How to Conduct B2B SaaS Concept and Pricing Interviews

Customer research · Customer interviews

Somewhere between having an idea and committing a quarter of engineering time to it, somebody has to find out whether customers understand the idea and what they would pay for it. A concept and pricing interview is the cheapest way we know to ask. It is also the easiest of the five interview types to contaminate. Anything you show a customer changes how they describe their own work, and any number you say first becomes the number they answer around, so a badly run round gives you back your own price, repeated by five different companies. Run carefully, these calls tell you which of your concepts customers can describe back to you accurately, and what range of money each one is worth to them. They will not tell you whether anyone will buy, and the limitations section below explains why. Read it before you book the first call.

What are concept and pricing interviews for?

Concept and pricing interviews tell you what customers understood and what they would pay

A concept interview takes an idea that does not exist yet, or one that exists but has never been priced, and puts it in front of the people who would buy it. Three things come out of the call: whether the customer can describe the idea back to you after hearing it once, whether they can point to a recent occasion where it would have helped them, and what range of money they attach to it before you say any number of your own.

The pricing half is usually why the call gets booked, and it is the half most likely to produce a clean answer that is wrong. The limitations section covers this at length, and the short version is worth having before you read the method. The most dependable outputs of a concept round are the ranking between concepts and the record of what people misunderstood. The prices are ranges, and if the decision needs a figure, a survey supplies it afterward.

The output is narrower than what a discovery round produces. Discovery gives you the material for your positioning and messaging. A concept round informs a go or no go decision on two or three specific ideas. It also gives you the candidate price units for a first draft of a pricing model, and it tells you which words to use when the surviving idea launches.

When should you conduct concept and pricing interviews?

Run concept interviews a few weeks before the decision

Two conditions have to hold: you have concepts worth testing, and the decision they inform is close enough that the answers can still change it.

Concepts written before a discovery round are built from the vocabulary you already had, so testing them tells you how customers react to your own assumptions. If you have not run discovery, run it first.

The timing against the decision matters in both directions. Research that arrives the week after a roadmap is locked changes nothing. Research run six months early gets filed and forgotten. We aim for two or three weeks before the decision.

There is one combination we rule out. Do not add a concept to the end of a discovery call because the customer is already there. From the moment somebody sees a concept, they describe their situation in its terms, so the discovery material from the rest of that call cannot be used.

Who should you talk to in a concept and pricing interview?

Talk to people who would hold the budget

You want the people who would pay, which usually means managers and decision makers more than daily users. A user can tell you whether an idea makes sense in their work. Most cannot tell you what their company would spend on it, and when asked they will guess, and a guess looks identical to data once it reaches a spreadsheet.

Decide how many calls each segment gets before you start recruiting. In our experience price perception moves with company size more than with anything else we segment by, so eight interviews split between startups and enterprises leave you with two samples of four. If you can only run eight calls, run them all in one segment and say which segment the answer applies to.

Eight to twelve calls per segment is where a concept round usually becomes readable. You are not looking for saturation in the way you would in discovery. You are looking for the same concept being misunderstood in the same way by three different people.

What is a concept and how do you write one?

Write each concept as sixty to a hundred words you can read out loud

A concept call can only be compared with the next one if the concept is written down, so this step comes before any interview is booked. A concept here means a short written description of what the thing would do for the customer, in plain language, with no feature list and no adjectives from your website. Sixty to a hundred words takes about thirty seconds to read out. Longer than that and you are describing features.

Write one card per concept with four things on it: what it is in a sentence, what it would replace or compete with, the description exactly as it will be read aloud, and the questions that go either side. The description has to be written out because three interviewers reading the same words produce comparable answers, and three interviewers improvising do not.

Figure 1. One card per concept, written before any call is booked. The concept is invented for illustration and describes no client.

Write it to be read aloud. Slides are allowed where a concept cannot be understood without a picture, but the default is verbal, because a customer looking at a mockup tends to critique the mockup: the color, the label, the word in the corner. If that starts happening, put the visual away and describe the idea out loud.

What does a typical concept and pricing interview look like?

The standard 45 minute format

Concept calls need more time than a 30 minute discovery interview. The first four sections establish who the customer is, what they do now, and what they pay for it, and those answers are what make the concept answers interpretable. The call has six sections, and about half of it goes to the concepts themselves.

Sections 1 to 4 are the first three sections of a discovery interview plus its pricing perception section, all compressed, with the cross sell section dropped. If you have already interviewed this person during discovery, cut sections 2 and 3 to two questions each and give the time to the concept block. If you have not, leave them as they are, because a price given by somebody whose situation you have not established cannot be interpreted afterward.

Section 1 · 3 minutes

Warm up and context

Open the way you open any call in this series. Say who you are, that this is research and not a sales conversation, what happens to the recording, and that nothing they say goes back to anyone with a quota. Then work out which of the three persona types you have in front of you, and whether a purchase of the kind you are about to describe would pass through them.

The second question does two jobs. It tells you whose frame you are getting the price in, and it gives you the names of the people to approach for a second interview in the same account.

Section 2 · 5 minutes

The compressed switch story

The full switch story has four forces, and the discovery guide covers all four. A concept call has time for two. Push and pull are the two we keep, because they tell you what it took for this customer to change supplier the last time they did it. In our experience a concept that answers neither gets compliments in the call and no budget afterward.

Cutting anxiety and habit costs you something. You will not learn what would stop this customer moving again, and a launch plan needs that material. Pick it up in the trade off question later, or in a separate round.

Section 3 · 5 minutes

Current value and frustration

Find the most recent occasion when the current way of working cost them something. The concept gets measured against that occasion later in the call, and customers who cannot recall one tend to answer the concept questions in the abstract.

Do not ask whether they are satisfied with what they have, for the reasons set out in the overview.

Section 4 · 4 minutes

Pricing perception before any concept appears

This section establishes the price range the customer already thinks in. Somebody who treats your category as a line item under thirty thousand dollars a year hears your concept differently from somebody who treats it as a six figure platform commitment.

The third question matters most for what follows. The unit a customer already buys in is the unit they will price your concept in, whatever unit you had in mind, and a concept quoted in an unfamiliar unit gets read as expensive before anyone has done the arithmetic.

Section 5 · 25 minutes

The concept block

Run two concepts if sections 1 to 4 ran long, and three if they did not. Each concept takes six to eight minutes and goes through the same seven steps in the same order. Do not improvise the order on the call.

1. Describe the concept.

Read the card's sixty to a hundred words, which takes about thirty seconds. Do not narrate features and do not sell.

2. Check comprehension before anything else.

If the answer repeats your description back, ask once more. A price given against a concept the customer did not understand looks identical to a real one in the analysis afterward. Write down how they misunderstood it, in their words. The same misunderstanding will reach your launch audience, and you will not be there to correct it.

3. Anchor the concept in something that happened.

Some customers cannot find an occasion. Write that down and move on, and do not help them construct one.

4. Ask the four price questions in order.

These are the Van Westendorp boundaries, and the numbers have to be theirs, never yours. The questions work because they ask the customer to describe a range instead of guessing the answer you want. When somebody refuses to give a number, ask what it would have to save them per month to be worth paying anything for, which gives you a lower boundary from the other direction.

The order differs from the one in the discovery guide, which asks for the two extremes first. A discovery customer is pricing a product they already pay for, so they know roughly where their ceiling and floor are and can name both straight away. A concept customer heard the idea thirty seconds ago and has no price for it yet. Asking from the top down lets them work along a single scale, with their own previous answer as the reference for each new one. Each number is still anchored on the one before it, but the anchor is the customer's own figure, and the rule this method depends on is only that the first number never comes from you. We have not found a published comparison of the two orders, so this is our reasoning and not a tested result.

5. Ask two intent questions at the customer's own prices.

The four boundaries give you a range and tell you nothing about where inside it a customer would move. Somebody who calls two hundred dollars expensive but worth considering, and then scores two out of ten at that price, is describing a range they would not act on. Somebody who scores eight would pay it.

The discovery guide asks a similar question on a five point scale, and we use zero to ten here for two reasons. A discovery customer answers it once, about a product they know. A concept customer answers it for two or three concepts at two prices each, which is up to six scores from one person, and five possible values leave little room for those scores to differ, so concepts the customer would treat differently end up tied. The second reason is the wording. A scale anchored on chance, from no chance to certain, predicted purchases better than a plain buying intentions question when the two were first compared (Juster, 1966), in a study of about 800 American households buying cars and household appliances in 1964 and 1965. Nobody has run that comparison on software bought by companies, as far as we know, so the second reason is weaker than the first.

Figure 2. One customer’s answers for one concept, in illustrative figures per month, from cheapest to most expensive. The two shaded boundaries mark the range this customer would discuss, and the intent scores show whether they would buy inside it: at $300 they probably would, at $600 probably not.

6. Ask the trade off question.

Then ask whether the expensive but worth considering price they gave would move, and by how much. Remove a different element in different calls. You are looking for the part of the concept that the price depends on, and you find it by taking pieces away one at a time and asking what happens to the price.

7. Close each concept with an open question.

Then move to the next concept, and change the order you present them in from call to call. The first concept a customer hears becomes the reference point for the ones after it, so a round where everyone saw concept A first will flatter concept A. Write down the order used in each call. Without that record you cannot tell a concept that performed well from one that was shown first.

Figure 3. Three concepts rotated across six calls. Each concept is shown first in two calls, second in two and third in two. Record the order used in every call.

Two further rules apply inside the block. The first is to test components before bundles. A customer who sees the bundle first has a bundle price in mind for every component question afterward, and the number you most want, the difference between what the bundle is worth and what the parts are worth, only exists if the parts were priced first.

Figure 4. One customer, illustrative figures per month. Price the components first, then the bundle, then ask the removal question. Taking C out of the bundle lowered its price by $100, although C was worth $250 to this customer on its own.

The second is that where a concept involves automation, you ask about trust instead of pushing harder on price. Use these in place of step 4.

Section 6 · 3 minutes

Wrapping up

Ask who else in their organization would have a view on what they have seen. The person who would hold the budget and the person who would use the thing rarely answer the same way.

Ask the open question and then stop talking. Count to seven before you speak again. In my experience the reaction a customer has to a new idea often comes out here, once the structured questions have finished.

limitations

What a concept and pricing interview does, and doesn't, tell you

Four things come out of a round reliably: which concepts were understood and how the others were misread, whether each concept connects to something that happened recently in the customer's work, a stated price range per concept, and some sense of whether the customer would act inside that range. For anything automated, you also get the boundary between the work they would hand over and the work they would not.

What you do not get is whether anyone will buy, and the published evidence on that point is weak.

Stated purchase intent predicts sales well for products that already exist and poorly for products that do not. Across sixty comparable product tests, intent correlated with sales at 0.751 for established products and at −0.177 for new products, the second figure resting on eight tests and not distinguishable from no relationship (Morwitz, Steckel and Gupta, 2007). New products are the case a concept test exists for, so the paper's weakest result falls on the use this guide describes. Those figures come from the paper's second study, sixty packaged goods tests from one manufacturer, so applying them to B2B software is an argument and not a measurement.

Prices from these calls are directional for a second reason. People overstate what they would pay when the payment is hypothetical. The main meta analysis puts the average overstatement at about 21 percent across 77 studies and 115 effect sizes (Schmidt and Bijmolt, 2020). Its studies are all of private goods and services bought by individuals, and no organizational purchase is in the sample. The B2B willingness to pay corrections in circulation therefore borrow a figure measured on a population that contains no B2B buyers. As far as we can establish, nobody has measured hypothetical bias in an organizational purchase. One finding from the same paper does transfer, and it supports the method here: indirect elicitation methods overstated more than asking directly.

Two pieces of folklore need correcting. The claim that people overstate by a factor of three comes from a 2001 meta analysis, and it is a mean pulled upward by a few extreme ratios. A 2005 follow up corrected sixteen coding errors in that dataset and reported a mean of 2.60 beside a median of 1.35, with about half the studies on public goods. The typical overstatement is closer to a third than to a factor of three. And Van Westendorp, the source of the four price questions, comes from a practitioner conference paper of 1976. There are papers comparing it with other methods and papers criticizing the logic of its curve intersections, but we have not found a published validation of those intersections against what people went on to buy.

None of that is an argument for skipping the price questions, and we have not stopped asking them. It does change what the answers are for, in three concrete ways.

Treat each number a customer gives you as the edge of a range they are willing to discuss. Somebody who says four hundred dollars a month would be expensive but worth considering has told you that four hundred is inside the territory where a conversation is possible, and nothing more precise than that. Write it down as a boundary and resist the arithmetic that turns twelve boundaries into an average and an average into a price.

Use the comparison between concepts, because as far as we can tell the biases described above apply to every concept in the round at roughly the same strength. If concept A gets a consistently higher expensive but worth considering figure than concept B from the same twelve people, the gap between them survives the overstatement in a way that neither figure survives on its own. Rank the concepts, and keep the individual numbers out of any deck until a survey has tested them.

Figure 5. Twelve customers’ expensive but worth considering figures for three concepts, in illustrative figures per month. Each dot is one customer and the dark line is the median. The order of the concepts survives the overstatement described above. The figures themselves do not, and they go into a survey before anyone quotes them.

Field a survey if the decision needs a defensible figure. A survey puts the same price questions to a sample large enough to show distribution, with each segment separated, and it asks them on a form instead of out loud. Two things inflate the prices from an interview: the payment is hypothetical, and the answer is given to a person. A survey removes the second and leaves the first. The interviews tell you which concepts and which price units are worth putting in that survey.

The calls produce a worksheet for that decision, with one row per concept per call: whether the customer understood the concept, the misunderstanding in their words if not, whether they found an episode, the four boundaries, the two intent scores, the element you removed and what the price did, and the order the concepts were presented in. Ranking the concepts means reading down the columns of that sheet, and it is what the marketing lead shows the person who has to make the decision.

Figure 6. The worksheet, one row per concept per call, with illustrative entries. Read the comprehension columns first, the order column second, and the price columns last.

What you should watch out for

Dos and don'ts in a concept and pricing interview

FAQs

Frequently asked questions about concept and pricing interviews

How many concepts can you test in one call?

Three at most, and two is more comfortable. Each one needs six to eight minutes for the full sequence. A fourth concept in a 45 minute call gets a shortened version of the method, and its answers cannot then be compared with the others.

How many concept interviews do you need?

Eight to twelve per segment is where a round usually becomes readable. The round is finished when the same misunderstanding has arrived from three different people, whatever number of calls that takes. Keep segments apart when you recruit, since a sample split across company sizes gives you two small samples.

Can you ask pricing questions in a discovery interview?

You can ask about pricing perception, which is what section 5 of the discovery guide does. It gives you the boundaries people describe in words and the unit they buy in. Neither that call nor this one produces a defensible number. For that you need a survey.

Is Van Westendorp reliable?

It is useful and, as far as we have found, unvalidated. The four questions work as a way of getting a customer to describe a range instead of guessing at a price. We have not found a published test, against purchase behavior, of the further claim that the intersections of the resulting curves identify an optimal price.

Should you show visuals or describe the concept out loud?

Describe it out loud by default, and keep visuals for concepts nobody could understand without one. Customers looking at a mockup critique the mockup, and the critique is usually about something you were not testing.

What do you do when a customer refuses to give a number?

Ask what it would have to save them per month to be worth paying anything for. That produces a lower boundary grounded in their own operation, and in my experience it is often the more honest of their answers.

How do you handle bundles?

Components first, bundle second, with a removal question at the end. A bundle price heard early becomes the reference point for every component priced afterward.