When a mid-market B2B company says its positioning is “not working,” the phrase may describe several different failures. The audience may be wrong, the problem may be too broad, the distinction may be unproven, the page may be difficult to understand, sales may be using another story, or the evidence may be too thin to choose a message. Rewriting a headline before diagnosing the branch often makes the problem harder to see.
This guide provides a practical troubleshooting path for positioning research. It moves from symptom to possible cause, evidence check, corrective action and review gate. It is not a market-size forecast, a promise of conversion, a search-ranking playbook or legal advice about claims.
1. Define the decision before diagnosing the message
Write what the company must decide: choose a primary segment, clarify a product page, resolve a sales objection, qualify a campaign audience or retire an unsupported claim. Include product version, market, buyer stage, commercial motion, owner and deadline.
If the decision is “make the brand clearer,” ask what action will change when clarity improves. A positioning diagnosis without a decision can collect interviews indefinitely while leaving the business unable to choose a page, segment or message.
2. Symptom: everyone is the audience
Possible causes: the product serves multiple roles, the team fears excluding demand, research was grouped by industry label alone, or the commercial owner has not chosen a first use case.
Evidence check: list the user, economic buyer, approver, implementation owner and disqualifier for each candidate segment. Compare the job, urgency, current workaround and proof requirement. If all segments share only a vague desire for “growth” or “efficiency,” the research is not yet specific enough.
Corrective action: choose a problem context and buying situation for the next message. Keep adjacent groups in a research backlog with a reason for deferral. A bounded audience can be widened later; a universal message rarely tells a buyer why the product matters now.
3. Symptom: interviews produce agreeable but unusable language
Possible causes: questions ask for opinions rather than recent behaviour, participants are too close to the vendor, notes remove context, or the researcher treats a compliment as proof.
Evidence check: for each observation, record what changed, what the participant did, what they tried, who was involved, what blocked a decision and what evidence would persuade an internal approver. Mark direct observation, interpretation, confidence and limitation.
The NIST Information Quality Standards provide a useful lens for utility, objectivity, integrity and correction. Use it to decide whether an interview note can support a message, needs corroboration or belongs only in an open-question register.
Corrective action: rewrite the research prompt around a concrete episode. Preserve disagreement and negative cases. Do not average contradictory evidence into a smooth persona statement.
4. Symptom: the message lists features but not a decision problem
Possible causes: product and marketing use different vocabularies, the research starts from the interface, or the team has not traced capability to workflow change.
Evidence check: write the chain capability → user action → workflow change → observable result → business interpretation. Circle the first step that relies on an assumption. “Automated routing” is a capability; “fewer manual triage decisions in a defined queue” is a testable operational result; “eliminates revenue risk” is a much stronger claim.
Corrective action: draft a problem statement in the buyer’s situation, then add the feature as the means. Ask a user, an approver and a delivery owner whether the same sentence describes their decision. If not, create role-specific proof rather than a longer feature list.
5. Symptom: competitors sound identical
Possible causes: research compares taglines instead of alternatives, the category language is genuinely crowded, or the distinction is based on an internal adjective such as “simple,” “powerful” or “flexible.”
Evidence check: map current process, spreadsheet, incumbent platform, agency, internal build and inaction. For each alternative, document the job it performs, why it persists, visible and hidden costs, switching trigger and unacceptable risk.
Corrective action: express a distinction as a situation and proof: for whom, in which workflow, with what evidence and under what boundary. Do not claim “only,” “best” or “unique” until a current comparison and reviewer support the wording.
6. Symptom: the page gets attention but conversations remain unqualified
Possible causes: the message attracts a broad curiosity audience, the next action is unclear, the page promises a different problem from the sales process, or the qualification rule is inconsistent.
Evidence check: compare page language, form questions, sales notes, disqualified reasons and the actual buyer stage. Separate traffic, comprehension, fit, permission and commercial progression. A click or form completion may be an observation; it is not automatically a qualified opportunity.
For search-facing pages, Google Search Essentials is a technical and content reference for making pages understandable to people and search systems. It does not validate a positioning claim or choose the right segment.
Corrective action: narrow the promise, add qualification context, state who the page is for and provide a next step that matches the buyer’s situation. Measure whether the audience understands the message before optimising for more visits.
7. Symptom: many near-identical pages are being proposed
Possible causes: a keyword list has become a publishing quota, the team expects location or industry swaps to create distinct demand, or no one owns canonical and overlap decisions.
Evidence check: compare the decision, audience, problem, proof, examples and expected action for each proposed page. If only the noun changes, the pages may be competing with one another and offering little additional value.
Use the Google Search spam policies page as one bounded reference when checking scaled or manipulative patterns. It does not replace editorial review or a genuine research decision.
Corrective action: consolidate overlapping topics, create one stronger page or add materially different evidence and use cases. Keep a written reason for publishing a separate page and a reviewer for the canonical choice.
8. Symptom: sales and marketing tell different stories
Possible causes: the positioning file is not versioned, product changes were not reflected, objection language is not fed back, or sales has learned a useful qualification phrase that marketing never captured.
Evidence check: sample recent calls, proposals, demos, win/loss notes and onboarding questions. Mark where the stated problem, proof, boundary and next action diverge. Do not turn one anecdote into a market conclusion; use it as a prompt for a wider check.
Corrective action: publish a message map with version, owner, permitted claims, objection responses, exclusions and evidence links. Give sales a short “when not to use this message” note. Revisit the map after product or market changes.
9. Symptom: a promising claim has weak proof
Possible causes: a target was written as a result, an illustrative example is mistaken for a typical outcome, a customer quote lacks permission, or an internal metric has no denominator or time window.
Evidence check: classify the statement as observation, customer statement, internal measure, target, example or public claim. Record population, period, method, limitations, permission, reviewer and expiry trigger.
The FTC Advertising and Marketing guidance is a U.S.-scoped prompt to keep advertising claims truthful and supportable. It is not a complete legal review for every market.
Corrective action: narrow the wording, add a qualification, obtain permission, gather more evidence or hold the claim. A quieter sentence that survives scrutiny is more useful than a dramatic sentence the team cannot defend.
10. Symptom: research insights do not change a decision
Possible causes: no owner was assigned, measures have no action rule, the review cadence is missing, or the team is reporting activity instead of uncertainty resolved.
Evidence check: for each insight, ask what decision changed, who made it, when it was made and which evidence was accepted or rejected. If the answer is “we added it to the deck,” the research loop is not operational.
The GOV.UK Measuring Success guidance is a process reference for linking measures to decisions and review owners. Use it to design a review record, not to import a universal positioning KPI.
Corrective action: set a decision date, assign an owner, state a stop rule and log unresolved questions. Retire insights that no longer affect a message, segment or research priority.
11. A practical diagnostic table
| Symptom | Root-cause branch | Evidence to collect | First correction | |—|—|—|—| | Everyone is the audience | Segment and decision undefined | Jobs, roles, disqualifiers, urgency | Choose one problem context | | Agreeable interviews | Behaviour and context missing | Recent episode, workaround, blocker | Reframe questions | | Feature-led copy | Capability-to-outcome leap | Workflow and observable result | Write a buyer problem statement | | Identical competitors | Alternative map too shallow | Inaction, incumbent and switching trigger | Specify a proof-backed difference | | Attention without fit | Message and qualification diverge | Page, form, sales and disqualifications | Qualify and align next step | | Near-duplicate pages | Publishing quota replaced research | Decision and evidence comparison | Consolidate or materially differentiate | | Unsupported claim | Evidence class confused | Denominator, permission, limitation | Narrow, prove or hold |
12. Copy-ready problem-solving record
“text Decision / audience / product version / owner / date: Observed symptom and affected surface: Root-cause branch selected and alternatives rejected: Evidence source / context / confidence / limitation: Buyer, user, approver and disqualifier: Capability-to-outcome chain and first assumption: Alternative map, including inaction: Message change / proof / boundary / permission: Measurement / action rule / stop rule / review date: Decision made / unresolved question / next research: “
Positioning research becomes useful when it can explain why a message is unclear, what evidence would change the diagnosis and which correction is safe to test. For a mid-market B2B company, that problem-solving discipline is a better foundation than endlessly refreshing slogans.
How did this article land?
Choose one reaction. You can change it anytime.
