[drop_cap]Y[/drop_cap]ou have three documents now. A company position, a competitive map, and three customer avatars. Each one is defensible on its own, and that is exactly the problem — a foundation is not three good documents, it is one set of claims that does not contradict itself anywhere a buyer can see. Most product companies have the three documents. Almost none have run the check.
This is the last stage of the Foundation band, and it does one job: it folds what you have written into a single grid, and it turns that grid into the three context blocks that every later stage of this playbook reads. Skip it and the rest of the playbook still works, but each stage will quietly invent its own version of who you sell to. By Stage 9 you will have an ad promising speed, an onboarding built for depth, a pricing page priced for scale, and a case study about a company nothing else in the funnel is aimed at.
What Misalignment Actually Looks Like
Misalignment does not announce itself. It shows up as a metric that will not move for reasons nobody can name.
The ad promises “set up in five minutes”. The signup asks for eleven fields, a company domain, and a credit card. Signup conversion is bad and the team tests button colours.
The positioning says the product is for teams of fifty and up. The free tier is generous enough for a team of three to live in forever, and the pricing page’s cheapest paid tier is priced for a solo user. Trials convert, but to the wrong plan, and the average revenue per account never rises.
The ICP says the buyer is a head of operations under pressure from a board. Every piece of onboarding content is written for a curious individual contributor who has all afternoon. Activation looks fine in aggregate and terrible among the accounts that could actually have paid you.
In all three cases nothing is wrong. Each decision was reasonable when it was made, in isolation, by someone who had read a different one of your three documents. That is the failure this stage exists to prevent: the funnel is one conversation, and it has to sound like one person is having it.
The Aligned Foundation Grid
Put your three documents into one table. Every row is a claim about the same buyer, and the test is whether you can read across a row without flinching.
| The buyer you named | The alternative they’d otherwise pick | What you promise them | |
|---|---|---|---|
| Trigger | What happens the week before they look | What they try first | The first line of your hook |
| Job | The job-to-be-done in their words | How the alternative does that job badly | The outcome your product produces |
| Fear | What they are afraid of underneath | What the alternative risks | What you say to defuse it |
| Proof | What would convince this person | Why the alternative feels safer | The proof you can actually show |
| Price | What they have authority to spend | What the alternative costs them | Your value metric and range |
Fill all fifteen cells with your own words, not with placeholders. Then read across each row out loud. A row that reads as one continuous sentence is aligned. A row where the middle column has nothing in it means you have not identified the real alternative — which almost always means the real alternative is a spreadsheet or doing nothing, and you have been comparing yourself to the wrong competitor.
The row that flinches is the one to fix, and it is usually Price or Fear. Price flinches when the buyer you named cannot approve the number you charge — a real misalignment, and the fix is either a different buyer or a different price, never better copy. Fear flinches when you have written down what your buyer says worries them rather than what actually does.
The Three Context Blocks
The grid is for you. The context blocks are for every stage and every agent that comes after, and they are the reason this stage is not optional. Write them once, in plain prose, and every later stage inherits them instead of re-deriving them.
CUSTOMER
Who they are, the trigger event that starts the search, the job they
are hiring the product to do, the words they use for their own problem,
what they are afraid of, and who else is in the decision.
DIFFERENTIATION
The alternatives they really consider, including the spreadsheet and
doing nothing. What you do that those cannot. What you deliberately
do not do. The one claim a competitor could not honestly copy.
VOICE
How you sound: the words you use, the words you refuse, the register
(peer, expert, or guide), and one sentence that is exactly right
alongside one that is nearly right but wrong.
The VOICE block is the one people skip and the one that pays off fastest, because it is the block that makes generated output usable. A tutor with a CUSTOMER block writes accurate copy. A tutor with a VOICE block writes copy you do not have to rewrite. The “nearly right but wrong” sentence does more work than any adjective list — it shows the boundary rather than describing it.
The Consistency Check
Before you leave the Foundation band, run three checks against what you have actually shipped, not against what you intend to ship.
The headline check. Read your landing page headline, then read your CUSTOMER block. Does the headline name the job in that block? If it names your category instead, Stage 1 has work to do and now you know why.
The seam check. Walk the funnel you have today — ad or search result, landing page, signup, first screen after signup, first email. Write down the promise each one makes in five words. Five entries that say roughly the same thing is an aligned funnel. Any entry that changes the subject is a seam where buyers are leaving, and it will cost you less to fix than anything you could add.
The plan check. Take the buyer from your grid, put them on your pricing page, and answer “which plan am I” as them. If you cannot answer in under ten seconds using only what is on the page, your pricing is misaligned with the buyer you just spent three lessons defining. Stage 5 fixes this properly; note it now.
None of these three checks require data. That is the point. You can run all of them in twenty minutes, before you have traffic worth measuring, and each one finds the kind of problem that no amount of traffic would have told you about.
AI Earns Its Place Here
Alignment is a comparison task, which makes it the stage where the tutor beside this lesson is most obviously better than a blank page. It has already read your company, market, and customer work. Ask it to fill the fifteen cells of the grid from what you wrote, and then — this is the useful part — ask it to name the row that contradicts itself. A person who wrote all three documents cannot see the contradiction, because each claim arrived with a reason attached. Something reading them cold can.
Then ask it to draft the three context blocks and push back on them until the VOICE block includes a sentence that is nearly right but wrong. That single line is the highest-leverage thing you will write in the Foundation band; it is also the line a model will not produce unless you ask for it specifically.
Save the grid and the three blocks to your playbook when they are done. Every stage from Hook to Refine reads them, which means the twenty minutes you spend here is the difference between ten stages that compound and ten stages that each start over.
You have finished the Foundation band. Next is Stage 1: Hook — a first message that names the job, not the category. Everything you just wrote is what makes that possible.