What a discovery phase actually costs, and what the fee has to buy
What you are buying with a discovery fee
A discovery phase buys an implementation estimate you can hold a vendor to, produced while you can still walk away. For mid-market work, expect a fixed fee between two and five percent of the implementation budget you think you are facing.
The fee pays for conversion work. An RFP answer is a bid written from the outside in a few days. Discovery turns it into a bottom-up plan: the current state mapped, requirements ranked by the people who will live with the result, every integration counted, and a cost model whose assumptions are written down where you can attack them. Across the 148 discovery engagements ACME Inc ran between 2021 and 2025, our own engagement data puts the median at 42,000 euro over four calendar weeks (acme-consulting.example). The smallest was 18,000 euro for a single-system replacement, the largest 76,000 euro for a group-wide ERP scoping.
That leaves one useful question, and it is the one this article answers: what does the money decompose into, and how do you know whether the number in front of you is honest?
The arithmetic behind the fee
A discovery fee is people times days times day rate, and all three factors are checkable before you sign.
Start at the bottom. The floor under any day rate is labour cost, and labour cost is public. Eurostat puts the average hourly labour cost across the EU at 33.5 euro in 2024, ranging from 10.6 euro in Bulgaria to 55.2 euro in Luxembourg (ec.europa.eu). A senior consultant billed at 1,800 euro a day is not being paid 1,800 euro a day. Between the labour cost and the invoice sit the firm's overhead, its margin, and above all its utilisation planning: at a planned senior utilisation of 70 percent, which is what we plan at ACME, the billed day also carries the three days in ten that nobody buys.
Then the shape of the team. Our median discovery is 30 person-days at a blended day rate of 1,400 euro: a lead consultant full time, a solution architect at half time, four to five calendar weeks. When a proposal quotes 60 person-days for a mid-market scoping, the question is not whether the firm is expensive. The question is which deliverable doubled the days, and the answer is usually a workshop programme that belongs in the implementation.
The honest cost drivers are countable: systems touched, integrations crossed, stakeholder groups that have to be heard, sites that have to be visited. A fee that moves without one of those moving is a fee drifting toward the implementation budget.
The cost of skipping it
Not paying for discovery does not remove the cost. It moves the cost to the other side of the signature, where it is larger and arrives as change requests with your name on them.
The industry-level numbers are not kind. The Project Management Institute's Pulse of the Profession research put the share of investment wasted through poor project performance at 11.4 percent (pmi.org). On a 1.5 million euro implementation that is 171,000 euro, roughly four discovery phases burned on rework, idle licences and revisited decisions. The same research programme found that organisations which undervalue project management as a discipline report 67 percent more of their projects failing outright (pmi.org).
Our own engagement data says the same thing in a smaller mirror. Implementations that followed a paid discovery closed within a median of 9 percent of the post-discovery estimate. Implementations sold straight off an RFP response overran their contracted value by a median of 24 percent, and their change request logs ran roughly three times longer (acme-consulting.example). The difference is not that discovery makes delivery cheap. It makes the argument happen early, on paper, while the argument is still cheap to have.
What the fee has to buy: deliverables and exit criteria
A discovery phase ends in a go or no-go recommendation you could act on with a different firm. Everything short of that was a workshop with a nicer name.
The deliverable list is short and it is not negotiable: a current-state map, a requirements backlog ranked by the people who will use the system, an integration inventory, a costed delivery plan with its assumptions written down, a risk register with named owners, and the recommendation itself. Each one belongs to you. The test of a discovery deliverable is whether it survives being handed to a competitor, because sometimes it is.
Reader question · answered by the byline
If we commit to the implementation up front, should the discovery be free?
"No, and treat the offer as information," says Marta Ellingsen, partner at ACME Inc. "A free discovery is priced somewhere, and the somewhere is the implementation bid you have not seen yet. When you pay for discovery, the fee is the whole of our interest in the answer, which is what lets the answer be no-go. Roughly a third of our discovery clients take the plan and implement with someone else. That is the product working, not the product failing."
Marta Ellingsen, Partner, ACME Inc. Answered by SMS during research and published under her byline.
The exit criteria matter as much as the artefacts. Each deliverable needs a date and a named acceptance test, and the phase needs a defined answer to the only question it exists to settle: build, buy, phase it differently, or stop.
Fixed fee or time and materials for the discovery itself
Buy discovery at a fixed fee against named deliverables. Save time and materials for the phase after, where the unknowns are the point.
Discovery is the rare consulting phase whose scope is known before it starts, because the deliverable list above does not change from client to client, only the depth does. That makes it priceable, and a firm that will not fix the price of its own scoping method is telling you something about the method. The statement of work should name the deliverables, the acceptance test for each, the team by name, and what happens to the fee if the recommendation is no-go. It should also say, in one sentence, that you own the outputs.
Two clauses are worth negotiating rather than assuming. First, a partial credit of the discovery fee against phase two is fine to accept and wrong to require, because requiring it quietly converts the discovery back into a sales cost. Second, rate transparency: the blended day rate and the person-day count belong in the statement of work, not just the total, so that when scope moves you can see which factor moved. Neither clause costs a serious firm anything. Both are cheap tests of whether you are buying analysis or feeding a pipeline.
Paid discovery and free vendor scoping, side by side
Read the comparison as a set of incentives rather than a ranking: the free option is not cheaper, it is paid for differently.
Row by row, this is the honest version of the difference:
| Question | Paid discovery | Free vendor scoping |
|---|---|---|
| Who pays for the work | You do, at a fixed fee | The vendor, priced into the later bid |
| Who owns the deliverables | You, reusable in any RFP | Often the vendor, tied to their proposal |
| Quality of the estimate | Bottom-up, tied to exit criteria | Top-down, anchored to winning the deal |
| Incentive of the author | Accuracy, the fee does not depend on phase two | Volume, scoping feeds the sales pipeline |
| When the numbers move | Before the contract, on paper | After signature, as change requests |
| Most common failure | Scope creep inside discovery itself | An estimate built to be low |
FAQ
How long should a discovery phase take?
Two to six weeks for most mid-market engagements; ACME Inc's median is four. Much shorter and you have bought workshops with a nicer name. Much longer and the implementation has quietly started without a contract that says so, which is the worst of both pricing models.
Is a percentage of the implementation budget the right way to price discovery?
As a sanity check, yes: two to five percent of the expected implementation spend is the band most mid-market work lands in. As a pricing mechanism, no. Buy it as a fixed fee against named deliverables, because a percentage gives the author of the estimate a share in the size of the estimate.
Can we run discovery with one firm and implement with another?
Yes, provided the statement of work says you own the deliverables and the documents are written to survive the handover. In ACME Inc's engagement data roughly a third of discovery clients implement elsewhere. A firm that resists that clause is pricing its discovery as a sales cost, and you have learned that before signing anything.