Chapter 3: Safe Is Just a Bet You Don’t Have to Defend

Chapter 3 · Part One: Why Your Best Work Keeps Coming Out Average

The thing I remember about asking the board for the money is that I wasn’t asked whether the campaign would work.

I was asked what happens if it doesn’t.

That’s a completely reasonable question and I want to be fair to the people asking it. This was a very large number for a company of that age, spent on a thing nobody in our category had ever done. If it fell flat, that isn’t a learning experience you write a thoughtful LinkedIn post about. That’s a real problem, with real consequences, for people who had put real money behind us.

What I noticed, standing there, is that nobody had ever asked me that question about the alternative.

Nobody had ever sat me down and said: Nate, walk us through what happens if we spend the same money on the standard program, the trade shows and the paid search and the content engine and the six-week email nurture, and at the end of the year we’re still the company that ninety-one percent of our market has never heard of. What’s the plan then?

That question doesn’t get asked. Not because the board is asleep, but because the standard program doesn’t present itself as a decision. It presents itself as the absence of one.

Psychologists have a name for this. Omission bias: we judge harm caused by acting far more harshly than the same harm caused by doing nothing. The standard program never registers as an action, so its failures never register as anybody’s fault.


Both of these are bets

We talk about risk in this industry as though it’s a thing you either take or you don’t. You take a risk when you do something strange. You don’t take a risk when you do the normal thing.

That isn’t how it works and I think most people know it isn’t how it works, which is why the belief survives by never being said out loud.

The normal thing is a bet. You’re betting that being indistinguishable from eleven competitors will produce a result. You’re betting that the buyer will pick you out of a lineup where everybody is dressed identically. You’re betting that the reason last year’s program underperformed was the market, and that this year’s market will be different.

Those are substantive claims about the world. You can be wrong about them. People are wrong about them constantly, at enormous expense, for years at a time.

So the difference between the strange thing and the normal thing isn’t that one is a bet and one isn’t. Here’s the honest version of the difference:

Option A. You do something different. It might not work. If it doesn’t work, everyone will know exactly whose idea it was, because your name is on the only thing in the plan that looks like a decision.

Option B. You do what the category does. It probably won’t work either. But if it doesn’t work, nobody will be able to tell, because your results will look like everybody’s results, and the story will be that the market was tough, that the buying cycle stretched, that budgets were frozen in Q3. All of which will also be true.

Read those again and notice that the difference has nothing to do with the odds.

It’s a difference in attributability. Option B isn’t safer. It’s quieter when it fails.

That’s the entire asymmetry, and once you can see it you can’t stop seeing it. We pick the option that doesn’t require us to stand up in a room and explain ourselves, and then we call that being responsible with the company’s money.


The real cost of the quiet failure

There’s a second-order problem here that took me most of my career to notice, and I think it’s worse than the first-order one.

A failure nobody can attribute is a failure nobody learns from.

When the strange thing fails, you get an autopsy. Everybody in the building knows what was tried, they know roughly what it cost, and there’s enormous organizational appetite to understand exactly what went wrong, mostly because several people want it on the record that they had concerns. That’s uncomfortable. It’s also the single most valuable byproduct of the whole exercise. I learned more from the failure in Chapter 8 than from either of the wins on either side of it, and it wasn’t close.

When the safe thing fails, there’s no autopsy, because there’s no body. The numbers came in a bit light. The team is already planning next year, which will be structurally identical, with a slightly different allocation between two channels and a new agency. Nobody writes anything down, because nothing happened, and nothing happening isn’t an event.

So the organization that consistently chooses Option B doesn’t just underperform. It stops accumulating knowledge about its own market. Five years of that and you have a marketing team with a great deal of experience and almost no learning, running a program that nobody currently employed there can explain the original reasoning behind.


We have to take big swings. From my 2022 CyberMarketingCon keynote.

The Bet Card

The point of this chapter isn’t “take more risks.” I’m not telling you to be reckless and I’m not telling you that strange is the same as good. I’ve made expensive strange things that worked, and I’ve been in the room for expensive strange things that didn’t, and the difference between them was almost never the size of the swing.

The point is narrower: know which bet you’re making, say it out loud, write down what would prove it worked, and then make it on purpose instead of by default.

Here’s the instrument. One page, filled in before you spend anything. It works for a booth, a campaign, a report, a rebrand, a webinar series, a sponsorship, or a hire.

1. What we’re betting. One sentence, in plain language, with no marketing nouns in it. Not “increase awareness in the mid-market.” Something closer to: “we’re betting that a buyer who sees this once will be able to describe us to a colleague a month later.”

2. What we’re betting against. The alternative use of this money and this quarter, named specifically. If you can’t name it, you haven’t made a choice, you’ve made a purchase.

3. What has to be true for this to work. Two or three conditions, each one falsifiable. This is where most bets quietly die, because writing down the conditions is what reveals that one of them is already false.

4. What would prove it worked. A number, a source, and a date. The number must be measurable by somebody other than the person who wants it to have worked.

5. What would prove it failed. Written at the same time as number four, and this is the one people skip. A bet with no failure condition isn’t a bet, it’s a purchase you intend to describe favorably later.

6. When we look. A date. Put it in the calendar now, with the people who will be in the room.

7. Who decides. One name. Not a committee, and not “we’ll discuss it.”

Two rules about the card.

Fill in five before four. If you write the success condition first, you’ll write the failure condition as its inverse, and you’ll have learned nothing. Write down what failure looks like as its own idea, from scratch, and you’ll often discover you have described the most likely outcome.

Do it before, never after. A bet card written after launch is a press release. The entire value of the thing is that it’s a commitment made while you’re still capable of being wrong.


How to defend a bet without claiming it’s safe

Back to the board.

The thing I didn’t do, and I want to be specific because I think this is the transferable part, is argue that the campaign was safe. It wasn’t safe. Everybody in that room could see that it wasn’t safe, and any attempt to dress it up would have cost me the only thing I actually had, which was credibility about my own assessment.

What I argued was three things:

  1. This is the best available strategy for the problem we actually have, which is that ninety-one percent of our market doesn’t know we exist and our sales motion can’t fix that by working harder.
  2. The alternative isn’t safety. The alternative is spending a comparable amount to remain in the ninety-one percent, and I’m willing to describe what that looks like in detail.
  3. We’ll know either way. Here’s the measurement, here’s the baseline, here’s when we take it again, and here’s the number that means I was wrong.

That third one is what actually moved the room, and it’s the one most marketers skip. An executive’s resistance to an unusual idea is very often not resistance to the idea. It’s resistance to unfalsifiable spending. They have been handed unfalsifiable spending before. They have watched a year of it produce a slide that says the brand is stronger.

When you walk in with a defined failure condition, you have changed what you’re asking for. You’re no longer asking them to believe you. You’re asking them to fund an experiment with a stated cost and a stated verdict date, which is a fundamentally easier thing to say yes to, and which also happens to be the honest description of what you’re doing.

One warning. Don’t offer a failure condition you aren’t prepared to honor. If you say you’ll kill it and then you don’t kill it, you’ve spent the instrument, and you won’t get it back at that company.


The version of this with no budget

Everything from here to the end of the book has a free version, and I’m going to list it at the end of each chapter, because the most common objection I get to this material is that it’s easy to talk about swings when somebody else has already given you the budget to take them.

That objection is fair about the examples and wrong about the method.

Here’s the free version of this chapter. You don’t need money to make a bet. You need something at stake, and attention is a thing you already control.

  1. Bet the format, not the budget. Next time you owe somebody a deck, send a one-page memo instead. Same information, different vehicle, zero dollars, and you’ll find out inside a week whether your organization rewards clarity or ceremony.
  2. Bet the words. Take the one comparative claim your category leans on hardest, the one everybody makes, and remove it from your next campaign entirely. No “faster,” no “smarter,” no “more accurate.” You’ll have to find something true to say instead. That’s the exercise.
  3. Bet on the people. Find the piece of work in your organization that somebody senior called too weird and that you privately thought was the most interesting thing you had seen all year. Put your name next to it. That costs nothing and it’s a real bet, because it’s attributable, which is the whole thing.

Each of those has a failure condition you can write on a card this afternoon.

And each of them is a bet you’ll have to defend, which is exactly why they’re worth making. The other kind was never free either. You were just not being billed for it in a way you could see.


The Comfort Check

Chapter 3: the annual plan

1. The comfortable assumption. The standard program is the low-risk option, and the unusual idea is the bet.

2. The true objective. Make a decision we can be proven wrong about, on a date we set in advance.

3. The redesigned approach. A Bet Card filled in before the money moves, with the failure condition written before the success condition, one named decision-maker, and a review date already in the calendar.