Lead scoring built from your closed-won data

AED 9,000 to AED 45,000 covers most of what we are asked to build. The number moves on how much historical data there is to analyse, whether fit and engagement models are both built, whether routing and alerting are in scope, and whether ongoing quarterly validation is included. Anything quoted before those are known is a guess. We scope first, then price, and the scope document is yours whether or not you proceed. These are market ranges rather than a fixed rate card — the proposal carries the real figure.

No, and it is worth being clear because the term is misleading. That first page is Gumloop's seven best tools, HubSpot's product page, ZoomInfo's ten tools compared and several similar. Everyone searching it is shopping for software. You almost certainly already have scoring capability inside the CRM you own — what is usually missing is the model, which is what this page is about.

Assigning a number to each lead that estimates how likely they are to become a customer, so the sales team works the best ones first. The number comes from two things: how well the company fits what you sell, and what that person has actually done.

A B2B software company might score a lead higher for being in a target industry with over fifty staff, higher again for visiting the pricing page twice in a week, higher again for requesting a demo, and lower for using a free email address or for having gone quiet for sixty days. The rep sees a number and works the top of the list.

Grading usually describes fit — is this the kind of company we sell to, often expressed as A to D. Scoring usually describes engagement — how much interest have they shown, expressed as points. The distinction matters because they need different responses, and collapsing them into a single number is the most common modelling error we find.

You need a reasonable number of closed deals for the analysis to mean anything. Below roughly fifty closed-won in a comparable period, the patterns are not reliable and we would recommend simple routing rules instead and revisiting later. We will tell you which case you are in before quoting.

By comparing closed-won against closed-lost and seeing which attributes actually differ. A signal that appears equally in both predicts nothing regardless of how sensible it seems. Then we review the result with sales, because they sometimes know something the data does not show.

Three to five weeks including the analysis, the review with sales and the build. The validation happens a quarter later and is part of the engagement rather than an extra.

High-scoring leads should close at a materially better rate than low-scoring ones. We measure that after a quarter. If they do not, the model is wrong and we rebuild it.

Then either the model is not predicting or the threshold action is not built. Both are fixable and both are our problem rather than a discipline issue. Usually it is that crossing the threshold does not actually create a task for anyone.

If you have the volume, it can outperform a rules-based model. It also needs enough closed deals to learn from and it is harder to explain to a sales team, which affects trust. For most Dubai mid-market companies a well-built rules model is the better starting point.