Why do people say they will buy and then not?

A customer asks for a quote. You send it. The price is not questioned. Then nothing happens, and when you follow up they are polite and still do not buy.

The usual conclusion is that the price was too high. It is usually the wrong conclusion, and acting on it — by discounting — makes the next one worse.

The reason is one of fifteen, and price is only part of two

Fifteen purchase drivers: pain, frustration, desire, fear, aspiration, objection, urgency, trust, social proof, status, convenience, saving, revenue, safety, belonging.

Most small-business marketing addresses saving and desire — cheaper and nicer — and leaves the other thirteen untouched. Meanwhile the thing that actually stopped the sale was often:

• Trust. They had no way to tell whether you would turn up.
• Safety. Nobody told them what happens if it goes wrong.
• Urgency, absent. Nothing made this month different from next month, so it became next month, repeatedly.
• Objection, unanswered. One specific worry nobody addressed, usually because nobody asked.

None of those are fixed by a discount. A discount on an untrusted supplier makes them more suspicious, not less.

Read off text, not guessed from an industry

The drivers are detected in text you supply: reviews, enquiries, support messages, the notes from sales calls.

That is the difference between this and the persona exercise every agency runs. A persona is assembled from assumptions about your market and then quoted back at you for two years. Reading your own customers' words tells you which drivers are actually live in your market this year.

And the rule is strict: it never fabricates a customer feeling. A driver not present in your text is not reported as present, and every score is labelled an estimate. If you supply three reviews it reports on three reviews and does not promote that to a finding about your customer base.

That is a smaller answer than a persona deck, and it is the one that is true.

From a driver to something you can publish

A target driver becomes a brief with four parts: angle, hook, proof, call to action.

The proof slot is the one that matters and the one most copy skips. If the driver is trust, the proof is not an adjective — it is the number of years, the named guarantee, the photographs of finished work, the fact that you answer the phone. If the driver is safety, the proof is what happens when something goes wrong, stated.

Which is also why this pairs with the claims check: the proof slot is exactly where an unsubstantiated superlative would otherwise go, and an unsubstantiated claim does not publish.

The cheapest version of this exercise

You do not need software to start. Take your last twenty enquiries that did not convert and write down, in the customer's words, the last thing they said.

Then count. If six of them mention timing, your problem is urgency. If four ask whether you are insured, your problem is safety. That is the whole method, and the platform's feedback clustering does it at the scale where counting by hand stops working.

The question to ask

Not "was the price too high". Ask: what did they need to be true that I never told them?

It is nearly always something you could have put on the page.