Positioning Research: How to Validate Your Product Positioning with Customer Interviews

The answer up front: Positioning research validates the strategic frame around your product — competitive alternatives, unique attributes, value, target market, and market category — using structured customer interviews with wins, losses, and never-evaluated prospects. It is fundamentally different from messaging testing (which evaluates specific copy) and brand research (which measures perception). The fastest, most evidence-rich way to run positioning research today is a hybrid program: 20–40 AI-moderated interviews using ranking and single-choice questions to quantify the competitive set and value attributes, paired with open-ended probes that surface the verbatim language buyers use. Teams using AI interview platforms like Koji ship validated positioning in 2–3 weeks instead of the traditional 3–4 months.

---

Why positioning is the highest-leverage research you're not doing

CB Insights' analysis of 483 failed startup post-mortems found that "no market need" was the #1 reason startups fail — cited in 35% of cases, ahead of running out of cash. "No market need" is almost always misdiagnosed product-market fit; the real failure is usually positioning. The product could serve a real job, but it's framed in a category buyers don't recognize, compared to alternatives buyers don't consider, or sold on benefits buyers don't value.

The same pattern appears mid-funnel:

• 40–60% of qualified B2B deals end in "no decision," and Matthew Dixon and Ted McKenna's analysis of 2.5M+ sales conversations found 56% of those losses are indecision-driven — confusion, complexity, fear of choosing wrong — not status-quo preference (HBR, 2022).
• 77% of B2B buyers describe their most recent purchase as very complex or difficult (Gartner B2B Buying Journey).
• Forrester found 86% of B2B purchases stall mid-cycle and 81% of buyers are dissatisfied with the provider they choose (Forrester, 2023).
• And on the upside: Matt Lerner (ex-PayPal, ex-500 Startups) documented signup conversion lifts from 0.5–3% to 8–40% — including one case of 3% → 30%, a 10x improvement — purely from getting positioning language right (First Round Review).

Weak positioning isn't a marketing problem. It's a revenue problem hiding inside marketing.

---

What positioning actually is (and isn't)

"Positioning is the act of deliberately defining how you are the best at something that a defined market cares a lot about." — April Dunford, An Introduction to Positioning

"Positioning is the single largest influence on the buying decision." — Geoffrey Moore, Crossing the Chasm

Positioning is the frame. Messaging is the copy that lives inside the frame. Brand is the emotional shell wrapped around it. Confusing the three leads teams to A/B test taglines for months while the underlying frame goes unvalidated.

See also: Messaging Testing Guide and Brand Perception Survey Guide.

---

April Dunford's 5 components — the framework positioning research validates

From Obviously Awesome, the canonical positioning model has five components plus an optional sixth. Positioning research's job is to put a customer-validated answer next to each one.

1. Competitive Alternatives — "What our target audience would use if our solution did not exist." (Could be a competitor, a spreadsheet, an intern, or "do nothing.")
2. Unique Attributes — "The features and capabilities that our product has that alternatives do not."
3. Value (and Proof) — "The benefits of those features to our target buyers and how they address their pain points," with provable evidence.
4. Target Market Characteristics — "The attributes of a group of buyers that make them care about the value of our solution."
5. Market Category — "The market we define our solution as fitting into to better help illustrate the value to our target buyers."
6. Relevant Trends (optional) — the "why now" context.

Source: April Dunford, A Quickstart Guide to Positioning; summary at Heinz Marketing.

Most teams know the framework. The gap is the interview discipline to populate it with real customer language — not just internal hypotheses dressed up as positioning canvases.

---

What positioning research actually tests

Five distinct hypotheses, each with a different interview technique:

1. Competitive set elicitation

What alternative does the buyer actually default to? Often shockingly different from the named competitor you obsess over.

Question: "If [our product] didn't exist tomorrow, how would you solve [the job]?"

Answers cluster into 4 types: a direct competitor, an indirect substitute (spreadsheet, intern, agency), a do-nothing baseline, or an internal build. Each implies a different positioning.

2. Unique-attribute weighting

Which differentiators do buyers actually care about?

Question: "Of these 5 capabilities, rank the importance from most to least essential."

Koji's ranking structured question type runs this in seconds, then probes the top and bottom items with open-ended follow-ups: why did this matter? When did you realize?

3. Value translation testing

Which benefit phrasing makes buyers lean in?

Question: "Which of these statements best describes the outcome you'd want?"

Note: don't ask buyers to choose between marketing taglines. Probe their language and let them describe the value in their own words. The verbatim you collect becomes the value statement.

4. Market-category fit

Which category does the buyer file you under in their head?

Question: "When you searched for a solution, what terms did you type?" And: "If you had to put us on a G2 or Gartner category list, where do we go?"

Category is the buyer's mental shorthand for price anchor, comparison set, and evaluation criteria. Get it wrong and you compete in the wrong drawer of the buyer's brain.

5. Trigger events / "why now"

What pushed them to start looking now instead of 12 months ago?

Question: "What changed in the last 90 days that made you start looking?"

The trigger event is the urgency narrative. Without it, your positioning describes a destination buyers feel no need to leave for.

---

23 positioning interview questions, organized by Dunford component

Competitive alternatives (5)

1. If [our product] didn't exist, how would you solve [problem] tomorrow?
2. What were you using or doing before you started looking?
3. Who else did you evaluate? Who almost made the shortlist?
4. What did the "do nothing" option look like — what would have happened if you didn't buy anything?
5. Was there a spreadsheet, internal tool, or person doing this job before?

Unique attributes (4) 6. Of features X, Y, and Z, which would you not give up if we removed it? 7. What can we do that [the alternative] can't? 8. Rank these 5 capabilities by importance to you. 9. Was there a single feature that tipped the decision?

Value and proof (4) 10. What measurable outcome did you expect when you bought? 11. Six months later, what actually changed for you or your team? 12. If you had to defend this purchase to your CFO, what's the one-line ROI story? 13. What proof would have helped you decide faster?

Target market (4) 14. Who else at your org pushed back on this purchase? Why? 15. What's true about your team or situation that made this matter so much? 16. Who do you know who needs this — and who definitely doesn't? 17. What would be different about a peer who would NOT buy this?

Market category (4) 18. When you searched for a solution, what terms did you type? 19. If you had to put us in a category on a G2 list, where do we go? 20. In your own words, describe what we do to a colleague. 21. What did you tell your boss this was?

Trigger / why now (2) 22. What changed in the last 90 days that made you start looking? 23. Why didn't you solve this 12 months ago?

---

Who to interview: wins, losses, and never-evaluated

The single most common positioning-research mistake is interviewing only current customers. Their experience is already filtered through the positioning you have. To validate or change positioning, you need three populations:

• Wins — recent customers (90–180 days). Best for value translation and trigger events. Dunford notes: "If we know how we win, we can figure out how to win more."
• Losses — evaluated but didn't buy, or churned within 12 months. Best for competitive-alternative truth and value-gap signal. Hardest to recruit traditionally.
• Never-evaluated — fit your ICP but never considered you. Best for category fit. They reveal which mental drawer you're not in.

Dunford specifically warns against positioning against phantom competitors — competitors you've never actually lost to. Loss interviews are how you find the real ones.

---

When to run positioning research

---

The 8 most damaging mistakes in positioning research

1. Confusing positioning with messaging — testing taglines instead of the underlying frame.
2. Asking customers to choose between marketing copy — they're not creative directors. Probe their language; don't poll yours.
3. Only interviewing current customers — you miss lost prospects and never-evaluated buyers, where category-fit signal lives.
4. Leading category questions — "Would you call us a CRM?" instead of "What category do you put us in?"
5. Asking what features to build instead of what alternatives were considered.
6. Polling positioning instead of probing — quant ranking with no qualitative "why."
7. Positioning against phantom competitors — naming competitors you've never actually lost to.
8. Skipping loss interviews because lost prospects "won't talk." With async AI interviews, they will — see below.

---

How to run a 3-week positioning research sprint with Koji

Traditional positioning research takes 3–4 months and a six-figure consulting engagement. Here's the AI-native version.

Week 1 — Setup and recruitment

• Define the 23 positioning interview questions (use the list above as a starting template).
• Configure a Koji study with a mix of structured types: ranking for value attributes, single_choice for category placement, open_ended for narrative-test ("explain to a colleague"), scale for buying-process pain.
• Segment outreach into wins (last 180 days), losses (lost or churned in last 12 months), and never-evaluated (ICP-fit but no sales touch).
• Send personalized interview links — async completion rates outperform scheduled video by a wide margin, especially for lost prospects who reject calendar invites.

Week 2 — Run interviews at scale

• Run 30–50 interviews across the three populations.
• Koji's AI-moderated interview approach asks every participant the identical question set the identical way, eliminating moderator drift across a long study.
• Watch real-time themes emerge as transcripts complete.

Week 3 — Synthesize and validate

• Use Koji's thematic analysis to cluster competitive alternatives, value statements, and category language.
• Identify which positioning components are validated, which need revision, which were dead wrong.
• Re-run a 10-interview "confirmation pass" with revised value statements to validate the new frame.

The output isn't a deck. It's a positioning canvas where every cell cites the customer language — and the count of customers — supporting it.

---

Why AI-moderated interviews are uniquely good for positioning research

Speed. Positioning research has to fit between strategic decisions — a launch, a rebrand, a fundraise. The traditional 3–4 month cycle means most teams skip it. 30–50 interviews in 2 weeks makes it actually feasible.

Reach lost prospects. Lost prospects won't take a sales call but will complete a 10-minute async AI interview because the friction is low and the incentive aligns. This is the most valuable positioning data — and traditional methods can't capture it.

Hybrid structured + open. Koji's six structured question types (open_ended, scale, single_choice, multiple_choice, ranking, yes_no) let you quantify category-fit and rank attributes in the same session as open-ended language probing. Traditional interviews force a choice between scoring and narrative.

Automatic language clustering. Theming across 30+ interviews reveals which positioning language repeats. This is the closest operational definition of language/market fit — and it's only practical when you have enough volume to cluster.

Iteration loops. Run the same script with revised category framing or value statements weekly. Positioning becomes A/B-testable instead of an annual offsite ritual.

---

Frequently asked questions

What's the difference between positioning research and messaging testing?

Positioning research validates the strategic frame — category, competitive alternatives, value, target. Messaging testing validates specific copy that lives inside that frame. Run positioning research first; messaging testing only works when the underlying positioning is right.

How many customer interviews do I need to validate positioning?

Most teams hit theme saturation between 20–30 interviews per population (wins, losses, never-evaluated). Aim for 30–50 total across all three. With async AI interviews, this is a 2-week project. With traditional moderated research, it's 3–4 months. See Data Saturation in Qualitative Research.

Should I interview current customers, lost prospects, or both?

All three populations: current customers (wins), lost prospects (losses or churned), and never-evaluated ICP-fit buyers. Each surfaces different signals. Only-current-customer research is the single most common mistake.

Can I use a survey instead of interviews for positioning research?

Surveys can quantify which attributes matter (ranking questions) but can't surface the verbatim category language and trigger-event narratives that positioning needs. Hybrid sessions — Koji's structured + open-ended questions in one interview — give you both. Pure surveys force a choice you don't need to make.

When in our company's lifecycle should we do positioning research?

Two best moments: (1) you have ~10+ closed customers and can mine win patterns, or (2) you're entering a new segment, repositioning, or your pipeline shows the 56% indecision signal. Pre-PMF, positioning is hypothesis-driven; post-PMF, it should be evidence-driven.

How do I test category language without leading the witness?

Open with "What terms did you type when you searched?" and "How would you describe what we do to a colleague?" Let buyers name the category in their own words before you offer any options. Only later, use single_choice to validate which of 3–5 candidate categories best fits — and probe the choice with an open-ended follow-up.

---

Related Resources

• Structured Questions Guide — Koji's six response types and how to combine ranking + open-ended for positioning research
• Messaging Testing: How to Find Copy That Converts (with Real Customers)
• Switch Interviews: The JTBD Method for Understanding Why Customers Buy (and Leave)
• Win/Loss Analysis: How to Learn Why You Win and Lose Deals
• Value Proposition Canvas: The Complete 2026 Guide
• Customer Discovery Interviews: The Complete Guide
• Pricing Research Interviews: How to Understand What Customers Will Pay

---

April Dunford's Obviously Awesome gave product teams the framework. Customer interviews are how the framework gets populated with truth. AI-moderated platforms like Koji are what make running enough of those interviews — fast enough — to actually validate positioning before the next launch instead of after.