Lead Scoring.

The purpose is ranking, not weeding out. When more enquiries come in than can be handled carefully, without a model it is the order of arrival or gut feeling that decides who gets a call first. Lead scoring replaces this with a transparent rule. The emphasis on transparent matters: a model whose scoring nobody can explain is ignored by sales, and an ignored model is worse than none at all, because it also creates maintenance work.

A simple model has four components. Fit points for characteristics that match the offer: industry, size, region, the person's function. Behaviour points for actions that indicate buying interest: viewing the pricing page, repeat visits, downloading a quote document, replying to an email. Negative points for signals that point the other way: unsubscribing from the newsletter, visiting the careers page, a competitor's address, a private free-mail address in B2B. And decay, which reduces points again after a few weeks of inactivity, so that old enthusiasm does not keep counting for ever.

An example: a consultancy receives around sixty enquiries a month and can handle about twenty of them thoroughly. Instead of working through them in order, it awards points for company size, canton, the role of the person enquiring and a visit to the fees page; job applications and enquiries outside its offer receive negative points. The twenty highest scores are contacted first, and the rest receive a quick, standardised reply. After three months it checks whether the orders actually won were at the top of the list. This review is exactly what turns a guess into a model.

Lead scoring is the wrong investment when the feedback loop is missing. If you do not record which contacts eventually became orders, you can never correct the weights; the points then remain a formalised opinion. Nor is it worthwhile at low volumes: with a handful of enquiries a month, a careful manual review is quicker and more accurate than any rulebook. Finally, a model needs maintenance. If the offer or the target group changes, the weights quietly go out of date; a review date belongs in it from the start.