Summary
Performance marketing is paid marketing where the only thing that counts is a measurable outcome. A lead, a sale, an install. Not impressions. Not applause. You put money in, you track exactly what comes out, and you scale what's profitable. This is how I actually run it. The economics that decide whether it works, the metrics that matter, the mistakes that cost me, and a step-by-step way to start without lighting money on fire.
The problem
Most people "do ads" and hope. Performance marketing is the opposite of hope. It's a system where every rupee is accountable.
The trap beginners fall into is optimising for the wrong number. They chase clicks, reach and CTR. They scale the campaigns that look busy. They quietly lose money.
Or they panic. Two days in, they judge a campaign and kill the ones that would have worked. Both burn cash.
Honestly, the fix isn't a secret tactic. It's knowing which number actually decides whether you win.
Background: what performance marketing actually is
Performance marketing is any paid marketing where success is a tracked action and you can tie spend directly to that action. The channel varies. The accountability doesn't.
Google Ads: search, shopping, YouTube. You reach people already looking.
Meta Ads: Facebook and Instagram. You reach people by interest and behaviour, and the creative does the targeting.
Programmatic / display / native: automated buying across the web. This is the one channel here where you're buying through a chain of intermediaries rather than from a single counterparty, which is why it needs its own operating discipline before you put budget through it.
Affiliate: you pay partners per outcome. (More on this below.)
The terms you have to know (this is the language of the game):
CPC / CPM: what you pay per click or per thousand impressions.
CPA / CPL / CPS: cost per acquisition, lead, or sale.
Conversion rate: the share of visitors who take the action.
ROAS: revenue divided by ad spend. The headline efficiency number.
CAC: the full cost to acquire one customer.
LTV: what a customer is worth to you over their lifetime.
Payback period: how long until a customer has paid back what you spent to get them.
The framework: economics, then message, then scale
Performance marketing works when three things are true. And they only work in this order.
Profit = (LTV − CAC) × Volume. Fix the economics before you ever touch volume.
1. Economics first. Do you make more from a customer than it costs to get one? If your LTV isn't comfortably bigger than your CAC, no amount of clever targeting saves you. You're just buying losses at scale. Know your numbers before you spend a rupee.
2. Message-market fit second. The right offer, to the right audience, with a message that converts. This lives in your creative and your landing page far more than in your targeting settings. A great product with a weak landing page still won't convert. You're paying to send traffic into a leaky bucket.
3. Scale last. Only once the economics work and the funnel converts do you pour in budget. Scaling a losing funnel doesn't fix it. It just loses money faster. Most people scale first and wonder why it breaks.
The metrics that actually matter
I watch a short list and ignore the rest:
CAC and LTV:CAC ratio: the real health of the machine.
ROAS against a break-even target: not ROAS in a vacuum. ROAS versus the number you actually need to profit.
Conversion rate: where most cheap gains hide.
Payback period: how fast your money comes back to reinvest.
Vanity metrics I ignore: impressions, reach, CTR on its own, and "engagement". They feel like progress. They don't pay the bills. Marketing metrics explained sets out what each of these actually measures, and what it doesn't.
The two numbers you need before you open an ad account
Almost every argument about performance marketing dissolves once these two are written down. And almost nobody writes them down.
Break-even ROAS = 1 ÷ gross margin. If you keep 40% of every sale after the cost of delivering it, you need 2.5x revenue for every 1x of ad spend just to get back to zero. Not 1x. The number of people running at "2x ROAS, so we're profitable" on a 40% margin is the single most common expensive mistake in the discipline.
Maximum affordable CAC = the gross profit a customer produces over their life with you. That's the ceiling. Anything you're willing to pay above it, you're paying out of capital rather than out of margin.
Everything else, bids, audiences, creative, placements, is an attempt to move your actual numbers toward those two. Without them you have no idea which direction is better.
A worked example: what "it's working" actually has to mean
Illustrative numbers throughout. Not a real account. The shape is what matters.
A business sells one product at ₹2,500 with a 40% gross margin, so each sale leaves ₹1,000 of gross profit. That immediately gives the two numbers above. Break-even ROAS is 1 ÷ 0.40 = 2.5x, and the most it can pay for a first purchase without losing money on it is ₹1,000.
It spends ₹200,000 in a month on search. Clicks cost ₹12, so that buys 16,667 clicks. The landing page converts at 1.2%, which is 200 sales.
- Revenue: 200 × ₹2,500 = ₹500,000
- ROAS: 500,000 ÷ 200,000 = 2.5x
- CAC: 200,000 ÷ 200 = ₹1,000
So the account is running at exactly break-even. On the platform dashboard this looks like a healthy, working campaign. Spend is stable, conversions are arriving, ROAS is 2.5x. In the bank account it has produced nothing at all.
Now the useful part: which lever? Two options. Worth comparing because they feel completely different and are arithmetically identical.
- Get clicks 20% cheaper: ₹12 down to ₹9.60. The same ₹200,000 buys 20,833 clicks, still converting at 1.2%. That's 250 sales, ₹625,000 revenue, ROAS 3.1x, CAC ₹800. Gross profit after ad spend: 250 × ₹1,000 − ₹200,000 = ₹50,000.
- Get the landing page converting 25% better: 1.2% up to 1.5%, clicks unchanged at ₹12. That's 16,667 clicks × 1.5% = 250 sales. Identical revenue, identical ROAS, identical profit.
Same result. That's the point. The difference isn't in the maths, it's in who owns the gain. The cheaper click is rented from an auction that your competitors can bid back tomorrow. The better landing page is yours. It applies to every channel you ever run, and nobody can outbid you for it. This is why the landing page is the first place I look and the ad account is the second. And why the biggest cheap wins in performance marketing are almost never in the campaign settings.
Then the question that decides whether any of it can scale. Suppose an average customer buys 1.8 times over their life with the business. LTV is 1.8 × ₹1,000 = ₹1,800, against a CAC of ₹1,000. That's an LTV:CAC of 1.8:1. Even in the improved version, at CAC ₹800, it's 2.25:1.
Both are under the 3:1 that gives you room to survive a rising CPM, a bad creative month or a seasonal dip. Which means the honest read on this account is: stop trying to scale it and go and fix retention or margin instead. No bid strategy fixes a 1.8:1 business. That verdict took nine lines of arithmetic and would have been invisible on any dashboard, indefinitely.
If you want this calculated for you rather than done by hand, the unit economics sheet is the version I keep.
Examples: three ways it plays out
The intent play (Google search). You bid on what people already type when they're ready to act. Which of the two big platforms to open first is its own decision. Google Ads vs Meta Ads works through it. Highest intent, usually the best place to start if people are actively searching for your solution. The demand already exists. You're just capturing it.
The creative-led play (Meta). Here the creative is the targeting. You're interrupting people who weren't looking, so you win or lose on the ad itself. The hook, the angle, the first three seconds. Test many creatives cheaply, then put budget behind the few that work.
The full-funnel play. Capture intent at the bottom, build demand at the top, and retarget the people who didn't convert the first time. If the stages themselves are unfamiliar, start with what a marketing funnel is. The channels compound instead of competing. Only attempt this once a single channel is already profitable.
Mistakes I learned from
Each of these cost me something to learn. I'm sharing the principle, not the paperwork. But a principle you can't recognise in an interface is useless, so each one below is written as the shape it actually has: what you see, what you check, and what it usually turns out to be.
1. Optimising for clicks instead of profit.
The shape: a Meta campaign where the reported CTR is the best in the account, the CPC is the lowest, and the conversion column is thin. It reads like the winner in every column you look at first. Because CTR and CPC are the columns platforms put in front of you.
What to check first: not the ad. The query or placement report. A cheap, high-CTR click is usually a cheap, high-CTR audience. Broad, curious, and nowhere near buying. On search, the same thing shows up as generic head terms and question queries outranking your commercial terms in click share.
The correct read: CTR measures how good the ad is at getting attention from whoever it reached. It says nothing about whether it reached buyers. A campaign can improve its CTR and worsen its CPA at the same time. Those two numbers aren't on the same side, and the direction that matters is CPA against your maximum affordable CAC.
What actually moved it: narrowing to commercial intent, which almost always makes CTR look worse and CPA look better. Accepting an ugly-looking account is the price of a profitable one.
2. Scaling before the economics worked.
The shape: something works for a fortnight, budget goes up sharply, and CPA rises with it rather than staying flat. The instinct is that the increase "broke" the campaign.
What to check first: whether the campaign was ever actually profitable at the small size, using the break-even ROAS above rather than raw ROAS. Usually it was marginally under and nobody had calculated the threshold. So the small version was losing a small amount and the big version is losing a large one.
The correct read: scale is a multiplier, not a fix. It multiplies whatever sign the unit economics already had. It also changes them for the worse on the way up, because the cheapest inventory is bought first. The incremental customer at double the budget always costs more than the average one did.
What actually moved it: proving profitability at a small budget against break-even ROAS, then increasing in steps small enough that you can see CPA drift before it becomes expensive.
3. Judging too early on noise.
The shape: two days in, one campaign has an excellent CPA and another has none at all, and the decision to kill one feels obvious. Or a strong first day makes a campaign look like a discovery.
What to check first: how many conversions the judgement is actually resting on. At a 1-2% conversion rate, a few hundred clicks produce single-digit conversions. And single-digit conversions cannot distinguish a good campaign from a lucky one. Both those campaigns can be the same campaign.
The correct read: decide the minimum number of conversions you'll judge on before you launch. Then let it get there. The decision rule has to exist before the data does, or the data will simply confirm whatever you already believed on day two.
What actually moved it: a written threshold. A conversion count and a minimum window at least as long as the buying cycle. And the discipline to leave the account alone until one of them is hit.
4. Ignoring the landing page.
The shape: weeks of creative iteration, a genuinely better ad, cheaper clicks, and a conversion rate that won't move. The account looks like it's improving and the revenue doesn't.
What to check first: the page, on a mid-range phone, on mobile data, from a cold start. Not on your laptop with everything cached. Then the drop-off between landing and the next step, and between the start of checkout and its end. One of those two gaps is nearly always where the money is going.
The correct read: the ad's job ends at the click. Everything after it belongs to the page. The worked example above shows why that matters so much: a 25% improvement in conversion rate does exactly what a 20% cut in click cost does, except you keep it. The ad is rented. The page is owned.
What actually moved it: matching the page to the promise in the ad, cutting the number of things asked for before the buying decision, and fixing the mobile experience. In that order, because they get progressively harder and progressively less impactful.
5. Not knowing LTV.
The shape: an endless, unresolvable argument about whether a CPA is "good", with no agreed number to compare it against. Every campaign review circles the same question and nobody can settle it.
What to check first: how many times an average customer actually buys, and over what period. Most businesses have never calculated it. Most that guess it guess high. The guess is usually the behaviour of their best customers, not their average one.
The correct read: without LTV you have no ceiling, which means both errors are available at once. Bid too cautiously and you leave volume to a competitor who did the sum. Bid past the ceiling and every additional customer makes the loss bigger. The worked example above is exactly this. A CAC that looks fine on the first purchase and an LTV:CAC that quietly says the business cannot scale yet.
What actually moved it: calculating LTV honestly on real repeat behaviour, then treating maximum affordable CAC as a hard limit rather than a target.
6. Set-and-forget.
The shape: a campaign that worked for a month starts drifting. CPM creeping up, frequency climbing, CPA following. With nothing having been changed. Because nothing was changed, it takes weeks to be noticed at all.
What to check first: frequency and CPM trend against the point where performance turned. On Meta, creative fatigue shows up as rising frequency and falling CTR on the same creative. On search, it's usually competitive. Google's auction insights report showing a new bidder, or your impression share falling while your bid didn't.
The correct read: nothing in performance marketing holds still. The environment degrades your results without you making a single mistake. Audiences saturate, creatives wear out, competitors enter and leave, and platform auctions reprice constantly. Telling those apart matters, because tired creative is a week of work and a genuinely exhausted audience is a quarter of it. Here's the order I check them in, and what to do when the channel really is finished.
What actually moved it: a fixed weekly check on a short list. CPA against target, frequency, CPM trend, conversion rate. And a creative pipeline that produces the next test before the current winner stops working, rather than after.
Step-by-step: how to start without burning cash
Know your numbers first. Work out your LTV, your target CAC, and the break-even ROAS you need. This is the homework everyone skips and everyone regrets skipping.
Pick one channel that matches your situation. Search if people already look for your solution, Meta if it's discovery or impulse. One channel, done well.
Set up tracking before you spend. If you can't measure the conversion accurately, you're flying blind. Get this right first.
Start small, with one clear offer and one audience. And a landing page you're genuinely proud of.
Give it enough data before judging. Resist the urge to react to every daily wobble.
Optimise toward profit, not clicks. Let the money metric make the decisions.
Kill losers fast, scale winners slowly. Cut what clearly doesn't work. Increase budget on winners in steps, not leaps.
Reinvest in creative and landing pages. That's where the biggest, cheapest gains almost always hide.
Checklist
Before you spend anything:
I know my LTV, target CAC and break-even ROAS.
One channel chosen, matched to how my buyers actually behave.
Conversion tracking set up and tested.
One clear offer, one audience, one strong landing page ready.
A minimum data threshold agreed before I judge results.
I'm optimising toward a profit metric, not clicks.
A plan to kill losers fast and scale winners in steps.
Resources
Platforms: Google Ads, Meta Ads Manager, and (for volume) programmatic and native networks.
Measurement: GA4 plus proper conversion tracking / attribution. Non-negotiable. Tracking fails silently more often than it fails loudly. See how affiliate tracking breaks.
The one spreadsheet that matters: your unit economics (LTV, CAC, payback). Build it before the ad account. There's a free one here.
FAQs
Performance marketing vs digital marketing: what's the difference?
Digital marketing is the whole field, including things you can't tie directly to a sale (brand, content, organic social). Performance marketing is the accountable, paid, measurable subset. It's where you can trace spend to a specific outcome.
Performance vs brand marketing: which should I do?
Both, but they do different jobs. Performance marketing gets a measurable response now and pays the bills. Brand builds demand over time and quietly lowers your CAC. Lean on performance early. Add brand as you grow. Where the two genuinely collide is when the ad that converts breaks your own brand rules — the decision rule for that is here, and it is a more common argument than the strategy version.
How much budget do I need to start?
Enough to get statistically meaningful data on one channel. Start small, but not so small you can never learn anything. The goal early on is learning, not scale.
Which channel should I start with?
Usually search if people already search for your solution (you're capturing existing intent). Meta if it's discovery or impulse and your creative can do the persuading.
Is affiliate marketing performance marketing?
Yes. Affiliate is a performance channel. You pay partners per outcome, not per impression. If you want the deep dive, read my full affiliate marketing guide.
Key takeaways
Performance marketing is paid marketing judged only on measurable outcomes.
Profit = (LTV − CAC) × Volume. Get the economics right before you scale.
Watch CAC, LTV:CAC, ROAS-vs-break-even and conversion rate. Ignore vanity metrics.
The landing page often matters more than the ad. Fix the bucket before adding water.
Kill losers fast, scale winners slowly, and never stop tending the system.
Related reading: Affiliate Marketing: the practical guide is the deep dive on one of the highest-leverage performance channels. Then SEO basics for the unpaid side of the same demand, and Google Ads vs Meta Ads for the channel decision above.
Top comments (0)