Reviews
What the product positioning best practices look like in practice
Positioning best practices are routines a team can run: evidence files, message tests, claim checks and a review cadence that retires what stops working.
What to take away
Product positioning best practices look like a set of files, tests, and review dates, not a slide.
- Write the product positioning statement as a claim with evidence attached: buyer, alternative, defensible difference. Use Geoffrey Moore's Crossing the Chasm template or April Dunford's competitive alternatives exercise.
- Rank claims by proof. A measured pilot result is strong. A customer's renewal reason is strong. An analyst category label is medium. A founder's market belief is weak until tested.
- Run win/loss interviews on a fixed six-question script within two weeks of a closed deal, before memory rewrites the reason.
- Keep a competitive intelligence file with a dated source for every entry. Delete entries you cannot source. Klue and Crayon are paid platforms that collect rival updates.
- Test message comprehension by segment with Wynter, UserTesting or Lyssna before launch day, not after sales reports confusion.
- Check review and testimonial practices against the FTC's five integrity points before naming any provider.
Write the claim, then the proof
A product positioning statement is one sentence naming the buyer, the alternative they would otherwise choose, and the difference you can defend. Write the evidence beside it in the same file.
Claim strength ranking
Strong
- Source
- Own pilot
- Example
- Measured result
- Action
- Launch message
Medium
- Source
- Analyst label
- Example
- Category label
- Action
- Use with care
Weak
- Source
- Founder belief
- Example
- Market opinion
- Action
- Research queue
Geoffrey Moore's template from Crossing the Chasm gives a working shape: For [target buyer] who [need], [product] is a [category] that [benefit]. Unlike [alternative], we [difference]. April Dunford's Obviously Awesome starts with competitive alternatives, then lists unique attributes, value, best-fit customers, and market category. Both methods end with a claim and a proof list.
Rank claims by strength. A measured result from your own pilot is strong. A customer's stated reason for renewing is strong. An analyst's category label is medium. A founder's belief about the market is weak until tested.
Weak claims do not go in the launch message. They go in the research queue.
Public examples show the shape. Stripe describes itself as financial infrastructure for the internet. Shopify describes itself as a commerce platform for entrepreneurs. Notion describes itself as an all-in-one workspace; its free plan covers individual use.
Miro describes itself as a visual workspace for distributed teams; its free plan includes three editable boards. Slack's early pitch focused on replacing email for teams; its free plan now keeps 90 days of message history.
Run the routines that produce the evidence
Positioning drifts when nobody owns the inputs. Four routines keep it current.
- Win/loss interviews.Script six questions, run them within two weeks of a closed deal, and record answers verbatim. Ask who else was considered and what tipped the decision. Gong and Chorus are conversation intelligence tools that can store call snippets; both use quote-based pricing.
- Message comprehension tests.Read the headline and the first paragraph to ten people in the target segment. Ask them to repeat the promise back. Track the share who get it right. Wynter tests messages with business respondents. UserTesting records video feedback from targeted participants. Lyssna offers five-second and first-click tests.
- Competitive intelligence reviews.Every entry carries a source and a date. Pricing pages, release notes, job postings and conference talks are checkable. Rumor is not. Klue and Crayon collect competitor updates and send alerts; both quote-based.
- Sales enablement inventory.List the assets the field actually uses, by deal stage. Assets nobody opens get cut at the quarterly review. Highspot and Seismic report asset usage; HubSpot's CRM has a free tier for basic tracking.
The product positioning work sits underneath all four. Each routine feeds the same claim-and-evidence file rather than a separate deck.
Set the cadence
| Cycle | What gets reviewed | Output |
|---|---|---|
| Weekly | Deals lost and the stated reason | Repair or a new interview |
| Monthly | Message comprehension by segment | Keep, rewrite or retire the line |
| Quarterly | Claims, evidence, enablement assets | Retest or cut |
| Event driven | A competitor claim that undercuts yours | Pause, recheck evidence, respond or concede |
Event-driven review is the one teams skip. A rival's launch, a regulator's action or a major customer's public switch all invalidate assumptions faster than a quarterly cycle catches. Klue or Crayon alerts can flag the event, but someone still has to judge the evidence.
Check the evidence you cite
Before a claim reaches a slide, confirm the source exists, is current, and says what you think it says. The GAO data reliability guide ties data quality to intended use and asks for a documented assessment, which is the right standard for a positioning file.
Record the source date and its limits next to the claim. A benchmark from a different segment is a limit, not a proof point. A Gartner or Forrester category report is a secondary source: check its definition, survey date, and sample before quoting it.
Keep reviews and testimonials clean
The FTC guidance for marketers using reviews names five things to avoid: fake feedback, selective requests that skew the sample, incentives conditioned on a positive review, hidden relationships between the reviewer and the seller, and paid ranking presented as organic.
Apply the same five to case studies and logo walls. A customer quote gathered under an incentive, or a ranking bought rather than earned, is a claim that will not survive a buyer's diligence. G2 and Capterra badges can be paid placements, so disclose sponsorship before using one.
Make the page match the claim
If the positioning promises a workflow, the page has to show that workflow. The W3C guidance on information and relationships asks that visual structure also be available programmatically, so a comparison table reads as a table to assistive technology. Use a table caption and header cells, not color alone.
Keep the review trail too. The CISA logging guidance for business systems explains why organizations retain and review event records. Keep access, change and correction records for the positioning file, without treating a log as proof of a market result.
Common questions
What is the first decision in a positioning review?
Name the owner and the buyer, then write the claim and the evidence that would force a rewrite. April Dunford's process starts with competitive alternatives, not your product. Without a stop condition, every review ends in agreement and nothing changes.
How often should positioning be revisited?
Monthly for message comprehension, quarterly for claims and enablement assets, and immediately when a competitor or regulator moves. A yearly offsite is too slow for a market that shifts between them.
What kills a positioning program fastest?
Claims that outrun their proof. Once sales stops trusting the message because a buyer challenged it and nobody could answer, the field writes its own version and the launch message is gone.
Where does competitive intelligence go wrong?
When entries lose their sources. A file of undated assertions about rivals looks like research and behaves like rumor, and it will steer a launch toward a competitor who never existed.
