From valuation to sales instruction: think in cohorts

A valuation request rarely leads to a sales instruction within the same quarter. The signed instruction often follows weeks, and sometimes more than a year, later. Counting by quarter makes that slow source score badly. So group leads into cohorts by the month or quarter in which they came in, and track how many sales instructions each cohort produces.

A valuation is orientation, not a decision to sell

A valuation request mainly tells you that someone is thinking about selling. It is not an instruction to sell. One applicant wants to know whether a larger home is within reach. Another is waiting on a divorce, an inheritance or a new-build project that will not complete for a year. A third just wants to know their home equity for a renovation.

As a result, the gap between the request and a signed agency agreement varies widely. Some sign within a month. Others only sign after twelve to eighteen months, or later still.

Valuations are widely offered. In our own measurement of 4,025 estate agency websites (23 August 2026), 44.4% offered a valuation on their site. The real share is probably higher, because we only looked at what agencies show on their website. Many agencies therefore bring in leads of this kind. The question is how you measure what they deliver.

Quarterly totals penalise the slow source

A quarterly total sets this quarter's sales instructions against this quarter's leads. For a fast source that works fine. For a valuation, you end up comparing two groups that have little to do with each other.

Take a sales enquiry through the contact form. Whoever fills it in usually wants to sell now. The sales instruction follows within weeks, so leads and sales instructions largely fall in the same quarter. With valuations, some of this quarter's sales instructions come from leads of months or a year ago.

That produces two distortions:

• If you start a new valuation campaign, the number of leads rises immediately while sales instructions lag for months. Conversion appears to collapse.
• If you stop the campaign, leads drop at once while sales instructions from old leads keep coming in. Conversion suddenly looks excellent.

Either way, you draw the wrong conclusion about the same source.

How a cohort works

A cohort is the group of leads that came in during the same period. You follow that group over time and count again each time how many of them have since signed a sales instruction. That way you compare leads only with their own sales instructions.

The table below is a worked example. The figures are invented to illustrate the method and do not come from an actual agency.

The numbers are cumulative. After three months, the valuation looks more than seven times weaker than the sales enquiry. After eighteen months, in this example, it delivers 20 sales instructions against 13.

Convert that into money and the difference becomes concrete. In 2025 a selling agent cost €5,800 on average (Krib, analysis of 23,000 quotes). At that amount, the valuation cohort represents €116,000 in commission after eighteen months, against €75,400 for the sales enquiry.

Read the curve, not the final number

A cohort also tells you how quickly a source matures. You see that in the shape of the series, not just in the last number.

A steep curve that flattens early, like the sales enquiry, calls for speed. If you do not respond within days, you lose the sales instruction to another agency. A flat curve that keeps rising for a long time, like the valuation, calls for patience and for contact that continues for months.

That has consequences for follow-up. A valuation lead closed as lost after two call attempts never appears in the figures again. Yet it is precisely those slow leads that can produce sales instructions with little competition, because other agencies have already dropped out by then.

Two rules for reading them:

• Compare cohorts at the same age. A cohort six months old next to one eighteen months old tells you nothing.
• Hold off judging a slow source until the oldest cohort is at least twelve months old.

First and last contact side by side

A lead that signs after a year has often come through several channels in that time. The valuation was the first contact. After that there may have been a newsletter, a viewing of another house and a phone call to the branch.

If you count only the last contact, you give the sales instruction to that phone call. The valuation that started the relationship then gets nothing. If you count only the first contact, you miss which channel tipped the balance.

You do not need a complicated allocation model. Show the first and last source side by side for each sales instruction. Then you see which sources open relationships and which close them, and with valuations those can be different channels.

What you record for each lead

A cohort analysis is only as good as the data beneath it. Record at least this for each lead:

Two pitfalls. If the same person comes in twice, through a valuation and later through a form, count them once, with the first date. Otherwise one sales instruction counts in two cohorts.

The second pitfall lies in the link with your agency software. Realworks has APIs for, among other things, contacts, properties and leads, but no API for the sales instruction or the commission. So you have to link the date of the agency agreement to the original lead within your own process. How to record the source itself reliably is covered in which lead source wins the most sales instructions.

From cohort to quarterly meeting

Cohorts do not replace the quarterly report; they complement it. The quarterly report shows what is coming in and signing now. The cohort table shows what last year's leads were ultimately worth.

So add one line per source to the report: the conversion of the cohort from twelve months ago. A layout for the rest of that report is in the quarterly report for a multi-branch agency.

KeyVue records the source of every lead and shows conversion by source, branch and agent. You build the cohort grouping on top of that yourself. If you want to see what that looks like with your own sources, book a demo.

Frequently asked questions