Pipeline · 7 min read

The pattern is so consistent I can predict it. An agent gets busy, stops prospecting, closes a run of deals, and then hits a dead quarter. Panic sets in. They buy leads. The leads are expensive and cold, they work them badly because they're rattled, and they conclude the leads were junk. Then they get busy again, and the cycle restarts.

Nothing in that story is a lead source problem. It's a cadence problem. And cadence is the only thing that makes a lead source compound instead of evaporate.

A compounding source has three properties: it gets cheaper per deal over time, it improves the more you do it, and it produces business while you're busy with other business. Judge every channel against those three. Most fail. Four don't.

1. Past clients — the one everybody claims and nobody works

Ask an agent where their business comes from and they'll say referrals and repeat clients. Ask them what they did for a past client last month and you'll usually get silence.

This is the highest-return activity in the business by a wide margin. Someone who has already trusted you with the biggest transaction of their life will do it again and will tell their friends — if you're still a presence in their life when the moment arrives. And the moment is unpredictable, which is exactly why the contact has to be routine.

What routine looks like: a real conversation twice a year, something useful about their specific property once or twice more, and a note on the purchase anniversary that isn't a sales pitch. Four touches. Per year. Per client. Nearly every agent underperforms this and then wonders why their referral count is flat.

2. Your sphere, worked like a database instead of a feeling

Your sphere is not "people who know me." It's a defined list of names in a system with a date next to each one. If it lives in your head, it isn't a lead source — it's a fond hope.

The discipline is unglamorous: write the list down, segment it by how likely they are to move or refer, and hold yourself to a weekly number of genuine conversations. Not messages. Conversations. Ten a week is 500 a year and it will change your business, which is why almost nobody does it consistently for a full twelve months.

3. A geographic or niche farm you commit to for three years

Farming compounds beautifully and fails constantly, for one reason: agents quit at month eight. The math of a farm is brutally simple — you spend for two years and you harvest in years three through ten. Anyone who stops in year two paid full price for nothing.

Pick something small enough to own. Three hundred homes you dominate beats three thousand you sprinkle. And it doesn't have to be geography; a niche works the same way — a specific community, a property type, a profession you understand from the inside. Own something narrow. Be the obvious answer inside it.

4. Published expertise

This is the one that changed most in the last few years. Whatever you publish — video, a local market letter, a podcast, written answers to the questions buyers ask you — it works while you sleep, and now it is also what AI assistants read when somebody asks for an agent who knows your market.

It's slow. It compounds harder than anything else on this list. And the bar is not production quality; it's usefulness and consistency. One honest, specific piece a month for two years will outperform a professionally produced series you abandon in March.

A lead source you work in bursts isn't a source. It's a rescue attempt.

The 90-day cadence

Here's the version I give agents to leave with. It takes about seven hours a week and it does not care whether you are busy, because the whole point is that it survives being busy.

Daily, one hour, before anything else. Conversations — sphere and past clients. Before email, before the transaction fires, before the day takes the hour away from you. The hour is non-negotiable and it goes first because by 2pm it will not exist.

Weekly, two hours. One block for the farm or niche, one for publishing. Same day, same time, on the calendar like a listing appointment.

Monthly, one hour. Look at the numbers: conversations had, appointments set, sources credited. Not to judge yourself — to see which channel is actually paying, since almost every agent is wrong about this.

Quarterly, half a day. Cut what isn't working, and recommit to what is for another ninety days.

That's it. It's not clever. Its entire power is that you can still do it during your busiest month, which is precisely when everyone else stops — and why the agents who hold the rhythm never seem to have a dead quarter.