For about a year, whenever anyone asked what I was building, I said “an AI email client.” It seemed like a good answer. It was short, it was accurate, and people nodded. It took me an embarrassingly long time to notice that the answer was the problem.
Here is what happens in someone’s head when they hear “email client.” They think of Gmail. They can’t help it. Over a billion people use it, and most of them have had the same address for fifteen years. Then, if they follow this sort of thing, they think of Outlook, and maybe Superhuman. And then they put me somewhere underneath all of that, in a drawer marked other. Nothing I could build would get me out of that drawer, because the drawer was assigned before anyone opened the product. It was assigned by my own sentence.
what you’re actually working on
The mistake is so easy to make that I want to spend some time on why.
If you’re technical, you have a deep and largely correct belief that the better thing wins. You’ve watched it happen. Better databases displace worse ones. A faster compiler beats a slow one. The belief is true within a fixed comparison, and the trouble is that the comparison itself is negotiable, and gets negotiated in the first second, usually by you, usually by accident. Founders spend years optimising the thing being compared while handing over the choice of what it gets compared to, free, in casual conversation, several times a day.
Ries and Trout’s book on this turns on one idea: [1]
“Positioning is not what you do to a product. Positioning is what you do to the mind of the prospect.”
The object you’re working on is the prospect’s mind, and the product is nearly incidental. I resisted this for a while because it sounds like the sort of thing marketers say to justify their existence. But it kept being true. What a customer means by “your product” is a two-word summary sitting next to some other thing they already understand, and the summary does most of the work of the sale before any of your work is even visible.
the mind is already full
The reason you can’t fix this by explaining has to do with how full the mind already is. A few facts, each of which kills a plan that founders reach for by instinct.
- Minds hold about seven brands in a category, and usually two. Try to name toothpastes. Two arrive instantly, a third with effort, and then nothing. There is no eighth slot sitting empty waiting for the best applicant.
- Minds accept only what connects to something already inside. This has the perverse consequence that a genuinely novel product is harder to sell than an obvious one, which is why so many original companies spend their first two years being mistaken for something they aren’t.
- Minds react badly to being corrected. People will listen with total equanimity about a subject they know nothing about. Tell them a belief they already hold is wrong and they stop listening. Advertising works by seduction; a debate just makes people defensive, and defensive people close the tab.
- Minds, once made up, stay made up. This isn’t stupidity. It’s how you avoid recomputing the world every morning. Yours does it too.
Put those together and the plan where you patiently explain to the market why you’re better dies on the spot. You can attach yourself to something already in there, or you can stay outside.
the ladder
The picture that made this concrete for me is a ladder. Imagine that for each category there’s a ladder in the customer’s head, with one brand per rung, in order. Beer has a ladder. Rental cars have one. Email clients have one. Some ladders have six rungs and most have two that anyone can name. Two rules come with the picture. The first is that the top rung takes roughly twice the share of the second, which takes roughly twice the third. The second is that you cannot dislodge the rung above you by any amount of effort. RCA went at IBM head-on in mainframes and wrote off something like $250 million in 1971. [2] They weren’t fools. They were climbing at a spot where somebody was already standing.
So the first question in this kind of work has nothing to do with what makes you special. It’s: which ladder do my own words put me on? Not the ladder you’d like. The one your sentence chose, whether or not you were paying attention when you said it.
the three moves when you’re not first
Being first into a mind beats being better, and since you’re never first, you get three moves. I’ll put them in order of how well they work.
The first and best is to find a ladder you can be top of. Volkswagen sold the small car and let everyone else fight over size; the whole strategy fits in two words, think small. Nyquil sold the night-time cold medicine and gave up the entire daytime market to get it, which sounds insane until you notice they got a rung nobody could take. You want to be a big fish in a small pond and then spend twenty years growing the pond.
The second is to attach yourself to whoever’s on top. 7-Up had no chance on the cola ladder, so it climbed on sideways by calling itself the uncola. Avis said out loud that it was second and trying harder, and turned thirteen straight unprofitable years into profit by admitting its own rank. Telling the truth about your position is a surprisingly underused move, probably because it feels like losing.
The third is to move the leader. When there’s no open rung, you make one by saying something true about the incumbent that changes how people see them. Tylenol spent sixty words on what aspirin does to your stomach before it got around to mentioning itself, and became the best-selling painkiller in America. This is a different thing from comparison advertising, which fails for a reason worth understanding: when you claim to be better, the reader silently asks why you aren’t first, and now they’re thinking about the leader again. Repositioning gives the reader a fact they didn’t have about a brand they already trust, and they update on the spot without ever feeling argued with.
The move missing from that list is the frontal attack, which is the one almost everyone tries first.
for something genuinely new
There’s an exception to all of this for the genuinely new. A mind has no shelf for the unprecedented, so if you’ve built something with no category, you have to hand the customer a familiar object and mark the single difference. The first car was a horseless carriage. Then lead-free gasoline, sugar-free soda, off-track betting, the uncola. Coining a new word for a new thing asks the customer to do work, and they won’t.
the name is the box
Which brings me to names, and to a part of this I’ve been avoiding.
Ries and Trout claim the name is the most important marketing decision a company makes, and I think they’re right, which is alarming given how most names get chosen. If it isn’t sold in a box, the name is the box. A good one tells you what the thing is, stays odd enough to be ownable, and carries some contrast. Atomic Habits sticks and Small Habits evaporates, and the difference is one word doing a lot of work.
My product is called Mailover. Mail plus over. I meant that your mail is over, finished, dealt with. Read cold, it can also mean a makeover for your mail, or a handover of it, and plenty of people hear the makeover. It passes the test the book cares most about, which is that it sounds fine when spoken, since minds run on sound. It fails a harder one: the name doesn’t start the positioning, so it needs a sentence behind it before anyone knows what they’re looking at. Right now the tagline is carrying that entire load, your inbox has a to-do list buried in it, which does more positioning work in nine words than the name does at all. [3]
I’m not going to rename a product that has paying customers. But I know exactly what it costs, and the cost is that every person who hears the name starts one sentence behind.
a worked example: positioning Mailover
Here’s the same reasoning applied to the actual problem, since worked examples are more useful than principles.
My customers are founders and operators at companies of two to fifty people. They run at least two mailboxes themselves, get a hundred emails a day, are personally on the hook for deadlines that arrive from outside the company, and have no assistant. That last clause turns out to do all the work, though I didn’t know it when I wrote it down.
1. the ladder I’m on today
“AI email client” puts me on Gmail’s ladder, and the word client does additional damage, because a client is the thing you open your mail in, so every prospect immediately wonders whether they have to leave Gmail. The answer is no. I’ve already made them think about it.
2. what’s already in the mind
So: what’s already in the customer’s head, near this problem, and attached to nothing at all?
A secretary. Everyone over about thirty knows what a secretary did with the morning post. Opened it, threw most of it out, put three things on the desk and said which one couldn’t wait until after lunch. Nobody in my market has one. The idea arrives fully formed, requires zero explanation, and no software company has claimed it. The customer built that slot years before I turned up, which is the best kind of slot to find. [4]
3. the position in one sentence
So the position is:
Mailover is the AI secretary for people who can’t have a real one. It reads your mail and hands you the list.
Check that against the rules and it holds up.
- It’s a ladder nobody has topped: plenty of tools summarise email and none of them have taken the word.
- It leaves Gmail entirely alone: in fact it makes Gmail the filing cabinet, a job Gmail is genuinely excellent at and has no interest in giving up.
- It hangs on something already in the mind: nobody needs a secretary explained to them.
- It borrows the old to describe the new: the way horseless carriage did. Mail, without the reading.
4. what it costs
Then comes the part nobody enjoys, which is the bill. A position you gave up nothing for won’t hold, because it isn’t distinguishable from a wish. Choosing this one means giving up everyone who enjoys their inbox, teams who want a shared one, anyone running a single mailbox, and all the search traffic for “email client,” which is real traffic that almost never converted anyway, because those visitors came looking for the thing I’ve stopped selling. It also makes an enemy, which is the point. The enemy is the inbox itself.
5. what changes downstream
What surprised me is how much of the value showed up downstream, in decisions I hadn’t connected to marketing at all. Anything that makes the inbox a nicer place to sit is now the wrong thing to build, however good the idea is, and this turns out to be the useful part: a position mostly functions as a rule for saying no on a Tuesday, when some plausible feature request arrives and you have no principled reason to refuse it. It also raises the bar on the product itself. A secretary who misses a deadline gets fired, so accuracy stops being one feature among several and becomes the whole thing. Claim something you can’t deliver and the position turns around and hunts you.
the part I’d bet against
I should say where I’d bet against myself. The word “secretary” sounds dated to some people and old-fashioned about gender to others, and a decade of voice assistants has taught the market that “assistant” means the thing that doesn’t work. The slot I’d defend to anyone is that someone reads your mail and hands back a to-do list and the highlights, with everything else filed into buckets: marketing, newsletters, financials, social. The word for it I’d test five ways before printing it anywhere expensive. And there’s a larger caveat covering everything above, which is that I worked it out alone at a desk, and a position is a claim about the inside of other people’s heads, and the only instrument that measures that is a phone call. All of this is a hypothesis worth about twenty of them.
how you know it’s working
There’s a cheap test for whether any of it is working. Ask five customers what your product is, without prompting them and without showing them the site. Five different sentences means you have a feature list that people are rebuilding from memory, differently each time. The same sentence five times, in their words rather than yours, means that sentence is your position, whether you chose it or not.
Then the hard part, which is holding still. A position only pays if you keep it for years, and the usual way companies lose one is boredom: you win the thing, get sick of your own line long before the market has encountered it once, and go hunting for a replacement. You live inside your message every day and your customer meets it twice a year. Change tactics as often as you like and leave the strategy alone. Real estate in a mind behaves like real estate everywhere else. Move out and someone moves in, and you rarely get the address back.
the template
If you want a way to start, fill in this sentence and notice which blanks are hard, because the hard ones are the findings.
[product] is the [category you can be first on] for [customer]. Unlike [the thing already in their head], it [the one difference]. Not for [who you’re giving up].
Then cover the whole thing and write down the single word you want to own. Volvo took safety. Nyquil took night. Avis took second. If you can’t get down to one word you’re still describing features, and the market will pick a word for you, and it will probably be one that puts you underneath somebody else.
The thing I keep coming back to is that this was never really a marketing problem. Choosing what you get compared to is the same decision as choosing what game you’re playing, which is the same decision as choosing what to work on, and I made mine four words at a time in casual conversation for a year without noticing I was making it at all.
I’m giving up the word client.
notes
[1] Positioning: The Battle for Your Mind, Al Ries and Jack Trout, 1981. Most of the examples here are theirs. The book is full of 1970s brands and is somehow not dated, which I take as evidence the underlying claim is about people rather than about advertising.
[2] The writeoff figure is the book’s. RCA is the expensive case, but the same thing happens at small scale constantly and doesn’t get written about, because a startup that spends two years attacking an incumbent head-on just quietly disappears.
[3] A reasonable objection: if the tagline works, who cares about the name? The answer is that the tagline requires a page, and the name travels alone, in conversation, in an App Store list, in someone repeating it to a colleague. The name is the version of your positioning that goes places you can’t follow.
[4] My own front page currently says “1,000+ working professionals,” which describes roughly everyone with a job, and is a good example of the opposite instinct. Widening the target until it holds nobody is the most common failure in this whole area, and it feels like caution while you’re doing it.