Step 3 of 5 on the chain

Propensity modelling

Also called Propensity scoring, Likelihood to convert.

Estimating how likely a lead is to do something: become qualified, reach an opportunity, close. The output is a probability rather than an amount.

Why it matters

It is the other half of a value, and it is often mistaken for the whole thing. Value is probability multiplied by size, so a propensity score on its own tells a bid strategy which leads are likely and nothing about which are worth having.

What goes wrong

Two leads equally likely to become customers can be worth ten times different amounts, and bidding on propensity alone buys the cheap certain one over the expensive probable one every time. That is the right answer if you are routing a sales team and the wrong one if you are spending a budget. The confusion is easy to make because the good version of a value model contains a propensity model: close rate is a propensity, and expected value is that probability times what the deal was worth when it landed. Google's own Journey-Aware Bidding is propensity modelling, and it says so; it models whether a lead will progress and leaves what the lead is worth to you.

What this product does about it

Close rate per cohort is the propensity half, measured from resolved deals, and it is multiplied by the average won amount rather than reported on its own.

How to check it on your own account

Ask what your scoring model would do with a lead that closes 5% of the time at £200,000 against one that closes 60% of the time at £900. If it prefers the second, it is a propensity model being used as a value model.

Next to this

The rest of this step

Measure this on your own data

The diagnostic reads a CRM export in your browser and reports your volume, your match rate and the spread between your leads against the thresholds in this glossary. Nothing is uploaded and no account is needed.

See what your own leads are worth

Read your closed deals and find out whether your lead values actually vary, and by how much. Nothing is stored, and your file is read in your browser.

Try it on a sample dataset