How do you avoid a pilot that runs indefinitely?

How to answer this, stage by stage

Seven moves. Name what a fixed date can and can't buy you before the story, or the answer sounds like a scheduling tip instead of a decision.

Let's learn

What happens to a pilot that never gets a real end date?

Say a company builds a tool that reads every public post and comment about a brand and scores how likely each spike in complaints is to turn into a real story, before a journalist ever calls.

Before the pilot, Halfmoon's four-person communications team spent about ninety minutes every morning combing through saved searches by hand. They usually caught a real problem about a day after it started, once thirty or more comments had already piled up.

With Beaconwell running, the tool scored every public mention in real time and flagged a likely brewing problem about fourteen hours earlier on average. The morning scan dropped from ninety minutes to about ten. By week six, nobody on either side was arguing about whether the tool worked. It clearly did.

Here's the turn. Whether the tool worked was never really the open question. The open question, whether Halfmoon would actually buy, still hadn't been answered five months later, because nobody had ever written down when someone had to decide.

At its worst, that gap doesn't just waste time. Every week the pilot ran unresolved, it looked, from the outside, like an active enterprise deal in progress. Beaconwell's sales team quietly deprioritized two other prospective customers because "we're deep in a complex pilot already." The pilot cost real customers who were never part of the story.

The cost that line was quietly predicting only became visible once somebody finally added it up.

Four reasonable stamps on the left. Real customers and real hours quietly gone on the right. Nobody signed off on either.

What I'd leave alone. A free trial that technically expires on its own when the license runs out doesn't need this. Neither does a one-person internal test with no external customer waiting and no other prospect's time being spent instead. The ceremony only earns its place where a real decision, and real resources, are quietly waiting on somebody to say something.

The lesson. A pilot with no end date doesn't fail to reach a decision. It just spends month after month being decided by default, in the one direction that never needs anyone's signature: keep going.

Now here is the same thing as a story

Use this version when you've got a few minutes. The short version is above. This is for when the decision needs to survive more than one meeting.

For five months, Isabel's answer to "how's the Halfmoon pilot going" was always the same word: fine.

She'd run eleven pilots in three years at Beaconwell, and she was good at the parts everyone worries about, tuning a model's sensitivity, writing a rollout plan, keeping a prospect's engineering team from going quiet on her. Halfmoon was supposed to be an easy one. A four-person communications team, a clear use case, a VP of communications, Elian Farrant, who asked smart questions in the first meeting.

At kickoff, in a small conference room, Isabel and Elian set up a recurring thirty-minute calendar invite titled "Halfmoon x Beaconwell check-in." Elian asked the one real question anyone asked that day. "So what happens at the end of six weeks?" Isabel said, "let's run it about six weeks and see how it feels, we'll know when we know." Elian nodded. It sounded reasonable, even careful.

The first month was genuinely good. Full readouts every two weeks, Halfmoon's whole comms team in the room, real numbers on the screen. Then the readouts thinned to monthly, just Isabel and Elian. By week ten, the recurring invite mostly ran itself, three bullet points copied forward from the week before, nobody prepping anything new for it.

Week six arrived, the day the invite had quietly been standing in for a decision. Elian said, "we're still finalizing budget on our side, can we do one more month?" Isabel said sure. It felt like nothing.

Week ten. "We just had a leadership reshuffle, give us another month." Sure.

Week fourteen. "There's a product launch we need to get through first, let's revisit after." Sure.

Week eighteen. Nobody even framed it as a question anymore. The invite just kept recurring.

Week twenty-two. Beaconwell's new director of engineering, Booker Sennett, sat in his first resourcing meeting and asked a question that wasn't meant to be pointed. "Wait, is Halfmoon a customer or not?" Isabel opened her mouth to answer and realized she couldn't. Not with a date. Not with a name. Just "we're still in pilot," the same sentence she'd been saying since week six.

She pulled the numbers that afternoon. Forty hours had been budgeted for a six-week pilot. The actual number, by week twenty-two, was three hundred and ten. Two prospective customers had been quietly told Beaconwell's engineering team was tied up in "a complex enterprise pilot" during exactly those months.

Two months earlier, at kickoff, Elian had actually asked the right question. "So what happens at the end of six weeks?" Isabel remembers answering, "let's run it about six weeks and see how it feels, we'll know when we know," and the room nodded, because it sounded reasonable, even generous. Nobody circled back to pin it down. It just quietly became the plan.

I would take that back. I'd have said, in that same kickoff meeting, "before we start the clock, I want a real date on the calendar, tied to your next budget cycle, and I want to know now who on your side signs off go or no-go when that date arrives. If we need more time later, that's a new, dated ask, not a default."

Here's the replay. Same six-week plan, but this time the recurring invite carries an actual date tied to Halfmoon's quarterly budget lock, with Elian named as the sign-off owner on Halfmoon's side and Isabel's VP of sales named on Beaconwell's. Week six arrives, and the invite that day isn't a check-in. It's a decision meeting. Elian either signs, or she doesn't, or she asks for a new, dated extension in writing. The total engineering cost stays close to the forty hours planned, because nobody quietly refills the tank four separate times without anyone noticing.

One design let the pilot answer for itself by never ending. The other made a person answer, on a date they'd agreed to six weeks in advance.

And the thing I'd tell myself, back in that kickoff room: "we'll know when we know" isn't patience. It's agreeing in advance that nobody has to decide.

LEAD, so a pilot can't just quietly keep going

This sounds like a question about scheduling. Underneath, it's still asking what makes "stop" cost something. That's LEAD, run on a pilot's own calendar instead of a dashboard.

And if you want to be sure it really works, try it somewhere else

Featherstone Growers Collective is a cooperative of about forty small farms in one growing region. Behrouz Obuya leads product for Halversen, a tool that reads field images and sensor feeds to flag crop-disease risk early enough for a grower to spray before an outbreak spreads. Featherstone agreed to an eight-week pilot across the region.

Featherstone's pilot was budgeted for twenty-five agronomist hours across eight weeks. By the time the growing season ended, twenty-six weeks in, it had cost a hundred and forty hours, and there was still no decision either way. Not a yes. Not a no. The co-op's calls just got shorter and further apart until they stopped.

Waiting for a co-op to go quiet never rings at all. A pre-agreed decision date rings on schedule, whatever the answer turns out to be.

Flashcards (click a card to flip it)

Check yourself Score: 0 / 0