Audit mode — critiquing existing positioning

Use when the user has an existing messaging framework, homepage, deck, or set of collateral and wants it reviewed, updated, or rewritten. Run this before Stage 0, because inherited language is the main source of drift and you need to know what you are inheriting.

Contents

• Why audit first
• The inventory
• The seven checks
• Presenting findings
• Deciding: patch or rebuild

Why audit first

A rewrite that doesn't name what went wrong reproduces the same mistakes with fresher adjectives. The audit also produces something politically useful: evidence. "Three of your decks claim X and your newest one says don't claim X" is an argument that moves a leadership team, where "I think the messaging is unclear" is not.

The inventory

Read everything before judging anything. Gather:

• Homepage and key product pages
• Every pitch deck in circulation, with last-modified dates
• Any messaging framework, positioning doc, or brand guide
• Recent sales collateral and one-pagers
• Customer-facing security or compliance summaries
• Recent customer calls or demo recordings, if available

Record the date of each. Vintage explains most contradictions. A deck from eighteen months ago isn't wrong, it's stale — a different problem with a different fix.

The seven checks

1. Hero count

How many different things does the product get called across the artifacts? List every one with its source and date.

More than one primary identity is the most common finding and the most damaging. Every downstream inconsistency usually traces back to it. Present the list as a table; the length of it does the arguing.

2. Category count

Same exercise for the category. Watch specifically for coined platform names. A coined category in an artifact is a signal the company attempted category creation at some point — check whether that decision was ever made deliberately or whether it drifted in from an investor deck.

3. Table stakes

Take the top three to five differentiators the materials claim. Search competitors' current marketing live. Mark each: still differentiating / now table stakes / never was.

This is the most valuable single output of an audit. Claims that were sharp eighteen months ago frequently aren't, especially where a large incumbent has bundled the capability at no additional cost. Include screenshots or direct quotes from competitor pages — the specificity is what makes the finding land.

4. Claim consistency and risk

Look for the same claim stated at different strengths across artifacts. Pay particular attention to:

• Security and privacy claims. Absolute statements ("never," "prohibited," "certified") in older collateral, versus carefully qualified language in newer material. This is a legal exposure issue, not just a consistency one, and it should be flagged as the highest-priority fix.
• Quantified outcomes. The same metric attributed to different subjects — "per engineer" in one place, "per PM" in another.
• Credentials. Years of experience, team background, certifications. These drift upward over time.
• Customer attribution. Named accounts in some materials, anonymized roles in others. Check whether a quote-clearance policy exists.

5. Sourcing

Every statistic: does it have a source and a date? Are internal estimates labeled as estimates?

Unsourced numbers are a liability in two directions — a buyer who checks and finds nothing, and a salesperson who can't defend the figure in the room. Where a number is unsourced, look for a credible published substitute rather than just flagging the gap.

6. Vocabulary

Build the two-column list: terms customers use versus terms only the company uses. Source the first column from real customer material — calls, transcripts, support tickets, reviews. If no customer says it, it is house language.

House language isn't worthless. It often does real work internally. The recommendation is usually "keep for internal alignment, retire from customer-facing surfaces," not "delete."

7. Specificity

Run the competitor-swap test on the homepage headline and the deck's opening slide. If a competitor's logo sits comfortably on top of the copy, the copy describes the category.

Presenting findings

Structure the audit as:

1. What's in the inventory — with dates, so vintage is visible
2. The core finding — usually one structural problem, stated in a sentence
3. Inconsistencies — evidenced, in tables, with sources
4. Risk items — anything legally or factually exposed, called out separately and first in priority
5. What's working — the strongest existing artifact, and why. There is almost always one asset that's better than the rest, and naming it gives the rewrite a template and the team a win.
6. The decisions that have to be made — the two or three forks that determine everything downstream

Be specific about which artifact says what. "Nine decks carry the absolute version of this claim" is actionable; "the messaging is inconsistent" is not.

Do not soften the structural finding. A team commissioning an audit wants the diagnosis, and a hedged one wastes the engagement. Deliver it plainly, with evidence attached, and without moralizing about how it happened — drift is normal and usually the product of a company moving fast.

Deciding: patch or rebuild

Rebuild when:

• There are multiple heroes or multiple categories
• Core differentiators have become table stakes
• The ICP has changed or is about to
• A strategic decision has been made that the framework predates
• The framework was built narrative-first

Patch when:

• The decision is sound and the expression is weak
• Only specific claims are stale or legally exposed
• The problem is distribution and adoption, not content

Say which you recommend and why. If a strategic decision has already been taken that the existing framework contradicts — a new category choice, a new segment, a decision about what the product is called — surface it immediately and prominently. It is almost always the real finding, and everything else is downstream of it.