you the refund rate, the chargeback window, or whether the product suits the people who read you.

The symptom. Your reported sales and your paid sales stop matching. Sales appear in the dashboard, then reverse 30 to 45 days later as returns and cancellations clear. Your "earnings" line looks healthy and your bank account does not.

The fix — three checks before you promote anything. First, would you send this to a friend who trusted you? Second, what is the refund or cancellation rate, and will the merchant tell you? A merchant who refuses to answer has answered. Third, does the payout survive the reversal window — that is, are you being paid on the sale, or on the sale that sticks?

Then there is the quiet version of this mistake: taking the one-off payout over the recurring one. Beginners almost always pick the bigger number today, because on day one it is genuinely bigger. Here is what that choice costs.

Worked example: one-off versus recurring

Two programmes in the same category, promoted to the same audience, converting at the same rate — you refer 10 customers a month, every month.

Programme A pays ₹4,000 once per sale. Nothing after that.
Programme B pays ₹600 a month for as long as the customer keeps paying (30% of a ₹2,000/month subscription).

Assume, for clarity, that nobody churns in the first year. Programme A earns a flat 10 × ₹4,000 = ₹40,000 every month. Programme B earns ₹600 × the number of customers you have accumulated, so month n pays ₹6,000 × n.

Month	A monthly	A cumulative	B monthly	B cumulative
1	₹40,000	₹40,000	₹6,000	₹6,000
6	₹40,000	₹2,40,000	₹36,000	₹1,26,000
7	₹40,000	₹2,80,000	₹42,000	₹1,68,000
12	₹40,000	₹4,80,000	₹72,000	₹4,68,000
13	₹40,000	₹5,20,000	₹78,000	₹5,46,000
18	₹40,000	₹7,20,000	₹1,08,000	₹10,26,000

The recurring programme overtakes on monthly income in month 7 and on total income during month 13. By month 18 it has paid roughly 43% more in total (₹10,26,000 against ₹7,20,000), and the gap widens every month after that because Programme A resets to zero with each new month while Programme B compounds.

Now make it realistic. At 5% monthly churn, Programme B's month-12 monthly figure falls from ₹72,000 to roughly ₹55,000 — still well above A's flat ₹40,000, and still compounding. Churn does not reverse the conclusion; it just moves the crossover a few months right. The lesson is not "recurring always wins" — it is that a one-off payout has to be roughly 12 to 18 months of the recurring one to be worth taking, and almost none of them are.

If you are choosing between programmes right now, work through a proper selection checklist rather than sorting by commission.

Mistake 2: Chasing volume when what you need is intent

Vanity numbers feel like progress but they do not pay. Reach, followers, impressions and pageviews are all upstream of the only number that matters — someone deciding to buy, today, and clicking your link to do it.

The cause. Volume metrics move fast and feel responsive. Intent-based traffic grows slowly and feels like nothing is happening for weeks. So beginners optimise the metric that gives feedback rather than the one that gives money.

The symptom. Thousands of sessions, a click-through rate that looks fine, and a conversion rate under 0.5%. If a lot of people click your link and almost nobody buys, you have not got a page problem — you have got an audience problem. Those readers were never in a buying frame.

The fix. Sort your content by where it sits in the buying decision, not by traffic. "What is X" articles bring volume and convert at almost nothing. "Best X for Y", "X vs Y", "X pricing" and "X alternatives" bring a fraction of the traffic and carry nearly all the revenue. One comparison page written for someone with a credit card open will outperform ten explainer posts. If the funnel logic is not familiar, this walkthrough of a marketing funnel is the shortest route in, and SEO basics covers how to find those queries.

The same mistake shows up as spreading too thin. Five half-built channels earn less than one finished one. A blog with 40 posts, a YouTube channel with 6 videos, an Instagram account posted to twice a month and a newsletter with 30 subscribers is not diversification — it is four things below the threshold where any of them work. Win one channel, then expand.

Mistake 3: Hiding the fact that you get paid

Hiding affiliate links breaks trust and often breaks the rules. Done properly, disclosure actually lifts conversions, because a reader who knows how you are paid stops looking for the catch and starts evaluating the recommendation.

The cause. Beginners assume disclosure kills conversions. It does the opposite when it is phrased as a reason to trust you rather than an apology.

The symptom. This one has no symptom until it becomes a serious problem — a platform strike, an ad network rejection, or a regulator's notice. That is what makes it dangerous.

The fix. Put the disclosure above the recommendation, not in the footer. One line, in your own words, stating that you earn a commission if the reader buys and that it costs them nothing extra. In India, the ASCI influencer guidelines require the label to be upfront and prominent — not buried in hashtags, not below a "read more" fold, and in the same language as the content.

Two categories need more care than a disclosure line:

Insurance. Soliciting or procuring insurance business for a commission is a regulated activity. Doing it lawfully requires IRDAI registration as an agent or intermediary — an affiliate link paying you per policy sold is not exempt just because it sits in a blog post. Take proper advice before you touch insurance offers.
Securities and investing. Educating people about how markets, index funds or demat accounts work is fine. Recommending specific stocks or securities, or providing investment advice for consideration, requires SEBI registration as an Investment Adviser or Research Analyst. Broker referral programmes exist and are legitimate, but they carry the broker's own compliance conditions — read them before you write the post, not after.

Mistake 4: Sending traffic to a page you have never bought from

You are borrowing someone else's checkout. A bad page kills good traffic, and you will be blamed for the result by everyone including yourself.

The cause. You evaluate the merchant's landing page as a reader, on a laptop, on fast broadband. Your audience meets it as a buyer, on a mid-range Android phone, on patchy 4G.

The symptom. Healthy clicks, poor conversion, and the merchant's own funnel report — if you can get it — showing drop-off at checkout rather than at the product page.

The fix. Go through the entire purchase yourself before you promote it. Specifically check: how long the page takes to load on mobile data; whether UPI is offered alongside cards; whether the form demands GST details or a company name from a consumer buyer; how many steps sit between "buy" and "paid"; and whether there is an empty coupon field on the checkout page.

That last one matters more than it looks. An empty coupon field sends your reader off to a coupon site to hunt for a code. They come back through the coupon site's link, and the last-click rule hands your commission to that site. You did the work, someone else got paid, and nothing in your dashboard explains why.

Mistake 5: Building an audience you do not own

Building only where the platform can change the rules overnight is the mistake that shows up all at once, usually at the worst time.

The cause. Followers are easier to get than email addresses, so beginners take the easy metric. But a follower is a permission granted by a platform, and the platform can revoke it — through a reach change, a policy update, or an account action you did not see coming.

The symptom. A month where your content performs exactly as it always has and reaches a third of the people. There is no notification for this.

The fix. Start collecting email addresses from your very first piece of content, before you think you deserve to. You do not need a sophisticated offer — a genuinely useful checklist or comparison sheet related to the post someone is already reading will outperform a generic "subscribe for updates" box many times over. Building a list from zero is the mechanics of it.

Email also solves the affiliate-specific problem that platforms cannot: it lets you follow up. Most affiliate purchases do not happen on the first visit. A reader who joins your list on a comparison post and buys six weeks later is invisible to every social platform and completely visible to you.

Mistake 6: Tracking you set up once and never verified

If you cannot see what drives a sale, you are guessing. The version that actually kills beginners is subtler: tracking that worked when you set it up, broke three months later, and never told you.

The cause. Affiliate tracking has more moving parts than any beginner realises — a redirect, a cookie, a cookie duration, an attribution rule, a merchant-side conversion tag, and a payout file. Each one is owned by someone else and can change without notice.

The symptom. Earnings that flatten while traffic holds steady. Not a crash — a plateau. That is why it survives so long: a plateau looks like a marketing problem, so people respond by making more content, which produces more untracked clicks.

The fix. Verify rather than assume, on a schedule. Use this as a diagnostic:

What you see	What it usually is	How to check it
Clicks in your analytics, far fewer in the network dashboard	Redirect or link cloaker dropping the affiliate parameter	Click your own link and inspect the final URL for your affiliate ID
Clicks recorded, zero conversions for weeks	Merchant-side conversion tag removed or offer paused	Make a small test purchase and confirm it appears in your report
Conversions appear, then reverse	Refunds, cancellations, or fraud screening	Ask the merchant for the reversal rate before you scale
Desktop converts, mobile does not	In-app browsers or tracking prevention dropping the cookie	Test the full path inside Instagram and YouTube in-app browsers
Sales from one page only, never the others	Missing or duplicated subID on your other links	Give every placement its own subID and check they all report separately
Earnings flat while traffic grows	Coupon extensions overwriting the last click	Check whether the merchant page has an open coupon field

Two structural points to settle early. Cookie duration decides how long after a click you still get paid — 24 hours and 90 days are wildly different businesses, and it changes what content is worth writing. Attribution rule decides who wins when several affiliates touch the same buyer; almost every programme is last-click, which is why mistake 4 costs real money.

Set a unique subID per placement from the start. Without it, you know that sales happened but not which article produced them, so you cannot tell which content to make more of. Affiliate tracking breaks quietly goes deeper on the failure modes, and if the underlying metrics are still fuzzy, CPC, CTR, CPA and ROAS in plain English is worth 10 minutes.

Mistake 7: Links you never go back and check

Every affiliate link you have ever published is a small piece of infrastructure that can rot. Most beginners publish a link and never look at it again.

The cause. Publishing feels like finishing. But merchants discontinue products, restructure URLs, migrate domains, close programmes, run out of stock, and geo-restrict offers — and none of them will email you when they do.

The symptom. An article that used to earn slowly stops. Or, worse, everything keeps looking normal because the link still resolves — it just lands on a generic homepage that pays nothing, or redirects through a chain that strips your tracking parameter on the way.

The fix. Once a month, run a link check across your whole site, not just your newest posts. You are looking for four things: hard 404s; soft failures where the link resolves to a homepage or a "product no longer available" page; redirect chains that drop your affiliate ID; and offers that have quietly expired inside the network. Do the same for the links in your email archive and your YouTube descriptions — old content keeps getting found, and dead links in it convert at zero forever.

Keep a single sheet of every live link: destination, programme, commission type, cookie duration, subID and last-checked date. It takes an afternoon to build and it turns a vague worry into a 20-minute monthly task. The same discipline applies to the rest of your numbers — one sheet that tells you whether your marketing makes money is the habit this sits inside.

The pattern behind all seven

Notice what these have in common: every one is a shortcut around trust or measurement. Slow down on those two and you skip the entire list. Recommend honestly, disclose openly, measure everything, and own your audience — that is the beginner who is still standing a year later.

It is worth being explicit about why these specific mistakes are the ones that do the damage. They are all silent. A bad headline tells you it is bad. A broken tracking link, an expired offer, a coupon extension eating your last click and a merchant checkout that fails on mobile all present as the same thing: a normal-looking week that earns less than it should. There is no alert. The only defence is a routine — a monthly hour where you click your own links, make a test purchase, reconcile your click count against the network's, and check what reversed.

That hour is not admin. In affiliate marketing it is the highest-paid hour in your month, because it recovers income you have already earned and are currently not being paid for.

FAQs

What's the single biggest affiliate marketing mistake?

Promoting for the payout instead of the person. It costs you the one thing you cannot quickly rebuild: trust. The financial damage is recoverable — a bad month is a bad month — but an audience that has learned your recommendations follow the commission will discount everything you publish afterwards, including the honest recommendations.

How long until affiliate marketing actually works?

Often near zero for the first few months while you build the asset. Meaningful, repeatable income is realistic by roughly month nine to twelve if you focus on one channel. Fast-income promises are a red flag. The gap exists because search traffic and audience trust both compound slowly, then both pay out at once.

Why am I getting clicks but no affiliate commissions?

Work through it in order. Click your own link and confirm your affiliate ID survives to the final URL. Make a small test purchase and check it registers. Test the whole path in an in-app browser on a phone. Check the merchant's checkout for an open coupon field. Most "no commissions" problems are one of those four, not a content problem.

Do I have to disclose affiliate links in India?

Yes. ASCI's influencer guidelines require a clear, prominent, upfront disclosure in the same language as the content — not hidden in hashtags or below a fold. Beyond disclosure, insurance and securities offers are separately regulated: insurance commission requires IRDAI registration, and investment recommendations require SEBI registration.

How often should I check my affiliate links?

Monthly for a full sweep across the whole site, plus an immediate check any time a merchant emails about a programme change, a site migration or a product being discontinued. Also check after you redesign or migrate your own site — that is when cloaked links and redirects most commonly lose their tracking parameters without any visible error.

Is it better to promote one high-paying product or several small ones?

Neither framing is right. Promote the product that fits the buying decision your reader is actually making, then prefer recurring commission over one-off where both exist. A recurring payout typically overtakes a one-off one within 7 to 13 months of steady referrals, and keeps compounding after that.

Key takeaways

Affiliate marketing fails silently — the dangerous mistakes look identical to a slow month, so you have to go looking for them on a schedule rather than waiting to be told.
A one-off commission needs to be worth roughly 12 to 18 months of the recurring alternative to be the better choice, and very few are.
Volume metrics and half-built channels both feel like progress; buyer intent and one finished channel are what actually pay.
Disclosure lifts conversions when it is placed above the recommendation, and in India insurance and securities offers carry IRDAI and SEBI obligations well beyond a disclosure line.
Test the merchant's full checkout on a phone yourself — an open coupon field alone can hand your last-click commission to a coupon site.
Verify tracking rather than trusting it: unique subIDs per placement, a monthly link sweep, and one test purchase will recover income you have already earned.