Chapter 2: The Sameness Machine

Chapter 2 · Part One: Why Your Best Work Keeps Coming Out Average

Read this sentence and tell me whose it is.

Blank uses agentic AI to triage, investigate and respond at machine speed, so your analysts can finally focus on the work that matters.

Whose is it? Nobody's.
Whose is it? Nobody's.

If you work in security marketing, a few company names just went through your head. Maybe yours.

It’s nobody’s. A model wrote it, in about nine seconds. How, and why that matters, is Chapter 13. This chapter is about why you recognized it.

For the keynote I built a wall of twelve homepages around sentences like that one. The names and logos are invented, but they’re built on what’s out there. Look at how different they are. Different typefaces. Different colors. Serif, sans, all caps, lowercase. Light, dark, gradient, terminal. Somebody designed every one of them on purpose, and no two of them look alike.

Now read them.

Agentic. AI SOC. Machine speed. Triage, investigate, respond. So your analysts can focus on the work that matters.

Twelve different brands. One sentence.

Twelve homepages. Invented names, invented logos, one sentence.
Twelve homepages. Invented names, invented logos, one sentence.

The real pages these are built on each had a different designer, a different agency, a different brand system, and a real budget. All of that money went into the part that’s different. None of it went into the part that’s the same.

If I swapped six of mine out for six real ones, some of them maybe yours, I don’t think you could tell me which six.

I didn’t put any real companies on the wall, and I’m not going to name any here. Most of them will be at the same conferences I go to, and naming them lets you file this as a fight between vendors instead of a description of a condition that includes me.


The part where I’m on the wall

And before anybody feels singled out: go look at 7ai.com. You’ll find at least three of these sentences on our homepage. I approved them.

I’d love to tell you I was the exception. I could defend every one of those sentences in a room for twenty minutes, and that’s exactly the problem. Each one is reasonable. Together they’re the category.


Mine took nine seconds

Mine took nine seconds.

The real ones took six months. A brief. A creative review. A workshop with sticky notes on a wall. Three rounds of stakeholder feedback. And somebody senior who looked at the final version and said, yes, that’s right, that’s us.

And every one of them got the nod.


Nobody in this picture was being lazy

This is the sentence I care most about in the first half of the book, so I want to spend a minute on it.

Nobody on those real teams was lazy. They cared. They stayed late. Somebody built three concepts and killed two. Somebody argued about whether the verb in the headline was strong enough, and somebody else was probably right about it.

Months of work. Real money. Real care.

And the output converged anyway.

If the sameness came from laziness, this would be a pamphlet instead of a book. Try harder. Care more. Hire better. Fire the agency. We’d all know what to do and most of us would already be doing it, because the people in this industry aren’t lazy and aren’t stupid and in my experience care a great deal about work that most of the world finds incomprehensible.

If effort were the answer, the problem would already be solved.

So something else is producing this. Something that operates on people who are all trying hard, in a dozen different buildings, at the same time.

You met him in Chapter 1. Now let me show you the machine we built for him.


Your review process is a fluency machine working perfectly

Most marketing organizations have accidentally constructed a process that selects for familiarity and filters out everything else. Not through incompetence. Through completely reasonable governance that nobody would remove if you described it honestly.

Watch what a round of stakeholder feedback actually does.

The concept goes to eight people. Each of them is being asked, implicitly, to find something. Nobody forwards a review with “no notes” and feels like they contributed, and in a lot of companies nobody forwards a review with “no notes” and keeps being invited to reviews. So each person finds the thing that snags them.

And the thing that snags a reader is almost always the unfamiliar thing, because from the inside, unfamiliarity feels exactly like friction, and friction feels exactly like a problem.

So the notes come back. Every note is reasonable. Every note, independently, sands off one unfamiliar edge. Round two is smoother. Round two is also closer to the category mean, because the category mean is by definition the smoothest available version of any statement in your market. Run three rounds and you have arrived.

The process didn’t fail. The process did precisely what it was built to do, which is remove friction.

Nobody ever told it that friction and distinction are the same substance.

Every round of review is a vote for the most familiar available version, cast by people who sincerely believe they’re voting for quality. That’s how twelve teams spend months and produce one website.


It doesn’t stop at your building

The review round explains how one company drifts to the middle. It doesn’t explain why a whole category drifts to the same middle. For that you need a sociologist.

In 1983, Paul DiMaggio and Walter Powell published one of the most cited papers in organizational sociology, asking why organizations in the same field come to look so alike. One of the forces they named is mimetic isomorphism: when organizations are uncertain, they copy whoever looks successful. Nobody orders them to. Imitation is just the cheapest available answer to “what should we do?” when nobody actually knows.

Nobody in our category decided to look the same. Uncertainty did it for us. And I can’t think of a category in more uncertainty right now than one where the product, the buyer’s budget line, and the analysts’ name for the market have all changed inside the same two years.


Why this costs more in our category than in most

Three reasons, and they compound.

The buyer is professionally skeptical. You’re selling to people whose entire job is to assume the thing in front of them isn’t what it claims to be. A security buyer reads a vendor homepage the way they read a suspicious email. When every vendor makes the same claim in the same words, the claim doesn’t get evaluated and rejected. It gets discarded as category-level noise before evaluation starts, and you’re underneath the noise floor with everybody else.

The buying window is short and rare. Nobody wakes up wanting to buy security software. Something happens: an incident, a board question, an audit finding, a renewal, a new CISO with a mandate. The window opens, a few names come to mind, the window closes. You aren’t competing for a decision. You’re competing to be one of the names that surfaces in a ten-minute conversation you aren’t present for. That’s a memory contest, and a homepage that reads like seven others has already lost it, months earlier, quietly.

The evaluation is a committee, and your champion has to describe you. Even when you get in the door, somebody inside that company has to explain you to four other people who weren’t on the call. If your description is indistinguishable from the incumbent’s, your champion has nothing to carry. You didn’t lose the bake-off. You lost the hallway conversation after it, and nobody will ever tell you that’s what happened.

Sameness in a consumer category costs you preference. Sameness in ours costs you recall, description, and repetition, which means it costs you the entire front of the process, invisibly, in a way that will never appear on a dashboard.


The Sameness Audit

Here’s the first real tool in this book. Ninety minutes, no budget, and you should run it before you go much further, because everything after this lands differently once you’ve seen your own category laid flat.

You aren’t looking for who’s bad at this. You’re looking for what your category has silently agreed on, because the agreements are the only things available to break.

Step 1: Pick the set (10 minutes)

Eight to twelve companies. Not your competitive battlecard list, which is organized around who you lose to. Organize this list around who a buyer encounters in the same week they encounter you:

  1. Three or four real head-to-head competitors.
  2. Two or three adjacent vendors a buyer would lump you with even though you don’t compete.
  3. Two or three names from the last sponsor roster of the show you all attend.
  4. Your own company. Not optional, and the entire reason the exercise works.

Step 2: Capture (20 minutes)

Same browser, same window width, same zoom, same day, light mode wherever offered. Hero down to just below the fold, not full page. Consistency matters more than quality here, because you’re comparing shapes and a reflowed responsive layout will lie to you.

Step 3: Score four axes (40 minutes)

Companies as rows, these as columns.

Vocabulary. Write out every noun and verb phrase in each headline and subhead. Count how many companies use each. Anything appearing three or more times is category vocabulary, which means it’s invisible. Anything appearing once is either a position or a mistake, and deciding which is your job.

Structure. Score presence or absence of: hero over image or gradient, one-to-two-sentence subhead, paired CTAs, logo bar within one scroll, three-column benefit section, stat band, testimonial carousel. Count the matches. Seven identical structural choices across ten companies isn’t a trend, it’s a template nobody wrote.

Navigation. List top-level nav items for each. Cheapest axis, most revealing, almost never run.

Proof. What does each company offer as evidence? Logos, a named case study, a number, an analyst mark, a certification, a product screenshot, a demo video, nothing. Note what almost everybody offers. Then note what almost nobody does, which is the more interesting column and the one people skip.

Step 4: The swap test (10 minutes)

Take the six most similar heroes. Strip every logo and brand name. Print them, shuffle them, hand them to somebody in your company who doesn’t work in marketing. Ask them to name the companies.

Don’t do this to be clever. Do it because you need to watch a colleague fail at it in front of you, and because their failure is the only version of this argument that survives contact with a skeptical executive. A slide that says “we all sound alike” is an opinion. Your head of sales unable to identify his own company’s homepage is a finding, and it will be repeated in that building for a year.

Step 5: Write the sentence

Finish by writing down the single sentence your category has agreed on without ever voting on it. For the AI SOC cluster in 2026 it’s roughly this:

We use AI agents to handle the alert work your overwhelmed analysts can’t get to, so they can do more valuable work.

That sentence is true. It’s true of my company. It’s probably true of yours. And it’s worth nothing, because it’s the noise floor, and every dollar spent saying it more beautifully than the last vendor is a dollar spent on a slightly more elegant way of being unidentifiable.

Once you’ve written it down, you have something to work against.

You can’t deliberately leave a place you haven’t located. That’s the entire purpose of the exercise, and it’s why the audit comes before any of the method in this book.


What this chapter isn’t saying

It isn’t saying category vocabulary is a sin. You need some of it or nobody knows what shelf you’re on, and a buyer searching for an AI SOC has to be able to discover that you’re one.

It isn’t saying the answer is strangeness for its own sake. I’ve made expensive strange things that worked and I’ve watched expensive strange things fail, and one of those failures gets a chapter of its own, in more detail than is comfortable.

What it’s saying is narrower and harder to argue with. Your category has converged on a set of decisions nobody in it can remember making. Those decisions are now invisible to the people inside them, including you, including me, right up until somebody puts a dozen of them side by side.

The work of finding them is cheap. It’s the finding that’s expensive, and only in the sense that you can’t un-see it afterward.


The Comfort Check

Chapter 2: the category audit

1. The comfortable assumption. We know what our competitors say, because we track them.

2. The true objective. Find the decisions our whole category made without noticing, because those are the only ones available to break.

3. The redesigned approach. Capture eight to twelve heroes under identical conditions, score four axes including the navigation, run the swap test on somebody outside marketing, and write down the one sentence everybody has agreed on.