A 30-day engagement, priced to be a single-call decision rather than a procurement exercise. The first three stores pay $1,500 in exchange for permission to publish the results. After that it’s $3,000.
For reference on the market: an annotated teardown with no implementation runs $250–$1,499 and stops at the document. A full CRO programme with implementation starts around $7,500 a month and generally wants a $10M brand and a 12-month term.
One-time. 30-day engagement.
The founding rate buys one thing from you: written permission to publish the reconciled results. That’s the actual trade, and I’d rather say so than invent a fake deadline.
Because the refund clause has already removed your financial risk, dropping the price further doesn’t buy either of us anything.
It’s high enough to signal that this is implementation work rather than a document, and low enough to be a decision you make on one call without assembling a committee. Below it, the only thing that changes is that I start competing on price with known brands who sell you a PDF — which is a fight worth losing.
There are three founding slots because I do the implementation myself, and three concurrent 30-day builds is what one person can ship properly. When they’re taken, the price is $3,000 and the guarantee is unchanged.
There is no upsell built into the front door. At the reconciliation call you’ll have a measured number and a ledger with lines still open on it. Then it’s a real choice: work through the rest yourself, or keep me running the ledger.
Which tier you belong in is decided by your traffic, not by what I’d prefer to sell. Above roughly 140,000 sessions a month, tests can reach significance and a programme is worth running. Below it, they can’t — and I’ll tell you that instead of charging you $7,500 a month to run tests that will never conclude.
For $1M–$3M brands, where A/B testing isn’t statistically viable yet.
For $3M+ brands, once the known debt is paid down and you’re genuinely at the frontier.
Debt accrues. Every product launch, every theme change and every app install adds new lines to the ledger — which is why the recurring model is honest here rather than invented. It is still month to month.
Not selectivity theatre. Under this floor, 30 days of revenue-per-session data is too noisy to stand behind, and a guarantee I can’t measure is worth nothing to you.
Paid traffic matters because it’s a stable, predictable source — it’s what makes a before-and-after comparison mean something. If you’re under the floor, say so in the application anyway; I’ll tell you honestly whether it’s worth doing without the guarantee attached.
Find it, price it, fix it in 30 days. If revenue per session on the pages I touched hasn’t gone up, you get every dollar back.