B2B Customer Surveys: Questions, Process, and What to Do With the Answers

Why B2B customer surveys often lead nowhere

Business customers rarely say on their own what bothers them. They buy less, get a second quote, or leave quietly. The article why customers leave describes this silence as the real warning sign. A customer survey is meant to break it. In practice, though, it fails at two points: either nobody ever asks because there is no time, or someone asks and the answers end up in a spreadsheet that nobody opens again.

So the question is: how does a small B2B business ask its customers so that the answers actually lead to better decisions? To answer it, we first clarify the terms and a simple mental model, then what research shows about the Net Promoter Score and open-ended questions, weigh two views, and finally set out a process and a rule for analysis.

A survey that leads to nothing is worse than none, because the customer has learned that their answer changes nothing.

Terms: survey, conversation, importance, and performance

A customer survey is a planned collection of customers’ assessments that is the same for all respondents, written or oral. It differs from a chance conversation in that the answers are comparable. In B2B there is an additional feature: the customer is a company in which several people have a say. The managing director judges something different from the person who works with your team every day.

Two terms from research help with analysis. Stated importance is what customers answer to the question “How important is … to you?” Derived importance is calculated: a topic is important if customers who rate it well are more satisfied overall or more likely to recommend. The two measures often do not agree (Bettencourt and Houston 2024). Customers like to name price as important, but their overall satisfaction often depends more on reachability or keeping deadlines.

The Net Promoter Score (NPS) is based on a single question: “How likely is it that you would recommend us to a colleague?” on a scale from 0 to 10. Those who give 9 or 10 count as promoters, those who give 0 to 6 as detractors. The score is the share of promoters minus the share of detractors. The approach goes back to Frederick Reichheld (2003) and is widely used in companies.

The mental model: the importance-performance grid

The mental model of this guide goes back to the importance-performance analysis by Martilla and James (1977). For each topic it asks two things: how important is it to the customer, and how well does the provider deliver it from the customer’s point of view? Plotting the two values against each other creates four fields, each with its own consequence. Bettencourt and Houston (2024) use the same logic in their study and speak of need gaps: topics that are important and poorly met.

The grid has an advantage that matters in small businesses: it translates answers directly into an order of priority. In the example, the problem is not technical quality but communication when something goes wrong. That is a typical finding. Customers take technical competence for granted and judge the provider by how it handles deviations.

What research shows about the Net Promoter Score

Research has examined the NPS critically for years. The criticism is directed less at the question itself than at the promise that a single number is enough. Two recent studies show where the score holds up and where it does not.

Bettencourt and Houston (2024) examined whether the open-ended comments that follow the recommendation question in NPS surveys deliver the right improvement topics. In the B2B part of the study, they analyzed the answers of 521 decision-makers at North American car dealerships about their industry software and compared them with a second survey of 228 decision-makers who rated the importance and performance of individual needs on closed-ended scales. About 42 percent of respondents in the second survey were owners or partners. The result: the topics mentioned most often in open-ended comments matched the topics that actually explained willingness to recommend only weakly to moderately.

Two side findings are particularly useful in practice. First, the comments of detractors and passives were closer to the actual drivers than those of promoters. Satisfied customers praise in general terms; dissatisfied ones are specific. Second, agreement depended on wording: the question of what the provider could do better produced more useful answers than the question about the reasons for the rating. The authors therefore recommend asking detractors in particular about improvements and supplementing open-ended comments with closed-ended questions on importance and performance. Their study concerns one industry and one product; whether the findings transfer to other industries without qualification is open.

Kim and Gruca (2024) tested the link between NPS groups and later purchasing behavior, though among consumer customers of casinos. Promoters did later spend more and were more likely to stay than detractors. Simple satisfaction questions, however, predicted behavior about as well, and the difference in later revenue was measurable but small.

Two views: one number or a conversation

Simplicity speaks for a single metric. Every customer answers one question in ten seconds, and a score collected the same way every year shows a direction. For a business that has not asked at all so far, that is a big step forward. The study by Kim and Gruca (2024) confirms that the groups say something about later behavior.

Against the single number speaks the fact that it does not say what to do. A score of 30 is neither good nor bad as long as nobody knows what it depends on. Small businesses also have few customers. With 40 answers, three customers shift the score by several points without anything having changed.

I therefore consider a combination to be right: a short written survey with the recommendation question and five to seven topics rated for importance and performance, supplemented by personal conversations with the customers whose answers stand out. The number shows the direction, the grid shows the order, the conversation shows the reason.

The process: a customer survey in six steps

The following process is intended for a business with a few dozen to a few hundred business customers. Spread over six weeks, it needs about two to three working days.

1. Define the goal and the decision

   Write down beforehand which decision the survey is meant to prepare, for example which process will be improved in the next six months. No decision, no survey.

2. Select respondents

   All active customers of the last twelve months, ideally two people per customer: the person who decides and the person who works with your team. How to rank customers by importance is shown in the guide ABC customer analysis.

3. Keep the questionnaire short

   The recommendation question, five to seven topics with one question each on importance and performance, and one open question about improvements. More than five minutes of answering time costs responses.

4. Announce it personally

   A short message from the owner or the regular contact person, saying when and how the results will be reported back. One reminder after a week is enough.

5. Follow up with conversations

   Hold a 20-minute conversation with about ten customers: with detractors, with large customers, and with a few who answered very positively.

6. Analyze and report back

   Calculate need gaps per topic, assign the conversations, decide on three measures, and tell customers what will change.

For the open question and the conversation, wording that asks about improvement rather than praise works best. It resembles the questions the guide win-loss analysis suggests for customers after a rejection.

1. “What should we change first to make working with us easier for you?”
2. “On what occasion in the past year were you dissatisfied with us?”
3. “What do you do yourself today that we could take over for you?”
4. “How can you tell that a supplier like us is doing good work?”
5. “Who in your company should we ask more often?”

What to do with the answers

The most important step is the last one. The article win-loss analysis describes a field experiment by Kaul and colleagues (2025): small shops that collected feedback from their customers and turned it into changes increased revenue and profit compared with a control group. The effect came from the changes, not from asking alone. This leads to the rule of this guide: every survey produces exactly three measures, each with a person responsible and a date.

• The decision the survey prepares is fixed before it is sent.
• Ideally two people per customer have been surveyed.
• Importance and performance have been collected for each topic, not just satisfaction.
• Detractors and passives have been asked specifically about improvements.
• There are three measures with a person responsible and a date.
• Customers have been told what will change.

Reporting back to customers is often forgotten, yet it is the part that builds trust. A short letter with the three most important results and the changes decided shows that the answer was heard. It is also a good occasion for the next conversation, as described in the guide growing existing customers. If you want to involve some customers permanently, you can later form a customer advisory board from your conversation partners.

Conclusion: questions that lead to decisions

How does a small B2B business ask its customers so that better decisions follow? It defines beforehand which decision the survey prepares, asks about importance and performance for each topic instead of only about satisfaction, seeks conversations above all with critical customers, and at the end decides on three measures that it communicates to customers. Research on the Net Promoter Score shows that a single number can indicate a direction, but open-ended comments alone are an unreliable compass for priorities.

1. This week: Write down the one decision a survey is meant to prepare and choose five to seven topics.
2. In two weeks: Send the short survey to all active customers, with a personal announcement.
3. In six weeks: Analyze the need gaps, hold ten conversations, and decide on three measures.

Customers do not expect a supplier to be flawless. They expect it to listen and do something with what it hears.