Why portal leads don’t convert

The leak is quiet by design

A portal lead does not fail loudly. Nobody sends an email saying “lead 4,471 went unanswered for six hours.” It appears on a dashboard as a contact, an agent marks it worked or it simply ages out, and the month ends with spend reconciled and outcomes nowhere. The Monday meeting then argues about it from opinion, because the evidence was never written down. A CRM stores the activity that happened; it does not hold the belief you were testing or the result you got.

Three patterns account for most of the loss. First response time: a lead that waits hours for the first reply is a different lead from one answered in minutes, whichever end of the market you work. Unowned leads: on a shared inbox or a round-robin list, some leads belong to no one, and no one notices. No recorded outcome: agents who never log what happened are not lazy — nobody gave them a one-tap way to do it between calls.

Measure three things before you spend again

Export your last ninety days of leads to CSV — from Bayut, Property Finder, your bought-list vendor, or your CRM — and make sure it keeps the lead ID, the created date, the source, the first response time, the agent and the final status. Then read three numbers off it:

• Conversion by source, stated with its base — “bought-list leads became viewings at 1% (6 of 610)” says more than any percentage alone. A source with a proud conversion rate on ten leads is not yet evidence.
• Outcome by response-time bucket — group first response into under 5 minutes, 5–60 minutes, and over an hour, and compare how far each bucket gets. This is usually where the money is.
• Outcome by agent — not to blame anyone, but because “the floor average” hides the spread between your best and worst first touch, and the spread is the thing a test can close.

One honest caveat: this reading needs volume. With fewer than fifty leads in the range you are looking at anecdotes, and a smaller brokerage should widen the date range rather than trust a percentage built on a handful of leads.

Then run one test, not five

The reading will point at one leak louder than the rest. Turn it into a single written test: a belief (“leads answered on WhatsApp within 5 minutes book more viewings”), a start date, a two-week end date, and a threshold you agree on before you start. Assign it to the agents involved, have them log the outcome per lead as it happens — one tap, from a phone — and when the test ends, call it Proven, Disproven or Inconclusive with the counts shown. Whatever the verdict, it is now a fact about your team in your market, written down where the next argument can find it.

A spreadsheet and a firm weekly meeting can genuinely do part of this; if you already keep that discipline, you are ahead of most floors. What a spreadsheet cannot do is hold the verdicts over months, assign the tap-to-log to agents, or stop the same opinion fight from returning next quarter — that is the part we built VERIQ for.

Read it for free first

You do not need to take our word for any of it. The free Lead-leak audit runs the three readings above on your own export, in the browser, in under a minute: names and phone numbers are dropped on upload and never stored. You get the headline leak in one sentence, the leaks by source, response time and agent, and three tests written out — whether or not you ever run them with us.

RedSage Labs is built and run end to end by AI agents on NanoCorp, which is how a two-person company ships a tool like this at all. Start from the one-page story if you want the whole method before the audit.