Broker Lead Lab · 2026-10-22 · OshyLabs

Which lead sources actually win instructions? A CRM export method for estate agents

Count signed instructions and listings, not enquiries. Export your leads from your CRM, group them by source, and compare how many became a valuation, a signed instruction or listing, and a sale. The source that sends the most enquiries is not always the one that wins the most business, and you only see the difference when you track outcomes.

This guide is for UK estate agency owners and managers, and US brokerage owners and team leads, who pay for portals, ads and referrals and cannot say which ones win business.

Want the report without building it? Broker Lead Lab reads the CSV your CRM already exports.

Why do enquiry counts mislead?

Imagine a portal that sends 200 enquiries and wins 2 instructions, and a referral partner that sends 5 and wins 3. Counting enquiries backs the portal. Counting instructions backs the partner. These numbers are invented to show the point, but the pattern is common enough that every agency should check its own.

Enquiry counts reward whichever channel is cheapest to feed. Instruction counts reward whichever channel pays for itself.

What do you need to export from your CRM?

One row per lead, with these columns:

  • Lead ID and the date it was created.
  • Source, as your team recorded it.
  • Campaign, or the portal listing or ad it came from.
  • Assigned agent.
  • Dates for each stage that applies: valuation booked, valuation done, instruction or listing signed, sale agreed or under offer, exchanged or closed.
  • Outcome, and the lost reason if you record one.

Spend by source usually lives in your finance records, not the CRM. Add it as a second table.

How do you work out which source wins?

  1. Clean the source names. "Rightmove", "rightmove.com" and "RM" are one source. Merge them before you count anything.
  2. Group by the month the lead was created. An instruction can take weeks or months to arrive, so judge each month's leads as a cohort once they have had time to convert.
  3. Calculate, for each source: leads, valuations booked, instructions or listings signed, the instruction rate (signed divided by leads) and the median days from enquiry to signing.
  4. Add spend. Cost per instruction or listing is the number that decides budget.
  5. Split by agent. If a source converts well for one agent and badly for another, the issue is follow up, not the source.

A worked example. The numbers are invented for illustration.

SourceLeadsValuationsInstructionsInstruction rate
Portal A2001821.0%
Website valuation form409410.0%
Walk in3010516.7%
Referral partner54360.0%

Five referrals is a tiny sample. Treat 60% as a pointer, not a verdict, and keep tracking.

What skews the numbers?

  • Blank or "other" sources. If a fifth of your leads have no source, you are guessing about the rest. Make the field required, with a short pick list.
  • The same person through two channels. Decide whether you credit the first source or the last, and apply the rule to every lead.
  • Time lag. Recent leads look worse than they are because they have not had time to convert. Compare cohorts of the same age.
  • Stages nobody updates. If agents do not log a valuation, the CRM will say it never happened.
  • Sales and lettings mixed together. They convert differently. Split them.

How do you test a change without fooling yourself?

Change one thing at a time. A portal package, an ad budget, a response time target. Give the change a fixed window, at least one full enquiry to instruction cycle by your own median, and compare cohorts of the same age rather than week against week.

Two changes at once and you will not know which one moved the number.

Common objections

"I do not have time for spreadsheets." Cleaning the source names is the work. Once the columns are clean, the pivot table is quick, and you only clean once.

"My CRM already has reports." Check whether yours joins source to the instruction or listing stage. If it does, use it. If it only counts enquiries, export the CSV.

Frequently asked questions

What is a good enquiry to instruction rate?

There is no honest universal figure. It depends on the market, the source and how your agency defines an enquiry. Measure your own baseline first, then improve against it.

How many leads do I need before I trust a source's numbers?

Small samples swing wildly. I would wait until a source has a few dozen leads before acting on its rate, and I would not cut a channel on five.

What is the US equivalent of an instruction?

A signed listing agreement. The method is the same: track each lead from source to signed listing, then to a closed sale.

What if my CRM does not record the source?

Start recording it today. A required field with a short pick list is enough. You will have usable data after one full cycle.

Next step

Broker Lead Lab imports the CSV your CRM already exports and shows which lead sources, campaigns and agents turn into signed instructions or listings. Then you test one change at a time. Visit Broker Lead Lab