Pilot Purgatory

Running an accumulating book of enterprise pilots that never convert to paid contracts because the buyer is interested but has no internal urgency to sign.

The pilots are real. So is the interest. A large company agrees to a proof-of-concept, assigns a team, gives feedback, and tells the founder the product is impressive. The founder reads a growing list of logos-in-progress as a pipeline and tells investors the same. The problem is that a pilot isn’t a purchase, and a curious buyer isn’t a committed buyer. The startup keeps each pilot alive with integration work, custom features, and weekly check-ins. The contracts still don’t close. In diligence, the motion that looked like traction reads as activity with no revenue behind it.

Symptoms

The trap hides inside activity that looks healthy, which is what makes it dangerous. Watch for these signs together, not one at a time:

• The pilot count grows but the paid-customer count doesn’t. New proofs-of-concept keep starting while few or none become signed contracts. The top of the funnel is busy; the bottom is empty.
• Pilots have no end date or success criteria. Nobody agreed in advance what the pilot must prove, by when, or who signs if it succeeds. An open-ended pilot is designed never to end.
• You’re talking to a champion, not a buyer. The enthusiastic contact is a user or a mid-level sponsor with no budget authority. They love the product; they can’t purchase it, and you’ve never met the person who can.
• Engineering is building for pilots, not for the product. Each pilot demands a one-off integration or bespoke feature as the price of continuing. The roadmap quietly becomes custom work for companies that haven’t paid.
• There’s no compelling event. Nothing forces the buyer to decide: no deadline, no expiring contract with an incumbent, no regulatory date, no budget that lapses. Without one, “not now” is free and indefinite.

A single long pilot is normal in enterprise sales. The cluster is the trap: many pilots, no conversions, no criteria, no economic buyer.

Why It Happens

Pilot Purgatory is rarely a failure of product quality. It’s a failure to distinguish a buyer’s interest from a buyer’s commitment, and the structure of enterprise sales makes that distinction easy to miss.

The first cause is that a pilot is cheap for the buyer and expensive for the startup. A large company can run a proof-of-concept on a discretionary budget, with no procurement, no security review, and no board sign-off, because nothing has been purchased. For the startup, that same pilot consumes scarce engineering and founder time. The asymmetry lets the buyer stay in evaluation indefinitely while the startup funds the work.

The second cause is selling to a champion who isn’t the economic buyer. Enterprise purchases require someone with budget authority to decide, and often a procurement, security, and legal gauntlet behind them. A founder who has won over an enthusiastic user mistakes that win for the deal, when the champion’s job is to advocate, not to authorize. The deal stalls not because anyone said no, but because the person who could say yes was never in the room.

The third cause is the absence of a compelling event, the deadline or forcing function that turns “someday” into “by Q3.” Most enterprise buying that doesn’t close dies not as a loss to a competitor but as a “no decision.” The status quo wins because changing it was never urgent. Without a reason the buyer must act now, a pilot can run until the startup runs out of cash.

The fourth cause is incentive. Active pilots are a more comfortable story to tell a board than a short list of closed deals, and a pipeline slide full of recognizable enterprise names performs momentum the revenue doesn’t yet support. The founder optimizes for the easier metric, logos in motion, rather than the one diligence will test: conversions.

The Harm

The damage is that the company spends its scarcest resources manufacturing the appearance of traction instead of the substance of it.

The most direct harm is to runway. Every pilot consumes engineering hours, founder attention, and sometimes infrastructure cost, all charged against revenue that never arrives. A startup with ten active pilots and zero conversions is operating a services business it isn’t being paid for, and the cash burns while the contracts don’t close.

The second harm is a distorted product. When continuing each pilot depends on a bespoke integration or a custom feature, the roadmap fragments into one-off work for non-paying companies. The product that emerges is a patchwork shaped by evaluators rather than by customers, harder to sell to the next buyer because it was built to satisfy the last one.

The third harm is the one that ends companies: a fundraising story that collapses under diligence. The pilot count looks like a pipeline until an investor asks the only question that matters: how many pilots became paid contracts, and how long did each take? A wall of pilots with a near-zero conversion rate reads not as early traction but as evidence the product has interest without demand. The round that the pilots were supposed to justify becomes the round the pilots make impossible to raise.

The Way Out

The exit isn’t to stop running pilots. It’s to refuse to run a pilot that has no path to a signed contract, and to convert the pilot from a free evaluation into a gated step in a sale.

First, gate every pilot on a paid contract or a clear path to one. Before committing engineering time, name the conditions in writing: what the pilot must prove, by when, who signs if it succeeds, and what the contract is worth. A paid pilot is best, because it filters for buyers with budget and intent. At minimum, the pilot agreement should specify the deal that follows success, so a successful pilot converts by prior agreement rather than re-opening the whole question.

Second, find the economic buyer before you build anything. Identify who controls the budget and what their criteria are, and insist on access to them as a condition of the pilot. A champion is valuable as a guide to that person, not as a substitute for them. If you can’t get to the buyer, you aren’t in a sales process; you’re in an evaluation that has no decision-maker.

Third, qualify for a compelling event, and walk away when there isn’t one. Ask what forces this buyer to act on a timeline: a contract expiring, a regulatory deadline, a budget cycle, a board mandate, a cost they’re bleeding now. If the honest answer is “nothing; they’d like it eventually,” the deal won’t close on a schedule a startup can survive, and the disciplined move is to deprioritize it rather than fund an indefinite evaluation. The go-to-market motion that escapes the trap treats the pilot as one qualified step in a defined sale, not as the relationship itself.

How It Plays Out

A seed-stage enterprise software startup lands three name-brand pilots in its first year. The teams on the other side are engaged: they integrate the product, send detailed feedback, and ask for two custom features apiece as the price of continuing. The founder puts all three logos on the pipeline slide and raises a bridge on the strength of “enterprise traction.” Eighteen months later the picture is unchanged. The same three pilots are still running, still unconverted, the engineering team is half-consumed by bespoke integrations, and the champion at the largest account has changed jobs, taking the only internal advocate with them. When the founder goes out for a Series A, the first question is the pilot-to-paid conversion rate. The answer is zero, and the round doesn’t happen. The product worked the entire time. Nobody with a budget was ever asked to buy it.

The contrasting case is the team that refuses the open-ended pilot. Offered the same proof-of-concept, the founder asks who signs the contract if it succeeds, what it’s worth, and what the company is using today that this would replace. When a prospect can’t name a buyer or a forcing event, the founder declines the pilot and spends the capacity on prospects who can. Fewer pilots run, but the ones that do are paid, time-boxed, and pointed at a named signer. They convert. The pipeline slide is shorter and the revenue line is real, which is the trade a startup selling into the enterprise has to make to stay alive long enough to win.

Related Articles

Sources

• Tom Eisenmann, Why Startups Fail (2021): the Harvard Business School research on recurring startup failure archetypes, including the resource and demand-misread dynamics that strand companies in non-converting customer engagements.
• Mark Roberge, The Sales Acceleration Formula (2015): the framework for building a repeatable, qualified enterprise sales process, against which an ungated pilot book is the negative case.
• The concept of the “no decision” outcome and the requirement of an identified economic buyer and a compelling event are standard enterprise-sales qualification doctrine, codified in widely used methodologies (MEDDIC and its successors) that emerged from enterprise software selling in the 1990s.