How to Conduct B2B SaaS Customer Testimonial Interviews

Customer research · Customer interviews

Many B2B SaaS companies already run testimonial interviews, usually as part of producing a case study, and treat the call as a source of marketing material and nothing more. A customer who sits down to explain why they chose you and what changed afterward is giving you quotes and a story, and those are what the call was booked for. In the same forty minutes the customer also tells you, in their own vocabulary, how they describe the problem you solve and which of your strengths they mention without being prompted. Set beside the words on your website and in your messaging, that vocabulary is a check on your value proposition, and it costs nothing beyond the call you were already running. The method below collects the marketing material and the research from one conversation. In our experience the research is the part companies drop: the quotes are published, and nobody compares the transcript with the messaging it was supposed to support.

What are customer testimonial interviews for?

Customer testimonial interviews produce marketing assets and evidence for your value proposition

The marketing half is familiar. You want quotes, a written story, a video if somebody is filming, and a named customer willing to stand behind all of it.

The research half comes free and is less familiar. Across six or seven of these calls you hear how customers describe the problem you solve, in their own vocabulary and with their own emphasis. You can then set that description beside your website and your messaging and see where the two differ. The method we use comes from T2D3, and T2D3's value proposition framework names customer testimonial interviews as one of its inputs. The question for the research half is whether customers repeat what you believe your strengths are, and what language they use when they do.

Turning those calls into something usable takes a spreadsheet. Every quote is tagged with the customer's ICP segment, the speaker's persona and the value proposition pillar it supports, so that each claim in your messaging framework has customer quotes beside it. The section on what happens either side of the call describes this grid in full.

The research half has one limit you should know before you start. The customers who agree to these calls are your happiest, so what they say gives you candidate messages to test with other customers, and it does not describe your market as a whole. The limitations section explains how to handle this.

When should you conduct customer testimonial interviews?

Run testimonial interviews soon after a customer gets a result

The trigger is a customer who has recently reached something worth describing: a completed rollout, a metric that moved, or a process that now runs differently. Recency matters because detail fades. Somebody three weeks past a successful implementation can still tell you what happened in the first month and in what order. Somebody two years past it gives you a summary they have told several times.

Start earlier than your deadline. The call takes forty minutes, but approval can take weeks, because a testimonial often has to clear the customer's communications team and sometimes their legal team as well. If the asset is needed for a launch, book the interview at least a month before the launch date.

Who should you talk to in a testimonial interview?

Recruit testimonial interviewees from your happiest customers

A testimonial round starts with your happiest customers, because they are the ones who agree to speak on the record. There is no way around this, and it is also the main weakness of the method as research: everything the round tells you comes from people selected for their enthusiasm. Recruit the enthusiasts anyway. The limitations section explains what their enthusiasm means for how you use the findings.

Within that constraint, choose for variety. Three customers from the same segment saying similar things produce one usable asset between them. One customer from each segment you sell into, with a different persona in the chair each time, gives you quotes you can put in front of different buyers and gives the research half something to compare.

Where the signatory and the person who lived through the change are different people, ask for the second one. The signatory gives you a sentence about strategy. The person who ran the rollout gives you the story with the detail in it: who was involved, what broke, and how long it took. In our experience readers who have bought software themselves find that detail more persuasive than a statement of strategic intent.

What does a typical testimonial interview look like?

The standard 40 minute format

The call has six blocks, lettered A to F. The other guides in this series use numbered sections taken in a fixed order. A testimonial call is organized around subjects instead, and the time moves between blocks depending on where the customer has the most to say. We publish a longer list of twenty four testimonial interview questions separately. The cards below hold the twenty we ask most often in this structure, and in most calls you will not ask all of those either.

Block D asks what went wrong, past the midpoint of the call and immediately before the outcome questions. The placement can look calculated, and there are two reasons for it. In blocks B, C and E the customer knows they are producing an endorsement and answers with that in mind, while block D asks about problems, and its answers are the part of the transcript least shaped by that knowledge. In our experience block D also makes the answers in block E more generous, because somebody who has just said something critical tends to balance the account without being asked.

The research half loses nothing from this placement. Approval does, since what block D produces is often what a communications team removes. Write the draft expecting some of your strongest quotes to be cut, and keep the ones that are cut for internal use.

Block A · 3 minutes

Permissions and framing

Say three things plainly at the start of the call, and then ask one question. Say that you would like to record the call, and what the recording is for. Say what you hope to produce from it: a written story, quotes on a page, perhaps a video. Then say that nothing is published until they have read it and approved it.

The promise of approval is what allows people to speak freely in block D, so repeat it at the end of the call and keep it afterward. The question about approvers matters because approval is the step that takes weeks. If you know in the first minute that a communications team has to sign off, you can send the draft to the right person first.

Block B · 5 minutes

Customer background

A reader needs to know who this company is before anything else in the story means much, and the customer describes themselves better than your write up will.

Keep this block moving. Customers are comfortable describing their own company and the block overruns easily, while the case study needs only about three sentences from it.

Block C · 9 minutes

Using the product

The first card has four questions about how the customer came to choose you, and we ask them one at a time. A single question about why they chose you produces a tidy summary. The four separate questions produce the material the summary was made from.

Customers often do not expect the third question in a testimonial call. Ask it anyway. Whatever nearly stopped this customer may be stopping prospects now, so your sales team will want the answer, and the section on the mapping grid explains where it goes. If the hesitation was overcome, it can also go into the published story, for the reason given under block D.

Work through these by laddering, which means taking whatever the customer gives you and asking why that was important for their business. The first answer is often a general one, such as price or ease of use, and one more question usually reaches the specific business reason behind it.

Block D · 9 minutes

Working with your team

This block covers the implementation and the working relationship with your team, including the parts that did not go well.

A problem that was resolved can go into the published story, and in our experience it makes the story more credible to readers who have implemented software themselves. A problem that was not resolved will not appear in the asset. It goes to your own team as a finding.

Block E · 10 minutes

Outcomes and return

Most of the quotes come from this block. For every benefit the customer claims, ask for the specific occasion behind it.

Always ask where a number came from, and ask politely. A figure given from memory may not match what the customer's finance team sees when they read your case study, and a statistic withdrawn after publication costs you the asset and some goodwill with the customer. Ask which system the number came from and what period it covers. If the answer is vague, use the story and leave the number out.

In our experience the question about a peer is where the quoted sentence usually comes from. It asks the customer to speak as they would to a colleague, so the answer comes out in their ordinary vocabulary.

Block F · 4 minutes

Close

Count to seven after the first question. Then repeat the promise about approval, say roughly when they will see a draft, and keep to the date.

What happens either side of the call

The testimonial process has five steps and a mapping grid

The interview is one of five steps. In our experience the other four are where testimonial programs stall.

1. Source the customer.

Ask customer success which customers have recently reached a result, or draw on your advocacy program if you have one. Make the approach through whoever holds the relationship with the customer.

2. Run the interview.

Use the structure above, and record the call with permission.

3. Write it up.

Draft within a few days, while you still remember how the conversation sounded. Quote accurately, tidy only the filler, and never improve a sentence into something the customer would not have said. Our guide to writing a customer success story covers the structure of the written piece.

4. Get approval.

Send the draft with a date for comments, and expect edits. We usually find that most of them fall on material from block D. Where a communications team softens a passage until it no longer says anything specific, drop the paragraph and do not publish the softened version.

5. Publish and reuse.

One interview gives you a case study, several quotes, a slide, and a video if you filmed it.

The mapping grid is the research half of the round, and it takes about twenty minutes per interview. Tag every usable quote with four things: the ICP segment the customer belongs to, the persona of the speaker, the value proposition pillar the quote supports, and, for the answers to "what almost stopped you" and "how was it resolved", the objection it answers. Tag in the same week as the call, because the tone in which a sentence was said is hard to recover from a transcript later.

Sales enablement takes the objection tag from the round. An objection the customer raised, together with the resolution they describe in their own words, goes straight into objection handling. Objections that stopped a prospect from buying cannot appear in this round, because those prospects did not become customers. Win, loss and churn interviews collect them.

Once the round is tagged, the grid shows each pillar of your messaging with the customer quotes that support it, and the pillars with few or no quotes are the main research finding. A pillar with no quotes against it after six interviews is a claim your customers are not making on your behalf. Either the claim is not true, in which case the messaging should change, or it is true and customers have not noticed it, in which case the product or the onboarding has to make it visible.

The number six comes from the qualitative research literature. In an analysis of sixty interviews with women in two West African countries, the basic elements of the main themes were present after six interviews, and saturation, the point after which further interviews added little that was new, was reached within the first twelve (Guest, Bunce and Johnson, 2006). That study covered one population and one subject, which is a more uniform group than most B2B customer bases. Read the gaps across the whole set of interviews. At one interview per segment, a gap within a single segment reflects what one person happened not to say.

limitations

What a testimonial interview does, and doesn't, tell you

A testimonial round gives you publishable quotes, the customer's own account of what was wrong before and what they compared you against, what nearly stopped them, and a messaging grid with quotes in it. For a company that has never run a research program it is an inexpensive place to start, because many companies already run these calls for their case studies.

The findings describe the customers you serve well, and they tell you nothing about the market that did not buy or about the customers who left. You are interviewing people who agreed to be interviewed about a product they still use and will praise in public, recruited by starting with your happiest accounts. No change to the questions alters who is in that sample.

The second limitation is the satisfaction effect described in the overview. A testimonial call is a conversation about satisfaction, held out loud, with somebody connected to the vendor, and people give a more positive account to a person than to a form. Pew Research Center randomly assigned 3,003 respondents to answer sixty questions either by telephone with an interviewer or on the web, in July and August 2014. The largest gap was on satisfaction with family life: 62 percent of telephone respondents said they were very satisfied, against 44 percent of web respondents, a difference of eighteen points; the average gap across the sixty questions was 5.5 points (Pew Research Center, 2015). Where answers can be checked against records, the gap turns out to be misreporting. In a study of about a thousand recent University of Maryland graduates, randomly assigned to a telephone interview, an automated voice questionnaire or a web questionnaire, answers to sensitive questions about their academic record were checked against university records, and the telephone interview with a live interviewer produced the least accurate answers and the web the most accurate (Kreuter, Presser and Tourangeau, 2008). Neither study had a vendor in the room, so applying the finding to a testimonial call is our extension of it. Block D offsets part of the effect, and we do not know how much.

Both limitations lead to the same way of using what the round produces, and it has two parts.

First, use the quotes, the story and the video for marketing as planned. Nothing above makes them less useful, because a customer willing to say in public why they chose you persuades prospects whether or not their account is representative of your market.

Second, treat the research half as a source of hypotheses. A phrase that recurs across five testimonial calls is a good candidate for your messaging, because five customers have reached for the same words without being prompted. It is a weak measurement of your market, because those five were selected for their enthusiasm. The difference decides where the phrase goes next. A candidate goes into the next survey, or the next round of win, loss and churn interviews, to find out whether people outside your happiest accounts use it too. Once it has passed that check, it can go on a slide in front of your board as a description of your market.

What you should watch out for

Dos and don'ts in a testimonial interview

FAQs

Frequently asked questions about customer testimonial interviews

How is a testimonial interview different from a case study call?

It is the same call with a second purpose attached. A case study call ends once you have the quotes. A testimonial interview as we run it also asks what almost stopped the customer and what went wrong, and the work ends with the quotes tagged against your messaging.

Who should run the interview?

The interviewer should be somebody with no relationship with the customer to protect, which is the principle the overview sets for every interview in the series. A customer asked about their experience by the person responsible for it gives the generous version, and block D produces little. The customer success manager or account manager should not run the call. They keep three other jobs: sourcing the customer, briefing the interviewer on the account, and reading the transcript afterward, since they will act on what it says. If you run a customer marketing or advocacy program, this is probably close to the division of labor you already use between the program and the account team.

How do you ask a happy customer what went wrong without souring the call?

Ask it as a matter of fact, past the midpoint of the call, after they have already said positive things. "Was there a point where you were frustrated with us?" is a normal question in a research conversation and an awkward one in a sales conversation. The approval promise in block A helps, because the customer knows they will see the draft and can remove anything before it is published.

What if the customer gives a number you cannot verify?

Ask which system it came from and what period it covers. If that produces nothing solid, use the story and leave the number out, for the reasons given under block E.

How many testimonial interviews do you need?

For assets, as many as you have customers willing to give them. For the research half, six or seven across different segments is the point at which a pillar with no quotes in the mapping grid becomes informative, for the reason given with the grid above.

Can you use the same call for a video testimonial?

Yes, with some adjustments. Tell the customer in advance that the call is being filmed, because people prepare differently, and expect block D to produce less on camera. Keep the interviewer off microphone so that answers stand on their own, and rely on the questions that produce self contained answers: the peer question, and the last time it showed up. Approval then means approving the cut as well as the transcript, so say that in block A. Run the research half from the transcript, and expect the candid material from block D and the usable footage to come from different parts of the call.

What do you do with the quotes that do not get approved?

Keep them for internal use. A sentence a communications team removes is often one nobody inside your company had heard before. Product, sales and customer success can still act on it, even though it cannot be published.