144,000 Agents Went Looking. Your Tech Stack Was on the Ballot.

Brokerage leaders describe agent churn as a recruiting problem. It is measured like one, staffed like one, and budgeted like one. There is a person whose whole job is bringing agents in, and usually nobody whose whole job is keeping them.

The volume argues for a rethink.

Per Mike DelPrete's analysis of the twelve months from June 2023 to June 2024, roughly 10% of all agents changed brokerages. That is about 144,000 people. Strip out the non-producers and the rate climbs to 14% of active agents. One in seven of the people actually writing deals packed up and went somewhere else inside a year.

Recruiting is not the counterweight to that. It is the treadmill it puts you on.

The churn is not where you think

The instinct is to picture the churn as dead weight leaving. New licensees who never sold anything, drifting off to another logo.

Partly true, and the data is more pointed than that. DelPrete finds the highest-risk group is agents at 12 to 23 months of tenure, alongside lower producers in the $1 million to $10 million range. Those are not quitters. Those are people who made it past the terrifying first year, learned the job, and then chose a different sign for year two.

And it does not stop at the bottom. Even top producers at $50 million and above churn above 10%. At that level a single departure is a revenue event with a name attached.

The most uncomfortable finding is structural. Per DelPrete, agents at large offices of 500 or more are about 33% more likely to leave than agents at small ones. The offices with the most resources lose people fastest. Scale, by itself, is not stickiness. It is often the opposite.

What an agent is actually deciding

Nobody leaves a brokerage over a split alone. Splits get compared, then rationalized. What actually moves someone is a quiet accumulation over the second year: leads that arrive and go nowhere, a CRM they were trained on once and now avoid, transaction paperwork that eats the evenings they meant to spend prospecting.

They do not describe it as a tech problem. They describe it as "I wasn't growing there."

That is the same sentence. An agent's growth is a function of how many opportunities reach them and how much of their week survives the administrative tax. Both of those are stack decisions the brokerage makes on the agent's behalf.

Which is why AI enablement is badly framed when it gets pitched internally as efficiency. Efficiency is a house metric. It sounds like cost-cutting, and agents hear it as something being taken away.

The retention frame is the honest one. If the brokerage's systems demonstrably grow an agent's business, the value stack gets harder to walk away from. If they do not, the competitor's recruiter has an easy story to tell, and they will tell it right at the 18-month mark.

The 18-month window

Here is the practical read for anyone running a brokerage.

You already know who is at 12 to 23 months of tenure. That list exists. It takes about ten minutes to pull.

For that specific cohort, the questions worth answering are narrow. How many leads reached them last quarter, and how fast. How many of those got a response inside five minutes. How many hours did they spend on transaction coordination that a system could have absorbed. And has anyone asked them, out loud, whether they feel like they are growing.

Answer those and you will usually find the churn risk clustered in a handful of names, driven by two or three fixable mechanics. That is a far cheaper problem than replacing them.

Recruiting fills the seat. Enablement is what keeps the seat from opening again eighteen months later.

Track A note: soft CTA below for the WholesaleAI version. Strip for any non-business placement.

An AI Operations Audit maps where your agents lose hours and where leads die before reaching them, cohort by cohort. Book one before the next round of recruiter calls lands.