follow-up

The Five Minute Window, and What the Research Actually Says

The lead response research is older and narrower than the slide decks suggest. Here is what it measured, what it did not, and what to do with it.

Every CRM sales page quotes the same number. Call a lead within five minutes and you are twenty-one times more likely to qualify it than if you wait thirty. The number is real, it comes from a real study, and almost everybody citing it has stopped reading one sentence too early.

What was measured

The finding comes from work by James Oldroyd, Kristina McElheran and David Elkington, written up in Harvard Business Review as The Short Life of Online Sales Leads. They looked at how long firms took to respond to web-generated leads and what happened next. Two things came out of it.

The first is the one everybody quotes: response time matters enormously, and the curve is steep near zero rather than gentle across the day. The difference between five minutes and thirty is not six times worse. It is far worse than that, because the lead is still at their desk, still on the page, still thinking about the thing they just asked about.

The second finding is the one that gets dropped, and it is the more useful half: most firms were nowhere near five minutes. A large share took more than a day, and a meaningful share never responded at all. The twenty-one-times figure is not a tip for optimising a good process. It is a measurement of how much money was sitting on the floor of an average one.

What it did not measure

It did not measure your trade. The study is about online sales leads broadly, and a roofing enquiry after a hailstorm, a dental appointment request and a software demo request are not the same purchase. Nobody has published the equivalent curve for a residential exterior contractor, and anybody who quotes one to three decimal places is quoting a sales deck.

It did not measure channel. Calling within five minutes and texting within five minutes are different acts with different rules attached to them, and the second one has a statute (see below).

It did not measure what happens on the second, fifth and ninth attempt, which for most field businesses is where the actual contact happens. The curve says be fast. It does not say be fast once.

So treat it as a direction with a very steep gradient near the origin, not as a formula.

The operational version

The useful translation into a working day is narrower and more boring than the statistic:

Know, within a minute, that the lead exists. Most slow responses are not decisions to wait. They are a form submission sitting in an inbox nobody is watching while the person who would have called is on a roof. The fix is a notification that reaches a phone, not a faster typist.

Have the first attempt belong to somebody. A lead that arrives to the whole team arrives to nobody. One name, one record, one next step with a time on it.

Make the second attempt automatic. First-attempt contact rates for inbound web leads are not high in any industry. The gain from speed evaporates if speed is followed by silence, and the cadence is the part a tool can actually own.

Write the attempt down where the next person can see it. Two calls from two people in ten minutes reads to the customer as disorganised at best. The record is what prevents it.

The part with a statute attached

Speed pushes people toward texting, because a text is faster to send than a call is to make. In the United States a text message to a mobile number is a call for the purposes of the Telephone Consumer Protection Act, and if it is sent with an automatic telephone dialing system or contains marketing, the consent requirements in 47 CFR 64.1200 apply.

The distinction that matters for a contractor is between a reply and a campaign. Someone who filled in your form and asked to be contacted about a roof has, in the ordinary case, given you the consent to contact them about that roof. That same person has not given you consent to be added to a monthly promotions list, and the written-consent standard for marketing messages is specific about what it requires: a clear and conspicuous disclosure, and agreement that is not a condition of purchase.

The practical rule is to keep the consent you actually have attached to the record, in the field, with a date on it, rather than in somebody's memory of the conversation. It costs nothing while it is unnecessary and it is the whole case if it ever stops being unnecessary. The FCC's own telemarketing and robocall material is the plain-language starting point; the statute is the authority.

What a good record looks like an hour later

If speed to lead is the input, the thing worth inspecting is the state of the record sixty minutes after the form arrived. Not the call log. The record.

It should say who the person is, what they asked about, what channel they asked through, who took it, what was attempted, what was said, and what happens next with a date on it. If the answer to the last one is blank, the lead is not worked, it is parked, and the five-minute call bought nothing.

That is the honest version of the statistic. Being fast is worth a great deal, and it is worth approximately nothing on its own.