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
An existing contract buys you access, a sponsor and a performance record. It does not buy you the agreement of people you have never met, and most expansion revenue goes missing in that gap.
The upsell that should have been easy is the one that teaches you the most. A client who has paid you for two years, renewed without argument and said in writing that the work is good, turns down a proposal to extend it. Nobody is unhappy. The sponsor still takes your calls. The proposal simply never comes back.
The usual reading is that the timing was wrong or the budget was spent. Sometimes that is true. More often the proposal was written as though the existing relationship had already done the work of getting agreement, and it had not. Access is not consensus, and the gap between the two is where most expansion revenue goes missing.
It feels like a continuation to you because you have been in the account for years. It is not one to the people who have to approve it. A new line of spend gets a new set of approvers, and the research on how those groups behave is not encouraging.
Gartner surveyed 632 B2B buyers between August and September 2024 and found that 74 per cent of buyer teams showed unhealthy conflict during the decision process. The same research puts buying groups at five to 16 people spanning as many as four functions, and found that groups which reached consensus were 2.5 times more likely to describe the deal as high quality. Read that last figure the right way round: consensus is not a nice outcome of a good deal, it is the thing that produces one.
None of that stops applying because you are already a supplier. The group is often larger on an expansion, because an existing contract has already taught finance, legal and whoever owns the budget that you are a line item they recognise. Being recognised is not the same as being wanted by all of them.
It buys three real things, and it is worth being precise, because overestimating them is the specific error here.
It buys you a meeting without a cold approach. It buys you a sponsor who will tell you the truth about timing if you ask a question they can answer. And it buys you a performance record, which is the single strongest asset in an expansion conversation and the one most sellers mention last.
What it does not buy you is the agreement of people you have never met. Your sponsor's goodwill does not travel into a budget meeting on its own, and a proposal written for your sponsor will not survive a room where three other functions are asking what this replaces.
Lumping them together is why expansion forecasts are so unreliable. They have different approvers and different odds.
| Type | What you are asking for | Who has to agree |
|---|---|---|
| More of the same | Extra volume of work already proven here | Usually the sponsor, inside an existing budget line |
| Adjacent scope | A service next to the one you deliver | The sponsor plus whoever currently owns that work |
| New category | Spend that did not exist in this account before | A full buying group, often including finance |
The first is a conversation. The third is a sale, and it should be resourced like one, with a business case and a timeline measured in quarters. Most disappointment comes from pricing and planning a new category as though it were more of the same.
The second row is where the hidden difficulty sits. Adjacent scope means taking work from someone, even if that someone is an overloaded internal team who would be glad to lose it. If you cannot name who currently owns the work you are proposing to take on, you are not ready to propose it.
Gartner's broader buying research makes the point that almost all B2B purchases are driven by organisational change rather than by a vendor's timing. That is as true inside an account as outside it. The expansion that lands is usually the one that arrives beside a new target, a new system, a departure, a reorganisation or a regulatory deadline.
So the practical work is not building a better deck. It is knowing early that something has changed. That means asking your sponsor what they are being measured on this year and what has moved since the last contract, and writing the answers down somewhere the whole team can see. Account notes that live in one person's head are the reason an obvious trigger gets noticed by a competitor first, and our six rules for CRM hygiene in a small sales team exist mostly to keep that from happening.
The quarterly review is the other place triggers surface, provided it is not a performance recital. We have written separately on running a quarterly business review that earns its hour, and the test is simple: if you leave knowing something you did not know about their year, it worked.
Your sponsor will have to argue for this without you. Most proposals are useless for that purpose because they are written to persuade the person who is already persuaded.
What travels is short and specific: what this costs, what it replaces or avoids, what happens if it waits a year, and the evidence from work you have already done here. One page. Written so your sponsor can forward it without adding a paragraph of explanation.
If the account already has somewhere that shows what has been delivered, use it rather than rebuilding the case from memory. Rivl's write up on scoping a client portal for a professional services firm is a good description of why that record is worth having before you need it, and the expansion conversation is exactly when you need it.
Expansion gets discussed through net revenue retention, and the benchmarks deserve less confidence than they are usually given. The published medians disagree with each other, they move substantially with deal size, and most of them come from software businesses with subscription mechanics that a services firm does not have. Importing one as a target is how a sensible team ends up chasing a number that was never about them.
The honest version is that your own trend is the only benchmark worth managing against, and you need several quarters of it before it means anything. If you want a comparison, compare this year's expansion in this account with last year's in the same account.
There is also a case for not upselling at all, and it gets skipped because it costs revenue to say. If the current work is not clearly succeeding, an expansion proposal reads as a supplier reaching for budget while the first job is unfinished. Fix the delivery, then ask. Our piece on where accounts really leave covers the other side of that, and the overlap is not an accident: an account you are about to lose and an account you are about to expand look similar from the outside and nothing alike from the inside.
If the commercial terms are the sticking point rather than the scope, that is a different conversation with different rules, and raising prices with existing clients without losing them is the one to read instead. Expanding what you do and charging more for what you already do are both asks, and running them in the same meeting usually loses both.
Tell us three things: what you deliver there today, what has changed in their business this year, and who signed the original contract. We will tell you which of the three upsell types you are actually in, and whether the trigger is there yet or you are a quarter early.
Book a Free Consultation