Chapter 8: When the Check Fails: A Post-Mortem

Chapter 8 · Part Two: The Comfort Check

If I only show you the ones that worked, I’m not teaching you anything. I’m doing a victory lap in front of you and calling it a book.

So here’s one from the same company, roughly the same period, same guy, same budget authority. It cost real money and it didn’t work at all.

What makes it worth a chapter isn’t that it failed. Things fail. It’s that it failed for the exact opposite of the reason everybody assumes, including me at the time, and the correct diagnosis took me years to arrive at.


What I wanted

More demos at the booth. More qualified people sitting down in front of the product.

That’s a completely ordinary objective and I want to note, before anything else, that it passes the test in Chapter 6 only barely. “More qualified demos” is a measure of the top of something. It’s closer to an activity than I’d write today. Hold that thought, because it comes back at the end.

What we landed

Then something genuinely good happened. We landed the customer I had wanted for years.

A car company. I’m not going to name them, which, as you’ll see, is the only part of this that worked.

The idea

Every person who came to our booth and sat through a real demo got one entry into a sweepstakes. The prize was the car.

The plan, as it looked in the deck.
The plan, as it looked in the deck.

And here’s the part I didn’t tell anybody outside the company, because this was the part I loved. At the end, we’d draw the winner live, on a webinar, from that company’s headquarters in Germany.

Which meant the drawing was also, quietly, the best customer testimonial we’d ever get. A global manufacturer, on camera, at their own facility, participating in our campaign. Not a quote in a press release. A thing happening.

I thought it was the most elegant idea I had ever had. I’m aware of how this sentence reads given where the chapter is going.


Ed

Before we spent a dollar, we hired a lawyer named Ed.

Ed does one thing. Ed does sweepstakes. That’s his entire practice, which I hadn’t known was a viable way to make a living, and which should have told me something about how well-trodden this ground was.

In the first conversation, before I had finished explaining it, Ed said: you have to offer a cash alternative, that part isn’t optional, and if you offer cash they’ll take the cash.

They always take the cash.

I heard him. I want to be completely clear that I heard him, because the version of this story where I didn’t listen is more flattering and less useful. I heard the man whose entire professional life is this specific thing tell me the specific thing that was going to happen.

And I decided he didn’t understand how good my idea was.


What happened

We launched.

About a week later, the car company came back and said they couldn’t be associated with it. After they had agreed to it. In writing.

So now there’s no reveal. No webinar. No Germany. No testimonial. It’s now a car giveaway at a trade show, which, if you have ever walked through a shopping mall, you’ve seen before.

What it became.
What it became.

We ran it anyway, because the money was already gone.

The winner took the cash.

And never became a customer.


The easy read, which is wrong

The easy read is: he did something weird and it didn’t work.

I want to close that off hard, because if you take that lesson out of this chapter you’ll have learned the opposite of the thing, and you’ll be more careful for the next three years in exactly the way that costs you the most.

It didn’t fail because it was weird. It failed because it wasn’t weird.

A car giveaway is the most familiar promotion in the history of commerce. It’s what a car dealership does. It’s what a mall does on a Saturday. It’s what a radio station did in 1974. The prize was larger than usual and that’s the entire extent of the innovation.

The one genuinely different thing in the whole plan was the live draw from their headquarters. And that element was invisible to everybody except me. It existed in a slide. In reality it was going to exist for one webinar, at the very end, in front of people who had already entered.

It was different in the deck. It was never different in practice.

That’s the sentence I wish somebody had said to me, and it’s the visibility test from Chapter 7, which I built afterward, out of this.


The nod, with a body attached

Now the part that connects to where this book started.

When I pitched it internally, everybody nodded. Immediately.

Of course they did. Everyone understood it in one second, because everyone has seen a car giveaway. There was no explaining to do. The concept required zero cognitive effort from any person in that room, and the room experienced that effortlessness as clarity, and reported it back to me as agreement.

The nod wasn’t telling me the idea was good. The nod was telling me the idea was familiar.

It was the warning label and I read it as a review.

One person in the entire process didn’t nod. Ed.

And I discounted Ed specifically because everybody else was nodding. That’s the trap from Chapter 1 and I walked into it with my eyes open. The volume of agreement in the room made the single dissenting voice sound like an outlier rather than a signal. A room full of people who had never run a sweepstakes outvoted one man who had run hundreds, and I let them, because they were many and he was one.

He votes twice

Years after the car, I watched a clip of Bill Belichick making a point I’d heard a hundred times and never really taken in. His point, roughly: you can’t improve if you can’t accept constructive criticism.

Every coach says some version of that, and everybody who hears it agrees, which by now should make you suspicious of how much work the agreement is doing.

The usual reading is that criticism is hard because it stings. Somebody is blunt, your ego takes a hit, you get defensive. That happens, but it isn’t what happened with Ed. Ed was polite. He was specific. He was right. There was nothing in the delivery to resent.

It was hard to take because it felt wrong, and it felt wrong for exactly the same reason the giveaway felt right.

Go back to Chapter 1. Processing fluency makes the familiar idea feel correct. Run the same machinery the other way and it makes the unfamiliar objection feel like friction, and friction feels like a mistake on the other person’s part. So he votes twice: once for the idea everybody recognizes, and once against the person who doesn’t. Neither vote feels like a vote. Both feel like judgment.

That’s why “be open to feedback” is useless as advice. Nobody believes they’re closed to it. What you can actually do is notice the specific sensation, the small sense that an objection is beside the point, and treat that sensation as the thing to investigate rather than the conclusion.

It’s also why the rotation rule in Chapter 1 comes with a warning. Nemeth’s research says the real thing works and the role-play mostly doesn’t. You can’t fake Ed.

If the only feedback you can hear is the kind that agrees with you, you aren’t getting feedback. You’re getting a nod.

If you take nothing else from my very expensive afternoon:

The room agreeing with you isn’t the signal. The one person who doesn’t is.


The mechanism, which isn’t the car company

Here’s where most post-mortems stop, and where this one gets useful.

The obvious story is that a partner pulled out and wrecked the plan. Bad luck. Unforeseeable. We had it in writing.

That story is true and it’s a decoy, because it lets the whole thing be somebody else’s fault, which is exactly what a post-mortem is supposed to prevent.

So run the counterfactual honestly. Suppose the car company had held. Suppose we got the webinar, the headquarters, the live draw, the testimonial, all of it.

The campaign still fails.

Because the mechanic was always going to select the wrong people. A prize doesn’t select for people who want your product. It selects for people who want the prize. Everybody who walked up to that booth and sat through a demo to get an entry was, by construction, someone whose reason for being there was the car. We built a machine for filling a room with people who wanted a car, and then we were disappointed that it filled the room with people who wanted a car.

The partner falling through cost us the one nice element. The design error cost us the campaign, and the design error was present on day one, in the pitch that everybody nodded at.

The same error, in cheaper clothes

You’re probably not giving away a car. You may well be doing one of these.

  1. The gift card for a thirty-minute call. Selects for people who want a hundred dollars. Reliably produces meetings, reliably produces nothing after them, and reliably survives in the plan because meetings booked is a number that goes up.
  2. The high-value conference raffle. Selects for badge scans. Your scan count will be the best in the hall and your pipeline won’t notice.
  3. The gated report promoted to a cold list. Selects for people willing to trade an email address for a PDF, which is a genuinely different population from people who have the problem the report is about. You didn’t generate demand. You purchased a list at an unusual price.
  4. Swag that’s better than your product story. The line at your booth is measuring the socks.

The general form, and the question to run on any incentive before you launch it:

Every incentive selects. What does mine select for, and is that the same population as my buyer?

If the answer is no, you aren’t accelerating demand. You’re paying to contaminate your own data, and the contamination will look like success for two quarters.


The check, with question one blank

This is the only Comfort Check in this book where the first line doesn’t get crossed out.

I never crossed it out. I never asked the question. The assumption sat there the entire time, unexamined, load-bearing, and completely invisible to me because it arrived wearing the clothes of an objective.


Failure is where we learn. From my 2022 CyberMarketingCon keynote.

The Post-Mortem Template

What I use now. Seven questions, run in writing, before the meeting rather than during it, and with rule five taken seriously or the whole thing is theater.

  1. What did we predict, in writing, before we started? If there’s no written prediction, stop. You aren’t running a post-mortem, you’re constructing a narrative, and the narrative will be built out of whatever everybody remembers believing, which is always roughly what happened.
  2. What actually happened? Facts and dates only. No causes yet.
  3. Where did one and two first diverge? Find the earliest date, not the most dramatic one. The dramatic moment is almost never the divergence, it’s the consequence becoming visible.
  4. What did we know at that point that we didn’t act on?
  5. Who said it, and why did we discount them? Name the person. This question is the whole template. Every failed project I have ever examined had somebody who said the thing, and the interesting information is never what they said. It’s the reason we found for not listening, because that reason is a property of the organization and it will be there next time. In my experience the reason has a shape, and it’s usually the one from earlier in this chapter: the objection felt like friction, so it felt like the other person’s mistake.
  6. What was the mechanism, as distinct from the event? The event is what you’ll tell people at the bar. The mechanism is the thing that would still have been true if the event had gone your way. If you can’t separate these, you haven’t learned anything you can carry.
  7. What would we have had to believe, on day one, to make the right call? Not what should we have known. What belief would have produced the right decision. That belief is the thing to install, and it’s usually one sentence.

For this project, question seven has an answer I use constantly now: an idea everybody understands instantly is an idea everybody has already seen.


The Comfort Check

Chapter 8: the dream car giveaway

1. The comfortable assumption. A big enough prize gets us the right people.

This one isn’t crossed out, because I never crossed it out. I never asked.

2. The true objective. Qualified demos that turn into pipeline.

3. The redesigned approach. I never asked this one.

And the question I never asked, which is the whole reason this chapter exists: a prize doesn’t select for people who want your product. It selects for people who want the prize.