Customer discovery interviews, and how many are enough
B2B customer discovery interviews: the sample size the research converges on, why your segments change that number, and the questions that get honest answers.
Read
The fee does not make a buyer commit. A field experiment that separated the two effects found strong screening and no consistent sunk cost, which puts the price in your qualification process rather than your onboarding.
The question gets asked as though it were about money, and it is not. A pilot fee of five thousand dollars does not change your revenue for the year. What it changes is which buyers sign, and that is a different mechanism with a different payoff, and most of the advice on this subject has the mechanism wrong.
The common claim is that charging makes the buyer committed. They have skin in the game, so they will push the rollout through their own organisation. It is a satisfying story. It is also the part of the argument with the weakest evidence behind it, and there is a careful field experiment that separates the two effects and finds only one of them.
Nava Ashraf, James Berry and Jesse Shapiro set out to answer a question with the same shape as this one: can a higher price increase how much a product gets used. They identified two distinct ways it might, and they are worth naming because the whole pilot debate collapses them into one.
A screening effect means the price changes who buys. The people willing to pay more are the people who were going to use it more anyway, so the price works as a filter and the average buyer you end up with is better.
A sunk-cost effect means the price changes the behaviour of a given buyer. Having paid, they feel the need to justify the spend, so they use the thing more than they otherwise would.
Those predict the same surface result, higher price and higher use, through completely different routes. Telling them apart needs a design where the amount offered and the amount actually paid can move independently, and that is what the authors built. Their paper is Can Higher Prices Stimulate Product Use? Evidence from a Field Experiment in Zambia, issued as an NBER working paper and later published in the American Economic Review.
The intervention was a door-to-door sale of Clorin, a household water purification solution, to about 1,000 households in Lusaka. Each household was offered one bottle at a one-time, randomly chosen offer price at or below the prevailing retail price. Households that agreed then received an unanticipated, randomly chosen discount, which let the authors vary the transaction price separately from the offer price.
About two weeks later they went back, asked about use, and chemically tested the household's stored water for the presence of Clorin. That last detail is what makes the study worth reading: the outcome is a chemical test, not a self-report about whether someone meant to use it.
So varying the offer price while holding the transaction price constant tests screening. Varying the transaction price while holding the offer price constant tests sunk cost. The surprise discount is the whole trick.
On screening, the result is strong. Holding the transaction price constant, the households that agreed to a higher offer price were, in the authors' words, statistically and economically more likely to use Clorin in their drinking water at follow-up. Willingness to pay predicted use. It held up after conditioning on household characteristics, and the authors note that willingness to pay was more predictive of use than an optimal linear combination of everything they could observe at baseline.
On sunk cost, there is nothing. They found no evidence that households paying a higher transaction price were more likely to use the product, and some point estimates pointed the other way. Their confidence intervals were tight enough to rule out effects of roughly the same order of magnitude as the screening effect they did measure. It held even among households that displayed sunk-cost reasoning in hypothetical scenarios.
| Claim about pilot pricing | Which effect it assumes | What the experiment found |
|---|---|---|
| A fee makes the buyer push the rollout internally | Sunk cost | No consistent evidence |
| Charging filters out buyers who were never going to adopt | Screening | Strong evidence |
| Willingness to pay tells you something you cannot otherwise see | Screening | Supported, and more predictive than observables |
| Paying something beats paying nothing | Either | Could not rule out no effect |
That last row matters for the specific question in the title, and it is the row most likely to be skipped. The authors tested, at practitioners' suggestion, whether paying something produces more use than paying nothing. They could not rule out no effect there either, although they note the sign and magnitude were at least consistent with the hypothesis. So the paid-versus-free comparison is the weakest part of the evidence, not the strongest.
This is a consumer health product in Zambia, priced in cents, bought by a household in a doorway conversation. Your pilot is enterprise software bought by a committee over eleven weeks. Nobody has run this experiment on B2B pilots, and the honest position is that the mechanism generalises better than the magnitude.
Why the mechanism should carry: the screening effect works because willingness to pay encodes private information that the seller cannot observe. That is more true in B2B, not less. A buyer who will put four thousand dollars through procurement for a ninety day pilot is telling you something about internal sponsorship, budget authority and urgency that no discovery call reliably extracts. The signal is the procurement process itself as much as the money.
Why the sunk-cost absence should also carry: the organisations that abandon pilots rarely do so because they stopped caring about the money. They do so because the sponsor changed role, a reorganisation landed, or the project was never resourced on their side. None of those respond to having paid a fee.
If the fee is a filter rather than a motivator, you should design it as a filter and stop expecting it to do work it does not do.
Price the pilot to be refused. A fee low enough that anyone says yes screens nobody. The number should be high enough that a buyer without real sponsorship declines, which is the outcome you want, early and cheaply. Compare it against what you already know about how much a late disqualification costs you.
Do not treat the fee as adoption insurance. If you were relying on the invoice to make the buyer show up to the kickoff, you need the thing that actually does that, which is a named owner and an agreed sequence on their side. That is the job of a mutual action plan, not of a price.
Watch who signs the cheque, not the amount. The screening signal is strongest when the money comes from the budget of the person who owns the problem. A pilot paid out of an innovation fund tells you much less, because nobody in the building feels it. This is the same read you apply to who is actually in the room.
Expect procurement to be the real cost. In the Gulf especially, getting a small pilot through vendor registration and payment approval can take longer than the pilot. If you make it paid, you have inserted that process before any value is proven, which is sometimes the correct trade and sometimes fatal. The timings are set out in payment terms and the real clock.
Three cases, and they are not edge cases.
You have no reference customers. Screening is a luxury of a seller with more demand than capacity. If you need three logos before anyone will take a meeting, the filter is working against you and you should want the loosest possible qualification.
The same question has been worked through in a quite different market, and it is useful precisely because the stakes are personal rather than corporate: UCourses compared what the money really buys when a course is paid rather than free, and lands in a similar place, that the fee changes who turns up more than it changes what they do once they are there.
The buyer is a design partner. If you are getting roadmap input, access to their data model, or the right to publish the case study, they are paying in a currency that is worth more than the fee. Charging as well is confused.
The pilot is really a demo. If nothing in it requires their time, their data or their people, there is no adoption to screen for and the fee is just friction.
We cannot tell you what to charge. The specific figures circulating for pilot pricing come from vendor blogs reporting their own practice, and we could not trace any of them to a dataset. Treat them as anecdote.
We also cannot tell you that a paid pilot converts better than a free one, and neither can anyone else from evidence. That comparison, paying something against paying nothing, is precisely where the experiment above returned a result it could not distinguish from zero. Anyone quoting you a conversion rate for paid against free pilots is quoting their own book.
What is solid enough to build on: a price selects buyers, and selection on willingness to pay carries information that your qualification questions do not capture. A price does not make a buyer adopt. Those two together say the fee belongs in your qualification process rather than in your onboarding, and that is a different place from where most teams put it.
Send us the pilot scope you are offering now, who pays for it on the buyer side, and your last ten pilot outcomes. We will come back on whether your fee is screening anyone and what it is quietly costing you in qualification you are doing later and more expensively.
Book a Free Consultation