Mehdi Khalile

Positioning is not the list of things you are good at. It is the list of things you have agreed not to claim.

Which is why most of it fails. Every review adds a word, nobody removes one, and the sentence ends up covering every service, every sector and every size of client. It is now true, defensible, agreed by everybody, and it distinguishes nothing.

The test is simple and uncomfortable. Read your positioning and ask which competitor would refuse to say it. If the answer is none, you have written a description of your industry rather than of yourself.

Something has to be given up. A sector you will not serve, a problem you do not solve, a kind of buyer you are wrong for. Naming one is what makes the rest believable.

Being for everybody reads as being for nobody, and buyers are quicker at spotting it than teams are.

“Nobody knows who we are” is the most common brief in marketing, and it is almost always the wrong one.

Plenty of people have heard of you. What they cannot do is repeat what you actually do, which is a different failure with a different fix and a much lower price.

There is a test that takes an afternoon. Ask three customers to describe you to somebody else. Not what they like about you, what they would say you are. Three different answers means every euro of reach is buying three different messages.

Awareness is the comfortable diagnosis because it costs money and nobody has to change their mind. Clarity costs an argument, usually with whoever wrote the sentence you are using now.

Get the sentence right, then buy reach for it. In that order the spend compounds.

Before anybody evaluates you, they have to file you. What is this, roughly, and what is it like.

That filing happens in a second or two, before a single feature is read, and it decides which competitors you get compared against and which budget you come out of. Get filed wrong and every advantage you have is being judged against the wrong alternatives.

Most companies leave it to chance. They open with what makes them different, which cannot be understood until the reader knows what they are different from, so the reader files them anyway, badly, using whatever is nearest.

Say the category first, plainly, even if it feels reductive. Then the distinction. In that order it sticks; in the other it does not survive the first retelling.

Being hard to categorise is not sophistication. It is a cost, and the buyer pays it first.

Somebody says we should be on that platform. Nobody disagrees, because disagreeing sounds like being behind.

Six weeks later there is an account, a posting schedule, and still no answer to why. A channel is a place. A strategy is a choice about who you are for, what you are claiming, and what you are refusing to do.

Three questions before opening anything. Which specific buyer is there, doing what. What behaviour does the place reward, and can you supply it every week without heroics. What are you dropping to pay for it.

That third one is the one that gets skipped, and it is the one that decides. A new channel is not free time. If nothing is being dropped, something is about to be done at half strength.

One channel done properly beats four kept alive. A dormant account is not neutral, and buyers read it exactly as it looks.

Most lost deals are not lost to a rival. They are lost to nothing happening.

The prospect agreed with you, liked the price, and then went back to the spreadsheet they have used for six years, because carrying on is free and changing is not.

Which means the case you have to win is not “why us”. It is “why now, and why is doing nothing worse than it looks”. Almost no marketing makes that second case, because it is uncomfortable and it is not about the product.

The cost of the current way is invisible to the person living inside it. Your job is to make it countable: hours, errors, the thing that goes wrong every March.

Beat the rival and you take a customer. Beat inertia and you make the market.

Nobody writes their best line at a desk.

They overhear it. A customer explains what they were trying to fix, in their own words, and the sentence is better than anything currently on the website.

Which is why ten conversations beat any number of surveys. A survey returns the options you already wrote. A conversation returns the sentence you would never have thought of.

Four questions do most of the work. What were you doing before you found us. What almost stopped you buying. How would you explain us to a colleague. What did you expect that did not happen.

Then write the answers down exactly and do not tidy them. The polish is the thing that makes copy sound like everybody else’s.

Marketing generated it. Sales called the ones they were given. Both teams did their job and can prove it, and the enquiry still sat for four days.

Because the gap between them belongs to nobody. Tools are bought per department, so they are configured per department, and the seam is the one place with no owner and no report. That is exactly why it is the most reliable place to find lost revenue.

Put a rule on the gap. Route by rule rather than by who is free, stamp the time so the delay becomes a number instead of an argument, and reopen anything nobody has touched by a deadline you set.

The qualifying and the calling stay human. That was the part the whole budget existed to make possible.

Measure the gaps, not the stages. Every team measures its own stage and every stage looks fine.

Automation does not fix a broken funnel. It scales it.

A tool has no opinion about the process you point it at. It runs that process faster and more often, so if you lose people at step three, you now lose them at step three at volume, and the reporting looks busier than it did before.

The first month always reads as a win. Sends are up, touches are up, tasks completed are up. None of those is the number anybody hired you to move.

So find the step where people actually leave, fix it by hand for a fortnight, and prove it moved. Then automate it, because now you are multiplying something you have watched work.

Speed is a multiplier, and a multiplier has a sign.

Button colours, headline variants, the order of three bullet points. All testable, all safe, and all downstream of anything that would actually move.

The reason teams live there is that those tests cannot embarrass anybody. Testing the offer can. Testing the audience can. Testing the promise means admitting the current promise might be wrong, and somebody in the room wrote it.

Order of magnitude lives upstream. Who you are talking to, what you are offering them, and what you are claiming it does. A better word on a button cannot rescue any of those being wrong.

So work down, not up. Audience first, then offer, then promise, then message, and only then the page furniture.

If a test cannot change what you build next, it is decoration with a confidence interval.

Somebody hears about you from a colleague, sees a post two weeks later, searches your name a month after that, and arrives through a link they cannot remember.

Your report will credit the search. Everything that made the search happen goes uncredited, and the budget follows the credit, which is how companies quietly defund the thing that was working.

You are not going to solve this. The paths are private, they cross devices, and the most influential part of most of them is a conversation nobody logged. Every attempt to close that gap costs more than the precision it buys.

So invert the question. Before commissioning any more measurement, ask what you would do differently if you knew. If no answer changes the decision, the measurement is entertainment.

Then use the crude instruments, which work: ask people how they heard, watch what moves when you switch something off.