How to build a voice of customer programme for B2B companies

A B2B voice of customer programme runs on four data sources: customer interviews (QBRs, win-loss calls, churn calls, NPS follow-ups), product usage data, support ticket patterns, and sales call recordings. Unlike B2C, you can interview most of your customer base directly. Tools like Skimle make it practical to synthesise across all four sources, connecting themes from interviews with patterns in tickets and usage data.

B2B VoC is one of those disciplines where most companies know they should be doing it well, and most are not. According to Bain & Company research cited by CustomerThink, only 22% of B2B companies consistently measure and act on customer experience. The gap between collecting feedback and actually changing decisions based on it is where most programmes stall.

This guide is for product managers, customer success leaders, and market researchers who want to build a programme that produces real change, not slide decks that get presented once and filed.

Why B2B VoC is fundamentally different from B2C

B2C VoC runs on statistical sampling. You have 50,000 customers and you hear from 2,000. The signal is in the aggregate.

B2B is the opposite. You might have 50 customers, each worth €50,000 or more per year. Losing two or three is a serious business problem. That changes the maths of VoC completely.

Three differences matter most:

You can talk to all of them. A B2B company with 80 customers can run annual deep-dive interviews with every account and quarterly check-ins with strategic ones. This is not sampling; it is complete coverage. The constraint is not volume but consistency.

Multiple people influence every relationship. B2B accounts involve at least three distinct roles: the economic buyer (who approves the budget), the champion (who advocated for the product internally), and the end user (who lives with it daily). Each has different concerns, different metrics for success, and different levels of visibility into what is working. A VoC programme that only reaches champions is systematically missing the people who control renewal decisions.

Relationships span years. B2B customers do not make a decision and move on. They re-evaluate constantly, and their context evolves. The concerns a customer had at onboarding are different from the concerns they bring to a renewal conversation eighteen months later. A point-in-time VoC survey captures only a snapshot of a much longer story.

According to research cited by Spotio, B2B buyers are 57 to 70 per cent of the way through their buying research before they contact sales. By the time your team is in a commercial conversation, the customer has already formed strong views about your product and your competitors. That pre-contact research is largely invisible to you unless you have a VoC programme running before the sales cycle closes.

The 4 VoC data sources for B2B companies

A well-structured B2B VoC programme draws from four sources, each capturing a different kind of signal.

The power of B2B VoC comes from connecting these sources, not from running them independently. A customer who files three support tickets about a particular workflow and rates their NPS at 6 is a very different risk profile to a customer who files the same tickets but scores 9. The interview tells you whether the tickets represent frustration with the product or just normal onboarding friction.

For a broader introduction to VoC methodology, see our voice of customer research guide.

Customer interviews: the richest source

Customer interviews are the foundation of a B2B VoC programme because they are the only source that captures context and causality. Usage data tells you what customers did; interviews tell you why.

The four interview types in a B2B programme serve different purposes:

• QBR conversations (quarterly business reviews): strategic check-ins with economic buyers and champions. Focus on whether the product is delivering against the customer's business goals, not just whether they are happy with features.
• Win-loss calls: interviews with recent buyers and prospects who chose not to buy. These are the highest-information interviews in your programme because they reveal how your product compares to alternatives under real evaluation conditions. See our guide on win-loss analysis for how to structure these systematically.
• Churn calls: conversations with departing customers to understand the real reason for leaving. The reason recorded in the CRM ("moving to competitor", "budget cut") almost never reflects the full picture.
• NPS follow-up interviews: short conversations with detractors and passives to understand the score. A 6 from a mid-market account is a different problem than a 6 from your largest customer.

Support tickets: volume signals you are probably under-reading

Support tickets are underused in B2B VoC because they are usually owned by a customer success or support function that does not have a direct line to product. The result is that ticket patterns inform support staffing but not product decisions.

Reading ticket themes systematically, even monthly, tells you which parts of your product generate the most friction and which customer segments are most affected. When combined with interview data, ticket patterns often reveal problems that customers mention politely in interviews but feel more acutely than their language suggests.

Sales call recordings: your window into the buying decision

Sales call recordings (from tools like Gong or Chorus) capture the evaluation conversation in real time. Prospects explain their current situation, their criteria, the competitors they are considering, and the objections they have to buying from you. This is primary customer voice, and most B2B companies treat it as sales training material rather than a VoC source.

Integrating sales call themes into your VoC synthesis tells you what the market thinks your product does and does not do well, before customers are customers.

How to run a B2B customer interview programme

Who to talk to: the three stakeholder types

The economic buyer, the champion, and the end user care about fundamentally different things. A VoC programme that only reaches one of them is missing two-thirds of the picture.

In practice, most B2B VoC programmes talk almost exclusively to champions. Champions are easy to reach (they are your main relationship contact), motivated to engage (they have internal skin in the game), and inclined to be positive (they advocated for the purchase). This creates a systematic positive bias in your data.

Economic buyers are harder to access but more valuable to interview at renewal time. Their view of the product is shaped by outcomes and costs, not by day-to-day usability. A champion who loves the product and an economic buyer who cannot articulate the business value to the CFO is a churn risk.

End users are the group most likely to reveal friction. They have less incentive to be diplomatic, and their workarounds (the things they do when the product does not work the way they need) are the most revealing signal in B2B VoC. Look for our guide on customer discovery interview design for practical question frameworks you can adapt for these conversations.

How often to run interviews

The cadence depends on the account's strategic importance and lifecycle stage:

• Strategic accounts (top 20% by revenue): quarterly conversations with at least two stakeholder types. Annual deep-dive covering all three.
• Mid-market accounts: annual interview, plus lifecycle-triggered calls at onboarding completion, first renewal, and any support escalation.
• All accounts: post-churn interview within two weeks of cancellation. This is non-negotiable, as waiting longer reduces recall and willingness to engage.

B2B VoC is not a quarterly project; it is a continuous programme with structured touchpoints. For a practical look at what continuous customer research looks like at scale, see always-on customer research.

What to ask: questions by interview type

The questions that work in a QBR are different from the questions that work in a win-loss call or a churn call. Using the same discussion guide across all interview types is one of the most common mistakes in B2B VoC.

QBR questions (focus: outcomes and strategic alignment)

• "What were the specific outcomes you were trying to achieve when you started using [product]?"
• "Looking at the last six months, where have you seen those outcomes? Where haven't you?"
• "What's changed in your organisation that has affected how you use the product?"
• "If we could change one thing about how the product supports your team's goals, what would it be?"

Win-loss questions (focus: evaluation and decision)

• "Walk me through how the evaluation decision was made: who was involved and what criteria mattered most?"
• "How did we compare to the other options you considered, on the things that mattered most to you?"
• "Was there a moment in the evaluation when the decision tilted one way or the other? What happened?"

For a full set of win-loss questions, see our dedicated win-loss interview questions guide.

Churn questions (focus: the real reason and timing)

• "When did you first start thinking about whether this was still the right solution for you?"
• "What was the final factor that made you decide to move on?"
• "What would have had to be true for you to have stayed?"

NPS follow-up questions (focus: understanding the score)

• "You gave us a score of [X]. Can you tell me a bit about what was behind that?"
• "What's one thing we could do differently that would move that score up?"

For the analytical framework behind NPS follow-up research, see how to analyse NPS verbatim comments.

Good interview design draws on jobs-to-be-done thinking: what is the customer actually trying to accomplish, and where does your product fit into that job? See our jobs-to-be-done interview methodology guide for how to apply this framework to customer interviews.

Synthesising B2B VoC across sources

Most B2B companies collect VoC data in multiple disconnected systems. Interviews are in call recording software or shared notes documents. Tickets are in Zendesk or Intercom. NPS scores are in Delighted or Typeform. Sales calls are in Gong. Nobody synthesises across them because it takes too long and there is no clear process.

The result is a set of parallel VoC streams that each tell a partial story. Product hears about friction from support tickets. Customer success hears about strategic concerns from QBRs. Sales hears about competitive positioning from deal calls. None of these teams are systematically sharing what they learn, and nobody is connecting the dots across sources.

Building a shared insight repository

The first step is structural: bring all the qualitative VoC data into a single place where it can be coded and cross-referenced. This means interview transcripts, call recordings (auto-transcribed), open-text ticket notes, and NPS verbatim responses all in the same analysis environment.

Skimle is built for exactly this pattern. You can import interview transcripts, call notes, and open-text responses into a single project, run automatic thematic analysis across the full dataset, and get every insight traceable back to the source quote. When a theme appears in both interview data and ticket data, you can see both sides of it in one view. The collaboration features let product, CS, and sales teams access the same repository rather than hearing about findings through presentations.

For how to approach interview analysis systematically, see how to analyse customer interviews.

Connecting interview themes with usage data

Usage data tells you what customers do; interviews tell you why. The combination is more powerful than either alone.

A useful analysis pattern: identify the 10 accounts with the highest support ticket volume in the last quarter. Pull their interview transcripts (if you have them) and look for themes that show up in both sources. When ticket themes and interview themes overlap, you have a signal that is both large enough (tickets show the volume) and deep enough (interviews explain the context) to bring to a product decision meeting with confidence.

The same pattern applies to churn risk. Accounts that show decreasing product engagement in usage data combined with specific friction themes in interview data are higher-risk than accounts where only one signal is present.

What to do with the conflicting signals

B2B VoC data frequently contains contradictions. A customer says in a QBR that they are satisfied, then files three tickets about the same problem over the next six weeks. A churned customer tells you price was the reason they left, but their usage data shows they stopped logging in two months before the cancellation.

The right response to contradictions is not to pick the signal you prefer; it is to interview more deeply. Contradictions are usually hiding something more interesting than either signal alone.

From VoC insights to decisions

Data collected and not acted on is worse than data not collected, because it destroys the willingness of future participants to engage. When customers invest time in an interview and nothing changes, they stop agreeing to interviews.

Product roadmap influence

VoC insights enter product decisions most reliably when they are attached to specific evidence, not summarised themes. A product manager is more likely to prioritise a feature when shown five direct quotes from customers explaining exactly what workflow breaks without it than when told "customers mentioned [feature] in interviews."

This is why source traceability matters in your analysis tool. The claim "our enterprise customers cannot complete X workflow without [feature]" is a different level of conviction than "customers mentioned X in interviews." The first is citable; the second is not.

Customer success plays

VoC data informs CS plays in two ways: early warning and playbook development.

Early warning: accounts whose interview themes shift from "productive friction" to "strategic misalignment" are approaching churn before any usage signal appears. A QBR where the economic buyer cannot articulate business value is a red flag that should trigger a proactive intervention, not a wait-and-see.

Playbook development: when you identify a theme that appears in the churn interview data ("we never got buy-in from the finance team for the renewal"), you can build a CS play that addresses it earlier in the customer lifecycle.

Sales messaging refinement

Win-loss interview themes tell you which parts of your sales story land and which do not. When multiple win interviews cite the same moment ("the demo of [feature X] was what made us decide"), that is a signal that your demo should lead with that feature. When loss interviews mention the same objection, that objection belongs in your sales training.

Strategy consultants doing commercial due diligence use exactly this pattern at scale: interview large numbers of customers and prospects to build an objective picture of a company's competitive position. For that use case, see how Skimle supports consultants and investors.

5 common mistakes in B2B VoC programmes

1. Confusing scores with insight

NPS, CSAT, and CES are health metrics. A declining NPS tells you something is wrong; it does not tell you what. Treating a score improvement as a VoC success metric is like treating your temperature as a health goal rather than a health signal. Scores are useful triggers for investigation, not substitutes for it.

2. Only hearing from champions

As discussed above, champions are systematically inclined to be positive, easy to access, and not representative of the full relationship. A B2B VoC programme that cannot point to regular interviews with economic buyers and end users has a coverage problem.

3. Running VoC in one function and not sharing

Customer success runs QBRs, product hears from support tickets, sales has the deal call recordings. These are all VoC sources, and they are almost never synthesised together. If there is no person or process responsible for pulling insights across all three, the programme is producing data but not intelligence.

4. Treating VoC as a project, not a programme

A one-time customer research initiative produces insights that are already ageing by the time you act on them. B2B markets move, competitive dynamics shift, customer contexts evolve. A programme runs continuously, with structured touchpoints, systematic analysis, and a regular cadence for sharing findings with decision-makers. See how companies structure continuous customer research for a practical model.

5. Not acting on findings, then wondering why participation drops

This is the most damaging mistake, because it is self-reinforcing. Customers who invest time in an interview and see no change stop saying yes to the next one. The quality of your data depends on the perceived value customers get from participating. Closing the loop (telling customers what changed because of what you heard) is not just good relationship management; it is how you protect the quality of future data.

Frequently asked questions

How many customer interviews should a B2B company do for VoC research?

There is no fixed number, but a useful minimum is: every strategic account (top 20% by revenue) at least once a year, plus lifecycle-triggered calls at key moments for all accounts. For a company with 50 customers, that might mean 30 to 40 interviews per year across QBR conversations, churn calls, and annual reviews. The aim is coverage, not sample size.

What is the difference between B2B and B2C voice of customer programmes?

B2C VoC relies on statistical sampling because customer counts are large (tens of thousands) and individual relationships are short. B2B VoC is built on direct relationships with a smaller, higher-value customer base where each account has multiple stakeholders and relationships span years. B2B programmes prioritise interview depth and stakeholder coverage over survey volume.

Who should own the voice of customer programme in a B2B company?

Ownership varies by company size, but the most effective structure keeps the programme in a function that has a reason to synthesise across customer success, product, and sales rather than optimising for one team's needs. Customer insights, market research, or a dedicated revenue operations function often works well. The worst outcome is three separate VoC streams running in parallel without anyone connecting them.

How do you synthesise VoC data from multiple sources?

Start by bringing qualitative data into a single analysis environment. Import interview transcripts, call notes, and open-text responses into a project, code them against a consistent set of themes, and look for patterns that appear across multiple sources. When a theme appears in both interview data and support tickets, the convergence gives you stronger evidence than either source alone. Tools like Skimle make this practical at scale by automating the initial theme extraction while keeping every insight traceable to its source.

Building a B2B VoC programme that actually influences decisions? Try Skimle for free. Bring your interview transcripts, call notes, and open-text responses into one place, and surface the themes that span your full customer dataset, with every insight traceable back to source.

Related reading:

• Voice of customer research: how to build a VoC programme: the full methodology behind VoC design and analysis
• Win-loss analysis: how to systematically learn from deals: turning sales call data into competitive intelligence
• How to analyse customer interviews: a systematic approach to interview synthesis

About the authors

Henri Schildt is a Professor of Strategy at Aalto University School of Business and co-founder of Skimle. He has published over a dozen peer-reviewed articles using qualitative methods, including work in Academy of Management Journal, Organisation Science, and Strategic Management Journal. His research focuses on organisational strategy, innovation, and qualitative methodology. Google Scholar profile

Olli Salo is a former Partner at McKinsey & Company where he spent 18 years helping clients understand the markets and themselves, develop winning strategies and improve their operating models. He has done over 1000 client interviews and published over 10 articles on McKinsey.com and beyond. LinkedIn profile

Sources

• How B2B voice of the customer maturity drives growth - CustomerThink
• B2B Sales Statistics - Spotio