Buy the definition before the month: the order the money should be spent in

Buy the definition before the month

A retainer bought before anyone has defined a lead is a monthly invoice funding the wrong number. That is not a dig at agencies — it is arithmetic, and the arithmetic is the same for a one-person operation like mine.

I sell three things, with public prices: a $900 Sprint (one-off), a $1,900/month Engine, and a $2,900 Full Build. The order they should be bought in is not the order they are usually bought in. Here is the case for buying the cheapest one first.

The month a retainer actually buys

Strip the channel names and every retainer is bought to produce four things:

1. A written definition of a lead. Not "inbound" and not "MQL" — one sentence with a trigger in it, signed by whoever will argue about it next quarter.
2. One owned number with its volume floor. A retainer that owns five numbers owns none. Below the floor, the month reports effort, not a result.
3. A stop rule per live test, written before the first contact of that test. A stop rule written after the fact is a justification, not a rule.
4. One artifact that still exists when the month ends — page, tool, dataset, read-out. If nothing shipped, the month should be free.

Everything else — which CRM, which sending stack, which sequencing tool — is downstream of those four. This is the part buyers skip, because tooling decisions feel like progress and definition work feels like paperwork. It is the reverse: the definition is the only part that changes what a hundred hours of work is worth.

The volume ceiling that no retainer can argue with

One strict two-variant test, 3% base rate, a +20% relative lift, 80% power, alpha 0.05, needs 13,914 contacts per arm — 27,828 in total. At 500 contacts a day that is 56 days. At 1,000 a day, 28 days.

So a retainer sitting on a few hundred contacts a day can finish at most one honest test per month. When a monthly report claims three significant wins at that volume, read it twice. The number of tests a month can contain is set by volume, not by how many channels someone is willing to run at once.

That single paragraph is the difference between a retainer that produces decisions and a retainer that produces activity. It is also why the $900 Sprint exists: writing the definition, wiring the measurement, and setting the stop rule are one-off tasks. Paying a monthly fee for them is paying rent on work that should have been finished before the first invoice.

Receipts from my own domain, including the boring ones

I run this on myself, so the numbers are checkable rather than asserted.

• Publishing: 76 articles on my dev.to account, 42 of them carrying a canonical tag back to my own domain, so the indexed original stays mine. Every page that sites an article gets a visible link to it, plus a line in a machine-readable index.
• Distribution is measured in server logs, not vibes: today's access log shows 3,139 requests and 109 hits carrying a UTM tag — 81 of them from my own email outreach, 9 from dev.to links, the rest from chat platforms. Analytics recorded 3 sessions and zero key events over the same window, because crawlers and most chat-preview fetchers do not run JavaScript. The honest reading: publishing is necessary, not sufficient, and 9 clicks from 55 articles in a day is the real size of the first step.
• The list: 4,988 verified company domains, each row required to have come with its own domain in the source post — no inferred addresses, no guessed mailboxes.
• The work I would rather not print: zero UTM-campaign sessions from content so far, one reaction and two comments across fifteen service articles. A month that reported only the top line would be a marketing month. A month that reports both lines is one you can actually renew or cancel.

Four decisions the retainer cannot make for you

• Which single number is owned this month, and the volume under it. Five numbers means zero accountability.
• What the retainer is allowed to switch off, and on what condition — no readable return in 48–72 hours, or a cost past a stated multiple.
• What is out of scope: media spend, subscriptions, data cost. Writing them out is what makes the invoice comparable with a freelancer's hour or an agency's deck.
• How the engagement ends. The renewal test is a list of artifacts, not a feeling: what exists this month that did not exist last month, what the numbers behind it were, and which channel was switched off and why.

A retainer that never reports a channel running below its floor is either very lucky or not measuring. I have both kinds of month printed on my own site.

The order

1. Define first ($900, one-off): definition, dashboard, stop rule, one channel live in five working days. If this produces nothing you would show a colleague, the monthly version was never going to.
2. Then the month ($1,900): one channel end to end, with a written record and an owned number each month. One honest test per month is the ceiling at moderate volume — plan for that instead of for three.
3. Then the whole loop ($2,900, one-time) if the pipeline the retainer reports on does not exist yet.

No media spend is included or marked up, the accounts stay yours, and the Engine cancels with 14 days' notice. Scope is written out line by line; the artifacts from running this on my own domain are on the proof page.

If you want the definition work before the retainer, that is exactly the order I would sell it in — start here.