Chapter 11: What Worked Before Is a Ceiling, Not a Floor

Chapter 11 · Part Three: The Practice

7AI launched in February of 2025. Black Hat that August was our first real show.

That year, AI in the SOC wasn’t yet the thing every company said. So we built a booth that was an art gallery. Not a booth with art in it. A gallery. The demo stations were disguised as the exhibits. Quiet, still, deliberately strange, sitting in the middle of a show floor that’s engineered end to end for sensory overload.

It worked. It worked really well.

The art gallery booth, Black Hat 2025.
The art gallery booth, Black Hat 2025.

That’s the whole first half of the story and it’s the least useful part of it.


The next year

The following year, everybody was an AI SOC company. Every booth. Every banner. Every hallway conversation. The category had arrived and it had arrived all at once, in the same words, which is Chapter 2 happening in three dimensions with carpet.

And here’s the comfortable move, the one my Neanderthal voted for before I had finished reading the floor plan.

Do the gallery again.

Look at the arguments for it, because they’re good arguments and you’ve made them:

  1. It worked.
  2. People loved it.
  3. We have the photos, the plan, the vendor, the drawings, the budget line.
  4. Most of the show hadn’t seen it, so to most of the audience it’s new anyway.
  5. Doing something else means starting from zero with four months to go.

Every one of those is true. Number four is even a reasonable argument on the merits. And taken together they’re the most persuasive case in marketing, which is why so many marketing departments are on year six of it.

The best thing we had ever done was about to become the ceiling on what we could do next.

That’s the sentence the belief rests on. Not that repeating yourself is lazy. That your own best work, left in place, quietly defines the upper bound of the next thing, because everything after it gets evaluated against it rather than against the problem.


What we did instead

We took over the Sphere. And then put almost no words on the booth at all.

Not a product billboard. Not our feature list at four hundred feet. The whole goal was one reaction, and it’s the title of the next chapter.

What the hell is that?

People came to the booth specifically because they had seen it and wanted to know what it meant. Curiosity first, explanation second. That order matters enormously and almost nobody runs it that way, and Chapter 12 is entirely about why.


The criticism

People had opinions.

That was the gist of a lot of what got posted. I can’t believe they spent that much. That money could have gone somewhere else.

They might be right.

I’m not going to tell you the Sphere was the correct allocation of capital for your company, because it almost certainly isn’t. I had a budget and an organization that would back a swing like that. Most people reading this don’t, and that has nothing whatsoever to do with whether they’re good at this.

I’m not telling you to do what I did. I’m telling you to use the strategy I used.

And the strategy was free. We noticed what everybody else was about to say, and we said something else.

What a tenth of it would have bought

Which is the only question worth asking after a paragraph like that one, so here’s the real answer rather than a shrug.

First, what it doesn’t buy, because this is where most people go and it’s the wrong read of everything in this book.

A stunt. Somebody stages a fake protest outside RSA. Somebody drives a branded truck in circles around the convention center. Somebody builds a thing purely to be photographed. I’m against all of it, and not on grounds of taste. A stunt is engineered to be noticed and nothing else, which means the noticing is the entire product. Once it’s over you have a photograph and no asset, and the only thing anybody remembers is that a company did a thing, not which company or why.

Different isn’t a synonym for attention-seeking. Go back to Chapter 7 and run the check properly. What’s the objective? Then what’s the cheapest route that actually reaches it?

Here are four we’ve used, all of them at a fraction of a Sphere, all of them available to a team with no budget at all.

1. Do real research and publish it.

Not a survey of four hundred CISOs about their top three concerns. Actual research: put somebody on a real question, let them find something nobody has documented, and write up what it is, why it matters, and what to do about it.

Two of ours. Last year our team found a QR-code phishing campaign that walked past every filter it hit. We called it Quish Splash, and we published what it was, why it mattered, and what to do about it. CRXfiltrate, an undocumented JavaScript execution backdoor running inside a network of Chrome extensions.

Both got written about by people who don’t owe us anything. Trade press picked up both (HackRead on Quish Splash, eSecurity Planet on CRXfiltrate), and Quish Splash ended up on a podcast we never pitched. That’s earned carriage, and no amount of paid media buys it, because the thing being carried is a finding rather than a message.

The cost is a person’s time and the discipline to let them actually dig instead of pulling them onto launch support in week two.

2. Find the right people instead of shouting at everyone.

Use AI to read your own site traffic, work out who’s actually showing up, build a lookalike from the ones that match your real ICP, and run LinkedIn ads at exactly that list.

Unglamorous, nobody will ever write about it, and it’s a completely different activity from buying category keywords and hoping. You aren’t increasing reach. You’re deleting the part of your spend that was never going to work, which is a bigger lever than most brand campaigns and costs a fraction of one.

3. Make the customer the hero, not the product.

Our Chief Innovation Security Officer campaign. The subject is the security leader doing something genuinely new, and we’re in the background.

The reason this works isn’t generosity. It’s that a customer will carry a story they’re the hero of, into rooms you’ll never be invited to, for years. A case study where the vendor is the hero gets forwarded by nobody, including the customer who’s quoted in it.

4. Start a podcast, and make it about them.

Do Human Work costs approximately nothing. Equipment and travel. That’s the whole line item.

What it produces is an hour of real conversation with a security leader about something they actually care about, and a number of those guests have become customers afterward. Not because it was a sales motion, which it isn’t and which would kill it immediately, but because an hour of genuine attention is rarer in this industry than anything you could buy.

What all four have in common

None of them buy attention. All four of them earn carriage, which means somebody who doesn’t work for you moves your message for reasons of their own. A journalist writes the research up. A customer tells their own story. A guest recommends you to a peer. Even the targeting one works by finding the people who were already going to care.

The Sphere bought attention, and it worked, and it was expensive. These earn it, and they’re slower, and they compound.

If I had a tenth of that budget I’d have put the whole thing into the first one and hired a second researcher.


Nobody decides to stop trying

Here’s why this matters past one booth.

Stagnation never feels like stagnation.

It feels like efficiency. It feels like using the thing that already worked, which is a completely reasonable thing to do, right up until it’s the only thing you do.

There’s a picture I keep saved in my phone under favorites, and I can’t tell you how many times I’ve pulled it up. A school cafeteria tray of macaroni, with a handwritten sign that reads “macaroni and fuck it.”

Nobody chose this. Somebody just did what worked last time.
Nobody chose this. Somebody just did what worked last time.

Nobody decided to serve that. Nobody stood in that kitchen and chose it as a strategy. Somebody just did what worked last time. And then did it again. And the thing about doing what worked last time is that each individual instance is defensible, and the aggregate is a sign on a tray that has given up.

Your annual threat report is on year five. Your webinar format hasn’t changed since before the pandemic. Your booth has the same shape and the same demo stations and the same shirts in a different color. Every one of those decisions was reasonable in the year it was made.

And look, I get it. Nobody can go all out on everything, every day. There are deadlines and budgets and things that just need to get done. So keep that picture in your head and ask yourself: is this good, or is it macaroni and fuck it? At least then you know which choice you’re making.


Why the old answer blinds you

There’s a classic experiment on this. In 1942, Abraham Luchins gave people a series of water-jar puzzles that all shared the same three-step solution. Then he gave them one that could be solved in a single step. Most of them kept using the long method. They couldn’t see the easy answer, because the old answer kept working. People who had never learned the long method found the shortcut right away.

It’s called the Einstellung effect. Your best idea doesn’t just sit there. It gets in the way of the next one.


The Ceiling Audit

Run this before planning season, not during, because during planning season every play on the list already has an owner and a slide.

Step 1. List every play you’re going to run again next year because it worked. Not everything you do. The things whose justification is that they worked.

Step 2. For each one, write the year it started and the year it last changed materially. Not cosmetically. Materially. New color isn’t a change. New format is a change.

Step 3. For each one, answer: what would we do if this play were banned?

Not “what would we do instead,” which produces a slightly smaller version of the same thing. Banned. Prohibited. Not available at any budget.

Step 4. Look at the answers. Somewhere between one and three of them will be better than the play you were going to run. That isn’t a hypothetical; it happens every time I’ve run this, and it happens because the ban is the constraint from Chapter 7 applied to your own history.

Step 5. Kill one thing that’s working.

This is the step people skip and it’s the only one that has teeth. Not the worst thing. Something that’s genuinely working, that you could defend, that nobody is asking you to stop. One per year.

The reason is arithmetic. A marketing calendar has a fixed size. If you only ever remove things that failed, the calendar fills up with things that worked once, and within three years you have no capacity left for anything new and a completely defensible explanation for why.


Criticism is a symptom, not a scoreboard

I need to be careful here, because there’s a version of this chapter that produces an insufferable person.

Criticism is a result of being different. It isn’t evidence that you’re right. Those are completely different claims and conflating them is how marketers end up defending bad work by pointing at the people who dislike it.

The honest version:

  1. If nobody reacts, you were probably invisible. Silence is a finding.
  2. If people react, you were visible. That’s all it establishes.
  3. Whether you were correct is settled by Chapter 9, not by the comment section.

The useful thing about criticism isn’t its emotional valence, it’s its content. Look at what the criticism is actually about.

Criticism of your judgment is expensive and worth taking seriously. Criticism of your taste is mostly noise. Criticism that takes the form of explaining your own strategy back to you incorrectly is genuinely valuable, because it tells you precisely which part didn’t travel.

And one thing that’s true and that I wouldn’t have believed ten years ago: the people who criticized the Sphere publicly could all describe what we did. That isn’t nothing. It isn’t success either. It’s a leading indicator, and it belongs in the leading indicator row of the measurement plan rather than in your feelings.


Strange is free

The Sphere is the expensive example and it’s the least useful one in this chapter. Here’s the catalogue of things that cost nothing and that almost nobody in this industry does, because the cost of them isn’t money.

  1. What you name things. Internal program names, feature names, the name of your report. Naming is free and it’s the highest-leverage unclaimed surface in B2B.
  2. What you refuse to say. Ban one category word for a year, across everything. The hole it leaves has to be filled with something true.
  3. Who speaks. Put the engineer, the analyst, or the customer on stage instead of the executive whose title says they belong there. Costs nothing, prohibited by nothing except habit.
  4. The format. A memo instead of a deck. A ten-minute webinar instead of forty-five. A one-page report instead of a forty-page PDF nobody finished.
  5. The ten minutes before. Everybody puts a countdown timer and a logo on screen before a webinar starts. That’s ten minutes of undivided attention from your most engaged audience, given away to a clock.
  6. What you send after a meeting. Every vendor sends a recap email and a deck. Almost nobody sends the one thing the buyer actually needs to forward internally.
  7. The back wall of the booth. It faces the aisle for three days and almost every company puts their tagline on it.
  8. What you refuse to gate. Ungating one thing that everybody gates is a position, not a tactic, and people notice.
  9. Your job titles. Free, and they’re on every email signature your team sends.
  10. Who you thank in public. Nearly unused in this industry, and it’s the cheapest distinctive act available.

Every one of those is available to a person with no budget and no authority beyond their own work. That’s Chapter 18’s belief arriving early, and it arrives here because this is the chapter where the money is most likely to distract you from the method.


The Comfort Check

Chapter 11: the second year

1. The comfortable assumption. It worked. So run it back.

2. The true objective. Be the thing nobody can walk past.

3. The redesigned approach. Go bigger and stranger than your own best work.