Engineers describe the machine; buyers want to know what changes for them. Fix it by naming the situation the reader is in before naming anything you built, then describing the outcome in their words, then the mechanism last and briefly. The test is whether someone who has the problem can repeat your explanation back to a colleague after hearing it once. If they cannot, they will not buy, however good the product is.
Somebody asks what you are building. You explain. Their face does the thing faces do when someone is being polite. You add detail, because detail has always worked before, and the face gets politer.
The short answer is that you are describing the machine and they are trying to work out what changes for them. Those are different questions, and the second one has to be answered first or the first one never gets asked.
What positioning actually means
Positioning is the decision about what category your product is in, who it is for, and what it replaces. It is not a slogan and not a coat of paint applied at the end; it determines what you build, who you sell to, and what price is credible.
The chronic technical-founder failure is articulating value in feature language. Features and value are not the same thing, and engineers conflate them more than any other group, because in engineering the feature genuinely is the value. A faster query is just better. That intuition stops working the moment you talk to someone who does not care how the query runs.
The order that works
Most founder explanations run mechanism, then features, then benefit, and run out of the listener's attention around the second one. Reverse it.
- The situation. Describe the state the reader is in, in words they would use themselves. This is not persuasion, it is recognition, and it is what buys you the next fifteen seconds.
- The change. What is different after, stated as an outcome rather than a capability.
- The mechanism. How it works, briefly, and only now. This is the part you want to lead with and the part that must come third.
- The proof. Why this is credible. If you have no customers yet, the honest version of this is an explanation of why the approach works, not a borrowed statistic.
Here is the same product both ways. Mechanism first: "It is an agentic platform with a multi-step orchestration layer that maintains context across tool calls." Situation first: "You built something good and nobody is using it, because finding buyers and writing to them is a full-time job you have never done."
The second one is not dumbed down. It is the same product, aimed at the question the reader actually has.
The same page, rewritten
Here is a homepage opening as most technical founders write it: "A unified platform for automating your go-to-market operations with intelligent workflows and deep integrations across your existing stack." Every word is accurate. Nothing in it tells a reader whether it is for them.
The same product, situation first: "You shipped it. Nobody is using it. Not because it is bad, but because finding the people who need it is a full-time job nobody taught you." Then the change: "This does that job." Then, and only then, the mechanism.
Notice what the second version risks. It is specific enough to exclude people, and excluding people is what makes the remaining reader feel spoken to. The first version excludes nobody and therefore reaches nobody.
Write the sentence they would say
There is a test that cuts through most of this. Imagine the person who has your problem describing it to a friend, in a bar, with no product in the conversation at all. Write that sentence down. Do not improve it, do not make it snappier, and do not add your category name to it.
That sentence is your opening line. It is almost never the sentence a founder writes, because founders write from inside the product looking out, and the sentence you need is written from outside looking in.
A useful habit: keep a file of the exact phrasing people use when they describe the problem to you. Not paraphrased, verbatim. Support messages, sales calls, Reddit threads, the way someone complained about it on a podcast. Those sentences are worth more than any copy you will write yourself, because they are already proven to be how the thought occurs to someone.
A second source, and the better one if you have any customers at all: read what they wrote when they first contacted you. The first message someone sends is written before they have learned your vocabulary, which makes it the closest thing you have to how a stranger thinks about the problem.
Founders tend to discard this because customers describe the product imprecisely. That imprecision is the data. If five people all describe it wrongly in the same way, that is not five mistakes, it is the category your product actually lives in as far as the market is concerned.
Use the comparison you are avoiding
Engineers resist explaining a product by comparison, because the comparison is never exact and inexactness is uncomfortable. Buyers rely on comparison, because it is the fastest way to place an unfamiliar thing in a known category.
| Approach | What it buys | What it costs |
|---|---|---|
| "Like X, but for Y" | Instant category placement | You inherit X's limitations in their head |
| "The X for people who cannot afford X" | Clear position and price expectation | Anchors you as the cheap option |
| Name the job, not the category | Works when the category is crowded | Slower; needs the situation line to land first |
| Invent a new category | No competitors in the frame | You pay to teach the market, which is expensive and slow |
| Describe the mechanism | Credibility with technical buyers | Almost nobody else finishes the sentence |
Inventing a category is the most tempting and the most expensive. It is occasionally right, and it commits you to explaining an unfamiliar idea to every single person you ever talk to, forever. If you do it, make sure the situation line is doing the work, because the category name will not.
One practical note on the second row of that table. Positioning yourself as the affordable version of a known tool works, and it is a decision about your ceiling as much as your floor: you will be compared on price forever, and moving upmarket later means fighting a reputation you built deliberately. It is a reasonable opening position and an uncomfortable permanent one.
The repeat-back test
Here is the only test of an explanation that matters. Explain it once to someone who has the problem. Change the subject for two minutes. Then ask them to describe it to you as though you were a colleague of theirs.
If they can, your explanation works and word of mouth is possible. If they reach for your words and cannot find them, it does not, and no amount of traffic will fix it, because every visitor arrives with the same fifteen seconds of patience the tester had.
Most explanations fail this test the first time. That is normal and is the point of running it: it converts a vague feeling that the pitch is not landing into a specific sentence that did not survive.
There is a weaker version worth running when you cannot get a real tester: read your own explanation aloud and stop at the first sentence that would make a busy person put their phone down. Almost always it is the sentence where you started describing capability. Delete everything from there to the end and see whether the remainder still says something.
A related habit is to watch where people interrupt you. An interruption is not rudeness, it is the moment they got enough information to form a question, and it tells you which sentence was doing the work. If nobody ever interrupts, nothing has landed yet and you are being heard politely.
Where this advice is wrong
Your buyer is an engineer evaluating infrastructure
If you are selling a database, a build tool, or anything an engineer will run in production, the mechanism is the value and the technical detail is the pitch. Leading with outcome language reads as marketing, which in that audience is a reason to distrust you. Lead with what it does, precisely, and put the benchmark above the fold.
You genuinely have a new category
Sometimes the comparison does not exist, and forcing one costs you more than it buys. The tell is whether informed people consistently misunderstand you in the same direction after hearing the comparison. If they do, the comparison is doing damage and you are better off naming the job and accepting slower recognition.
The problem is not one people admit to
Some problems are embarrassing, and naming the situation plainly makes people defensive rather than interested. "You cannot sell" is one of them. The fix is not to hide the problem but to locate it outside the person: it was never your job, nobody taught you this, it is a different skill. Same recognition, no accusation.
How this changes what you write
Once the explanation is right, most copy problems solve themselves, because almost every page is a variation on the same four moves.
- The headline is the situation or the change, never the mechanism.
- The first paragraph earns the second by being about the reader, not about you.
- Features appear after the reader has a reason to care what the features do.
- The call to action names what happens next in concrete terms, not what the reader should feel.
The reason to fix the explanation before the traffic is arithmetic. Doubling your visitors doubles the number of people who bounce off a confusing page. Fixing the page changes the fraction, and the fraction compounds across every visitor you ever get, including the ones you already paid for.
If the harder problem is that nobody is arriving at all, that is a different diagnosis and worth separating cleanly from this one. We wrote about telling them apart in distribution problem or product problem.
One warning about applying this. Fixing positioning is not a copywriting pass, and it will not survive being delegated to one. If the change is real, it shows up in which features you build next, which customers you chase, and what price is credible, because those all follow from the decision about who this is for.
A page rewritten without that underlying decision reads better for a month and then drifts back, because each new sentence gets written from inside the product again. The sentence you tested is the artefact worth keeping; the page is just where it currently lives.
Key takeaways
- A feature list answers how it works. Nobody asks that until they have decided they want it.
- Name the reader's situation first: recognition buys you the attention that explanation needs.
- Your explanation is not a description of the product, it is a description of the change.
- If a listener cannot repeat it to a colleague, it does not work, and word of mouth cannot happen.
- Comparison to something familiar is the fastest shortcut, and the one engineers most resist using.