Make every deal
prove it is real.

An ideal customer profile is the set of conditions under which you reliably win — not a description of the companies you would like to sell to. Size and industry are what Meridian screens on today, and both are equally true of the deals it loses.

• The winning condition is a failure, not a firmographic. Five of the last eight wins had a warehouse system the vendor had stopped supporting. That is a date on a calendar, and it is why those deals closed in eleven weeks against a five-month average.
• Disqualification is the output nobody writes down. Below $40M in revenue the deal clears procurement but never clears the price floor. Meridian worked nine of them last year and closed one.

The prompt refuses to build a profile from open deals — only closed outcomes count — and it states the sample size it had to work with.

Every line is graded Sourced, Inference or Unverified, and Sourced means named and dated. A brief that is mostly Inference has not been researched, and the grading makes that impossible to hide behind formatting.

• Two facts, one date. Brightwater’s warehouse system went out of support in March and a new VP Supply Chain arrived the same month. Separately each is a detail. Together they are a project with a budget.
• The job ads are evidence. Brightwater is hiring two inventory analysts to support multi-site reporting — which is the work the software does, priced at two salaries.
• The buying group is mapped by what each person is measured on. The VP is measured on service levels, the Director of Ops on headcount, IT security on nothing going wrong. Three different arguments.

Had no trigger been present, the brief would have said so and recommended leaving the account alone until one appeared. No invented quotes, no “I saw your post” unless the post was pasted in.

The delete test is the method: name the exact sentence that would make a busy person delete the email, then rewrite that sentence and show both versions.

• The angle is a verifiable fact, not a compliment. “Your warehouse support ended in March and peak season starts in November” can be checked. Fake personalisation is banned outright, along with invented mutual connections.
• No touch repeats the previous ask. Meridian’s old sequence asked for the same thirty-minute call five times in different words. This one moves from a question, to a peer example, to a specific dated proposal.
• Unprovable claims are flagged, not softened. The line about cutting late shipments carried [NEEDS PROOF] until a named customer agreed to be referenced.

If the only honest hook is weak, the prompt says the account is not ready and names the trigger to wait for — which is a cheaper outcome than five touches nobody answers.

Cost of inaction is what the current situation costs per year if nothing changes, and it only works when the buyer produces the number. A figure the seller supplies is a figure procurement argues with.

• The arithmetic is theirs. Eleven late shipments a month at $4,200 in contractual penalties is $554K a year — Brightwater’s contract, Brightwater’s count. Meridian only asked the questions that produced it.
• Three situation questions, maximum. Anything answerable from the annual report is not asked. The time goes to impact and consequence instead.
• The last-purchase question predicts more than any other. Brightwater’s previous rollout slipped two quarters on security review — which is why the plan in step 10 starts there rather than ending there.

The prompt also lists what not to ask: the leading questions and the pain-prodding that reads as manipulation to a senior buyer.

An objection is a real constraint, a missing piece of information, or a proxy for something the buyer has not said — and the response depends entirely on which. Most sales training skips that step and goes straight to a rebuttal, which is why most rebuttals make the deal worse.

• The test question separates them. “If the price were half, who would sign?” A name means it is price. A pause means no budget owner has been established, which was the case here.
• Conceding the true part buys the next sentence. Year one genuinely costs more than the upgrade Brightwater just paid for. Saying so plainly is what makes the $554K comparison credible.
• Prevention is the part that improves the next ten deals. This objection appears whenever the economic buyer is met after the proposal rather than before it.

No rebuttal scripts and no manufactured scarcity. Where the objection is a genuine disqualifier, the prompt recommends withdrawing and names what would have to change.

A battlecard is used live, thirty seconds before a call, by someone who is nervous — so the moment reality contradicts it, the rep stops trusting the whole card.

• Honesty is the feature. Tallgrass deploys faster and integrates natively with the ERP Brightwater runs. A rep who knows that walks away from the wrong deals and spends the week on the right ones.
• The questions for the buyer have to be fair. “Show me a month-end close across six sites” comes from the complaint pattern in Tallgrass’s own one-star reviews. A gotcha gets rebutted and costs credibility.
• Price is compared like for like. Tallgrass’s quote excludes $31K of integration work Brightwater would pay in month one. Every figure is labelled Sourced or Unverified.

Nothing goes on the card that Meridian could not show the buyer the source for. Claims without proof are listed separately as [NEEDS PROOF] rather than quietly used.

A conservative case models the result when things go worse than planned, and it is the only case a finance reader trusts. Meridian’s old proposals led with the best case and were read as marketing.

• Every number traces to something Brightwater said. The $554K came out of their own contract during discovery. Meridian’s assumptions are labelled as assumptions, for the buyer to confirm or correct.
• The exclusions section prevents the month-three fight. What is not included is stated in the document the buyer signs, not discovered after the kickoff.
• One recommendation, two alternatives. A menu of five options is a decision handed back to the buyer, and it is why proposals sit unanswered for three weeks.

If the business case does not hold at conservative assumptions, the prompt says so before the proposal is sent rather than after it is rejected.

Follow-up either carries information the buyer did not have, or it trains them to stop opening your email. Losing to a competitor, losing to no decision and losing to a reorganisation need three different responses, so the diagnosis comes before the writing.

• The causes are ranked with evidence, not assumed. The rep assumed price. The proposal was opened four times in one afternoon by three different people, which points at an internal priority fight rather than sticker shock.
• Multithreading is done in the open. The message to the VP goes out alongside a line to Dana telling her it is happening. Doing it quietly is how a champion becomes a blocker.
• The honest direct message recovers more deals than the four before it. It asks plainly whether this is still real and makes “no” the easiest possible reply.

No invented deadlines and no discount expiring on a date Meridian would ignore. Where the read is that the deal is dead, the recommendation is to close it and record why.

Pipeline coverage is open pipeline divided by target, and it is the most comfortable number in sales because it is the easiest to inflate. The inspection asks one question of every deal — what evidence exists outside the rep’s optimism — and re-cuts the total on the answers.

• Two of the four largest deals rest on one contact who started in March. That is not a risk note. It is $1.1M that disappears with one role change, and it is in the commit column.
• Twenty-two of thirty-four deals have no next step with a date agreed by both sides. That is the same 58% that ends in no decision, visible three months before it happens.
• The gap gets three named actions, not “increase activity”. Three deals, three people, this week — chosen because together they close the largest share of the shortfall.

The prompt does not invent probabilities to make the arithmetic work, and it names the fields missing from the export that limited the analysis.

A mutual action plan is one schedule both sides agree to, worked backwards from the buyer’s own date, with an owner against every step including theirs.

• Brightwater’s last rollout slipped two quarters on security review. That is why the questionnaire sits in week one of the plan rather than turning up in week six as a surprise.
• The confirmation message exists to get corrected. Sending the plan and asking the buyer to fix the dates is what converts Meridian’s assumption into Brightwater’s commitment.
• Price is the last thing traded. Term length, start date, scope and a reference all move first, and the walk-away is decided before the call rather than during it.

No manufactured deadlines. If Brightwater needs more time for a legitimate reason, the plan reworks the dates around it and states what the delay costs them.