Demo Teardown: The Quiet One in the Room Who Killed It Three Days Later

Six names on the invite. Six squares on the grid. Five of them talked, laughed at the right moments, asked good questions, and stayed two minutes past the hour. One of them said hello at the start, said "no, I'm good" when asked if he had questions at the end, and otherwise sat there with his camera on and his face still.

The deal died the following Tuesday. Not a loud death. An email with the word "internally" in it.

What follows is a teardown of that demo. It's a composite — the specific call I have in mind blends with a dozen others I've sat on and coached, because the pattern repeats so exactly that the details stop mattering. If you run multi-stakeholder demos, you have already lost a deal this way. You may not know it yet, because this failure mode is invisible from inside the call. The call feels great. That's the whole problem.

The call

Software buy, mid-market, six attendees. A VP who owned the budget and did most of the talking. Two managers from the team that would use the thing daily. An ops person who asked sharp questions about reporting. A finance-adjacent person who joined late and asked about billing cadence. And one man from IT who was listed on the invite as "engineering" and never spoke after the first minute.

The rep ran a clean demo. Genuinely clean. She opened with a recap of the discovery call, confirmed the agenda, ran a workflow start to finish instead of clicking around a feature tour, handled the reporting questions without deflecting, and closed with clear next steps that the VP agreed to on the call. If you graded that demo on execution, she'd score well.

She optimised for the loudest voice in the room. The VP was engaged, the VP was senior, the VP kept saying things like "this is exactly the problem we have." So the rep leaned in. Every time the VP spoke, the rep followed. Every time the VP nodded, the rep moved forward. By minute thirty the demo had become a two-person conversation with four spectators.

And the silent one owned the integration. The whole business case rested on the product pulling records out of a system nobody in the room had looked at in eighteen months. The VP assumed it was a connector. The managers assumed the VP had checked. The rep assumed that if it were a problem, someone would have said so.

The engineer knew it was a problem. He knew there was no clean API, that the last vendor who tried it had built a nightly file drop, and that the team who maintained that system had a freeze on new integrations until a migration finished. He didn't say any of it. He wasn't being difficult. He was outranked, he'd been added to the invite that morning with no context, and he'd learned that raising an objection in front of a VP in a vendor meeting is a good way to become the person who kills things.

So he waited. Three days later, in a Slack thread the rep will never see, he wrote two sentences. Those two sentences ended the deal.

The loudest-voice tax

Here is the thing that makes this hard. Optimising for the loudest voice is not a mistake in the moment. It feels like good selling. The person talking is giving you signal, they're senior, they're enthusiastic, and every instinct you have says follow the energy.

But enthusiasm from the economic buyer is the cheapest signal in the room. It costs them nothing. They are not the one who has to build the thing, maintain the thing, or explain to their own boss why the migration slipped. The VP on that call was sincere. He wanted it. He just didn't have the information that would have told him he couldn't have it.

The people who hold the deal-killing information are almost never the loudest. They are the ones who will have to do the work. And in most organisations, the person who has to do the work has learned that meetings are not where you say no. Email is where you say no. Slack is where you say no. The hallway after is where you say no.

Your job in a multi-stakeholder demo is to move that no forward, into the call, where you can still do something about it. That is the entire discipline. Not to avoid objections. To relocate them.

Reading the silence

Before you can act on a quiet attendee you have to work out what kind of quiet you're looking at, because the response differs completely. In my experience there are four, and they are distinguishable if you're watching.

Bored

Bored looks like a person doing something else. Eyes tracking sideways to a second monitor. Typing that doesn't correspond to note-taking rhythm. Camera off early. Bored is usually someone who was invited out of politeness, has no stake, and is waiting for the call to end so they can get back to their actual day.

Bored is the least dangerous silence and the one reps worry about most. It won't kill your deal. Don't spend the room's time trying to entertain it. If you're confident someone is genuinely peripheral, let them be peripheral and give them permission to drop — "Dave, I know a lot of this is downstream of you, feel free to jump if you need to" buys you enormous goodwill with the rest of the room and costs you nothing.

Outranked

Outranked looks like attention without contribution. Camera on, eyes on the screen, occasional nods, no words. Sometimes a hand-half-raise that gets abandoned when someone senior starts talking. This person is engaged and has opinions and has decided that this room is not where the opinions go.

This is the dangerous one. It was the engineer in our teardown. Outranked silence is loaded silence.

Sceptical

Sceptical is different from outranked because sceptical usually does talk — just not much, and always in questions that have a floor under them. "How does that behave when the record is incomplete?" Then nothing for ten minutes. Then another one. Sceptical people are testing you and keeping score, and they will keep the score to themselves until they've decided.

Sceptical is workable. It's the posture I see most often in technical and regulated buying rooms, and it's why the freight and 3PL demo script is built around a room that has arrived specifically looking for where the product breaks. When someone is sceptical, giving them a real answer — including "it doesn't handle that well, here's what people do instead" — converts them faster than anything else you can do.

Blindsided

Blindsided is the rarest and the easiest to miss. This is the person who didn't know what the meeting was about, is hearing for the first time that their team's workflow is being replaced, and is processing that in real time while you talk about time savings. They go very still. They sometimes go pale. They will ask one question near the end that sounds administrative and is actually a boundary test — "so who would be administering this?"

Blindsided silence turns into a detractor within a week unless you address it on the call. Nobody supports a change they found out about from a vendor.

Three things to do mid-demo

The fix for all of this is not a better demo. It's structural. Three moves, and you can run all three inside a standard hour without losing time on the product.

The round-the-room open

Spend the first four minutes going around every attendee by name and asking what they want to get out of the call. Not "any questions before we start" — that gets you nothing. Named, individual, one at a time.

"Priya, I know from our last call you're focused on the reporting side. Anything specific you want to make sure we hit? … Marcus, I've got you down as engineering — what would you need to see to feel good about this?"

Two things happen. First, you get an early read on every voice in the room, which means the quiet one at minute forty is a deviation you can notice rather than a baseline you never established. Second, and this matters more, you have publicly established that this person is expected to speak. You've given them a mandate in front of their VP. When you come back to them later, you're not putting them on the spot, you're following up on something they were already licensed to contribute.

The cost is four minutes. In a room with real seniority spread, it's the best four minutes you will spend. I've written a longer version of this for legal buying committees in the managing partner and risk committee demo script, where the seniority gap in the room is so wide that skipping the round-the-room guarantees you only hear from one person.

The name-and-ask at minute twelve

Around the twelve-minute mark — early enough that there's time to change course, late enough that you've shown something real — stop and call on the quietest person by name, with a specific question tied to their domain.

Not "Marcus, any thoughts?" That's a yes/no dressed as a question and you will get "no, looks good." You want something he cannot answer with a nod.

"Marcus, this part depends on us reading from your records system. Walk me through what that system actually is on your side and who owns it."

That's the question that would have saved this deal. It's not a demo question. It's a discovery question you're running mid-demo because you didn't get access to this person before the call. It hands him the floor on a topic where he is unambiguously the expert in the room, which lets him talk without appearing to challenge anyone. He gets to be helpful rather than obstructive. That reframing is everything for an outranked attendee.

Minute twelve is not magic, but early is. If you find the integration problem at minute twelve you can spend the rest of the hour on it, bring in a solutions engineer, or reset the next step honestly. If you find it at minute fifty-eight you have found it after the VP has already committed enthusiasm in front of his team, and now somebody has to walk that back. They will do it on Tuesday, by email, without you.

"What would make this a bad idea for your team?"

This is the one I'd keep if I could only keep one. Ask it directly, to a named person, in those words or close to them.

"Marcus — what would make this a bad idea for your team specifically?"

The question does something structurally clever. It gives explicit permission to object, and it scopes the objection to their team, which removes the risk of appearing to overrule the VP. He isn't saying the project is bad. He's answering a question about local constraints, which is his job. Most people, given that opening, will take it.

The first answer is often soft. "Probably just timing." Sit in it. Don't rescue them. "Timing how?" And then it comes out — the freeze, the migration, the team that owns the source system and hasn't been told.

Buyers who are looking for failure modes respond to this better than anyone, which is why it's central to how I'd approach a claims room in the insurance demo script. Technical and operational evaluators trust a vendor who asks where it breaks far more than one who insists it doesn't.

The four-minute pre-demo attendee map

Everything above is recovery. Here's prevention, and it takes four minutes the day before.

Open the calendar invite. For every single attendee, write three things:

What do they own? Not their title — what system, budget, team, or process do they control that this deal touches. If you can't answer this for someone, that is the finding. That's your first question in the round-the-room.

What do they lose if this goes ahead? Everyone in a buying room loses something. Time during implementation. A tool they liked. Control over a process. Headcount justification. Write the actual loss. If you write "nothing," you haven't thought about it hard enough.

Can they say no, and how? Some people can veto directly. Some can veto by taking eight weeks to respond to a security questionnaire. Some can veto with two sentences in a Slack thread three days after the call. Mark which kind each person is.

Run that on our six. The VP owns budget, loses nothing, can say yes directly. The two managers own the daily workflow, lose their current process, can slow-walk adoption. The ops person owns reporting, loses her spreadsheet, has real influence. Finance owns the cadence, loses very little, can delay. And Marcus — you'd have written "owns: unclear, listed as engineering." Unclear is a red flag. Unclear means you're walking into a room with an unmapped person who was added for a reason nobody told you.

The fix from there is a ten-minute call with him before the demo, or an email that says: Marcus, I saw you've been added to Thursday — I don't want to waste your time on parts that aren't relevant. What's your involvement, and is there anything on the integration side I should be ready for?

He would have told you. People tell you things in writing that they will not say in front of a VP. That email is the cheapest deal insurance in the job.

What I'd do next

If I were running this back, I wouldn't read about it, I'd rehearse it. The round-the-room open and the name-and-ask both feel intrusive the first time you do them live, and the discomfort is exactly what makes reps skip them and default to following the loudest voice. That's a reps-under-pressure problem, not a knowledge problem. It's the kind of thing we built DrillCall for — running the multi-stakeholder room over and over against AI buyers who behave like real committees, including the one who sits there and says nothing until you go and get him. Do it twenty times in practice and the twelve-minute pivot stops feeling rude and starts feeling like the demo.

The deals you lose to a competitor, you usually hear about. The deals you lose to the quiet one, you just get an email on Tuesday. Go find him before Tuesday.