Blueprint & Knowledge: test the bets your engine runs on

The question

We wrote down who we sell to and how we win months ago. Is any of it still true?

Every revenue team runs on a handful of decisions: who buys, what you say to them, how you beat a competitor and how you decide a deal is real. Most teams write them once, then stop checking. The documents stay confident long after the evidence has moved.

Where it sits on the loop

Each decision feeds a different part of the loop. Your ideal customer profile (ICP) decides who enters at Imported, and shows up again in how many reach Workingand Engaged. Your positioning is what turns Engaged into Replied. Your competitor answers and the way you qualify deals decide how Pipeline turns into Won.

That makes them likely causes when a stage struggles. They belong in a quarterly review, and in any week when the first constraint lands on a stage they feed.

How to think about it

Treat each decision as a bet with a test. A bet you can’t test is an opinion; a bet with a test and a review date is something the team can learn from.

• Four bets. Who buys (the ICP). What you say (positioning). How you win against competitors (battlecards). How you qualify (your framework). Write each one down where the team can see and use it.
• Each bet gets one test. The ICP should separate buyers: leads that fit it should convert clearly better than leads that don’t. Positioning should earn replies. A battlecard should keep your win rate against that competitor close to your overall rate. Deals that answer your qualification questions should win more often than deals that don’t.
• Wait for enough evidence. A test on five leads proves nothing. Decide the minimum before you look, and say “not enough data yet” until you reach it.
• Revise with a reason, then measure. When a bet fails its test, change it, write down why, and watch the same test for a fixed window afterwards. What changes after a revision is consistent with the revision; it isn’t proof the revision caused it.
• Bets are causes, not the constraint. A stale ICP explains why a stage might be failing. The stage is still the thing you measure.

Common mistakes

• Writing the ICP once and filing it. The market moves, the team learns, and the document doesn’t. Leads keep being scored against last year’s guess.
• Judging a bet on a handful of deals. Two losses to a competitor trigger a new battlecard; two wins declare it fixed. Neither is evidence yet.
• Editing a bet without saying why. Months later nobody knows what changed, so nobody can tell whether the change helped.

In Summit53

You write your bets in Knowledge and check them in the Blueprint. The Blueprint shows each bet, the evidence for it, and the checks that flag when a bet looks out of date.

Getting there

• Blueprint: on the Revenue Engine, switch the masthead toggle from Engine to Blueprint (or add ?view=blueprint to the address). The masthead stays; everything below it becomes your Blueprint.
• Knowledge: Knowledge in the main navigation, at /knowledge.

The Blueprint opens with The bets your engine is built on, and whether the evidence backs them. Revise a bet and we’ll track what changes after., then shows, in order: the ICP card and ICP checks, the positioning card, the battlecards card and Battlecard checks, the methodology card, and Other checks.

The four bet cards

Every card has an eyebrow naming the bet, the asset’s title and version (for example v2), a status and, where there is one, a review badge:

• Status: On track, Off track, Building evidence (not enough data to judge yet) or Not set up (nothing published).
• Review: Review by <date>, Review due (within 14 days), Overdue, or Reviewed <date>. Publishing an ICP, positioning or battlecard sets its review date 90 days ahead, and Mark reviewed resets it.

A one-sentence verdict follows, then the evidence:

• ICP · who buys. Leads by fit (fit tier, leads, converted, conversion) and Accounts by fit (accounts, deals, win rate), then How you’ll know it’s working with its test, by default Strong-fit leads convert at least 2× weaker-fit leads, marked Met, Not met, Building evidence or Not measured yet. It needs 20 strong-fit leads with 5 conversions, and 20 weaker-fit leads, before it compares. If leads were scored on an older ICP, a line says so with a Rescorebutton.
• Positioning · what you say. The share of engaged leads who reply, against a benchmark, over the last 90 days, broken down by sequence version and by positioning version. It needs 20 engaged leads, and 20 contacted per version. Only replies that can be tied to one sequence count, and the card says how many couldn’t be.
• Battlecards · how you win against competitors. Your overall win rate on closed deals with a competitor recorded, then Deals by competitor, last 365 days: closed deals, win rate, whether a battlecard covers it, closed deals with and without the card, and open deals. Lost-deal reasons sit under each competitor. A competitor needs 5 closed deals before it gets a win rate. The card is off track when you win 15 or more points below your overall rate against a competitor you have a battlecard for.
• Methodology · how you qualify. Closed deals by score, last 365 days, split into deals scored complete, incomplete, and with no score before close. A deal counts as complete when it scores at least 50% on <your framework> before it closes. The test is that complete deals win at least 1.5× as often as incomplete ones; each group needs 5 closed deals.

A bet with nothing published shows Not set up, a short explanation and a Set up … button (for example Set up your ICP) that opens a new asset in Knowledge. The ICP, positioning and methodology cards end with Review(open the asset), Mark reviewed and Revise; the battlecards card ends with Add a battlecard.

Revise a bet

1. Click Revise on the card. For battlecards, use Revise on the competitor’s row.
2. The dialog (Revise the bet) explains what happens: We’ll open a draft copy to edit. Your current <bet> stays live until you publish the draft, then we track …
3. Optionally fill in What do you want to change? (optional) (up to 500 characters), then click Start revision.
4. Summit53 opens the draft in Knowledge. Edit it and publish it when you’re ready. It replaces the live version as the next version number, and the old one is archived.
5. The card now shows Revision being measured, with the publish date, the end of the window and the starting point.

What each revision is measured on, and for how long after you publish:

• ICP: whether strong-fit leads convert better, over 45 days.
• Positioning: whether more of the leads you contact reply, over 45 days.
• Battlecard: whether you win more deals against the competitors it covers, over 90 days.
• Methodology: whether deals that score complete win more often than incomplete ones, over 90 days.

While the draft is unpublished, the card shows Revision in progress with Continue editing. When the window ends, the result appears in Did our advice work?.

Blueprint checks

Below the cards, checks flag a bet that looks out of date. Each one is firing(with a severity of High, Medium or Low, the evidence and a link to fix it), still building (named after the check, with Not enough data yet: <have> of <need> …) or passing (collapsed into N checks passing).

Where a check matches a Fix playbook, its row has Fix this: <playbook>, which opens the playbook over the Blueprint. See Fix playbooks. Positioning and methodology have no checks yet.

Checks are also listed as LIKELY CAUSES in the Stage Inspector: ICP checks on Imported (and on Working or Engaged when that stage is the first constraint), battlecard checks on Pipeline, and anchor customers on Renewed. They never decide the first constraint; see Diagnostics.

Knowledge

Knowledge holds your bets and the rest of what Summit53 draws on: Your ideal customer profile (ICP), positioning, competitor battlecards and sales methodology, in one place. Summit53 uses what you publish here to score leads, draft outreach and prepare deals, so keep it current.

• Types: Company Positioning, ICP Definition, Product Information, Pricing & Packaging, Security & Compliance Posture, Brand Voice & Messaging, Sales Methodology Definition, Competitive Battlecard, Case Study / Win Story, Team play, Reusable Snippet and Custom.
• Team play is a short, repeatable response to a specific situation, such as a deal going quiet or a new decision maker appearing: a clear trigger, three to five steps, and what success looks like. Your team can find it through Ask Summit53.
• Statuses: Draft, In review, Published and Archived. Only one ICP, one positioning and one brand voice can be published at a time.
• + New asset opens a guided form with Guide: <type> (what it is, why it matters, what good looks like) and, for most types, Insert starter template. The ICP has structured fields (industries, company size, locations, buying triggers, and who is not a fit) that must be complete before it can be published. A methodology asks you to pick the framework your team qualifies with (MEDDPICC, BANT or SPICED).
• Published bets change through a revision. On a published ICP, positioning, battlecard or methodology, Edit (opens a draft revision) opens a draft and leaves the live version in place until you publish. A revision started this way isn’t measured; use Revise on the Blueprint when you want the change tracked.

Who can do what

• Everyone in the workspace can open the Blueprint and Knowledge, and can create, edit, publish and archive assets, mark bets reviewed and start a rescore.
• Revise (a measured revision) is for workspace owners and admins, and for members with a leader role. Others don’t see the button.

Review your Blueprint

1. On the Revenue Engine, switch to Blueprint.
2. Read each card’s status and verdict. Skip any that say Building evidence: there isn’t enough data to judge them yet.
3. Work through the firing checks, starting with High. Use each row’s link, or Fix this where it appears.
4. For a bet that’s off track, click Revise, say what you want to change, edit the draft and publish it.
5. For a bet that still holds, click Mark reviewed so its review date moves forward.

When the data is thin

• Building evidence, not a verdict. Each test waits for its minimum sample: 20 strong-fit leads with 5 conversions and 20 weaker-fit leads for the ICP, 20 engaged leads for positioning, 5 closed deals per competitor, and 5 closed deals in each methodology group. Below that, the card says Not enough data yet and how far it has got.
• Partial coverage is stated. Cards say what they couldn’t count, for example Competitors are recorded on 31% of closed deals, so this only covers those. or how many replies couldn’t be tied to a sequence.
• Unknown versions. Sequence versions published before your positioning was tracked show as unknown.
• After, not because. A revision’s result shows what changed after you published it. It doesn’t prove the revision made the difference.
• As of when. The Blueprint is read when you open it and after each action you take on it. Checks listed in the Stage Inspector come from the daily signal snapshot, so they can trail the Blueprint by a day.
• What changes. Revise opens a draft and records the revision; nothing goes live until you publish. Rescore rescores leads and accounts against the current ICP.

Do it this week

1. Write each of your four bets in one or two lines: who buys, what you say, how you beat your most common competitor, and how you qualify.
2. Next to each, write the test that would tell you it’s wrong, and the minimum number of leads or deals you need before you’ll judge it.
3. Pull the numbers you have. Mark each bet holding, failing, or not enough data yet.
4. For one failing bet, write the change and the reason, and the date you’ll check the same test again.
5. Put a review date on every bet, 90 days out.

Troubleshooting

• A card says Not set up. Nothing of that type is published. Use its Set up … button, then publish the asset in Knowledge.
• There’s no Revise button. Revise is for owners, admins and leaders. You can still change a published bet with Edit (opens a draft revision) in Knowledge, but that change isn’t measured.
• I can’t publish my ICP. Its structured fields aren’t complete. Open it in the editor and finish the fields listed under Before you can publish.
• The ICP card says leads were scored on an older ICP. Click Rescore. It can take a few minutes.
• A competitor shows Not enough data yet. It has fewer than 5 closed deals in the last 365 days. Make sure reps record the competitor on deals.
• My revision isn’t being measured. The window starts when you publish the draft. Until then the card shows Revision in progress.
• A check fired but the canvas doesn’t show it yet. The canvas updates from the daily snapshot. Check again tomorrow.

Related

• Diagnostics: find your first constraint
• Fix playbooks
• Revenue flow canvas: likely causes in the Stage Inspector
• Signals & the Loop: revision results
• Build your engine: publishing your first bets