← All posts

Agent Retention Alerts: What to Watch and How to Act

By Phillip Gagnon · November 11, 2026 · 13 min read

Broker listens during an agent support check-in

An agent retention alert is a flag that tells a broker which agent to talk to this week, before that agent starts talking to another brokerage. Good alerts are built from data you already have: closings, listings, pending deals, meeting attendance. They are worth something only when a real conversation follows. This guide covers what to watch, what recent research says about it, and a simple routine you can start on Monday.

TL;DR

A retention alert is a prompt for a supportive conversation, not a verdict about an agent. Routine monitoring reports what happened. An alert says who needs you now.
The strongest signal in the research below is pipeline timing: months since the last closing and the number of active listings.
Failed listings are a weak signal. Do not build a program around them.
Treat every alert as a question to ask the agent, then fix the friction they name.
Tell agents what you track and why, and log who was flagged, who you spoke with, and who stayed.

What retention alerts are, and how they differ from routine monitoring
Why agents leave: the factors behind the alerts
Signals worth alerting on
From signal to first move
How to respond to an alert
A weekly routine you can start this week
Keep the process transparent and fair
How to tell whether your alerts are working
Perspective
Put alerts to work with 3 Data Pulse
FAQ
Sources

What retention alerts are, and how they differ from routine monitoring

Most brokerages already look at production reports. A monthly report tells you who closed what. It is backward-looking, covers everyone equally, and usually gets read after the damage is done.

A retention alert works differently in three ways:

It is selective. Out of every agent in the office, it surfaces the few whose pattern has changed.
It is timed. It fires while there is still time for a conversation to matter, not after a license transfer request lands.
It points to an action. The output is a name, the reason the name appeared, and a suggested first question.

The point of the system is the agent, not the dashboard. An alert should make you think “what does this agent need from us?” and never “how do I stop this agent from leaving?” That framing changes the conversation. Agents are independent businesspeople who choose where to hang their license, and they stay where their goals are supported.

Why agents leave: the factors behind the alerts

Before choosing signals, it helps to know what actually drives agents out. HousingWire’s October 2026 reporting says agents are more likely to leave over leadership, culture and support than a sales slowdown. Reported factors include leadership direction, protection and support for their business, professional development, compensation, technology, and culture.

An earlier HousingWire survey found that 43% of agents named culture, leadership and relationships as their main reason to stay, against 13% who named pay. The same report found that 87% of agents who saw a clear brokerage vision were happy, versus less than half of those who did not. Technology matters when agents compare brokerages (four in five rated it important), but only one in five named it as the main reason to switch. Training was rated critical by 70% of respondents.

What this means for alerts: the data tells you who has gone quiet, but the cause is usually a support gap. Your job is to find it.

Signals worth alerting on

Pipeline timing: the best-supported signal

A HousingWire study of more than 625,000 agents in MLS coverage areas found that switching risk rises with every month since an agent’s last closing and falls with every additional listing in their pipeline. The spread was wide:

Agents who had closed within the previous three months and had three or more active listings: 2.7% changed brokerages over the next year.
Agents who had gone at least six months without a closing and had no active listings: 15.6% switched, nearly six times higher.

That gives you a practical rule. An agent who has not closed in several months and has nothing listed is the first person to call, regardless of how large their past production was.

Failed listings: a weaker signal than you might expect

It is tempting to flag agents whose listings expired or were canceled. The same study found the opposite pattern. At every level of pipeline activity, agents who had listing failures were less likely to change brokerages than agents with none. Among agents with no active listings, 9.7% of those with no cancellations switched, compared with 8.0% of those with two or more.

An expired listing may be a good reason for a supportive call, because that agent probably wants help with pricing or marketing. But do not treat it as a flight warning.

Early-career agents and low production

A Relitix analysis reported by RealTrends looked at 3,489 agents who changed offices in Arizona during the first 11 months of 2020. It found that production level had little effect on the tendency to move, except for agents under $1 million a year, who moved at much higher rates. Agents in their second and third years faced dramatically higher risk. It is one state and one period, so use it as a prompt to look at your own cohorts, not as a rule.

The lesson: your top producers are not the only people who leave. Your newer agents often leave because they were not supported enough early on. Our agent onboarding checklist from pre-start to day 90 is a good place to tighten that window.

Engagement changes you can see yourself

Data is not the only source. Many brokers notice these before any report does:

An agent who used to attend sales meetings stops coming.
An agent who was active in team chats goes silent.
Questions about their database, listings or transfer process start appearing.
Their requests for help, marketing or transaction support drop off.

No single item proves anything. Together with the pipeline signals above, they tell you whose week to start with.

From signal to first move

Alert	What it may mean	First question to ask
No closing in 3 to 6 months, thin pipeline	Lead flow has stalled or confidence has dipped	“What part of your business feels hardest right now?”
No active listings, previously active	Listing appointments are not turning into signed agreements	“Where do your listing conversations tend to stall?”
Stopped attending meetings	The meetings may not be giving them value	“What would make our meetings worth your time?”
Second or third year agent, flat results	Needs structure beyond onboarding	“What do you want your business to look like in a year?”
Went quiet in team channels	Feels disconnected	“Who in the office do you work with most?”

Notice that every question is about the agent’s business, not the brokerage’s targets.

How to respond to an alert

The worst response is a call that sounds like an audit. HousingWire’s reporting on weekly one-on-ones advises asking what the agent dropped and where it fell, rather than just how many calls were made, then using a short weekly conversation to remove specific friction in their workflow.

Verify before you assume

An alert is a hypothesis. The agent may be on a family leave, handling a large deal that has not yet shown in the MLS, or working a referral business that does not appear in listings. A friendly check-in costs nothing and tells you which it is.

Hold a stay conversation

HousingWire’s October 2026 report describes stay interviews as an upstream alternative to exit interviews. The questions it recommends are worth using almost word for word:

What keeps you at the brokerage?
When did you last think about leaving, and what was happening then?
What would make your day easier?
What do you want your business to look like in a year?

The full piece is here. Ask one or two, then listen. Write down what you heard and what you promised.

Agree on one owned action

End every conversation with one specific thing, and a date. It might be a listing presentation role-play, an introduction to a transaction coordinator, a marketing piece, or a goal review. NAR’s guidance on top agents says they want accountability paired with support, documented commitments and reliable follow-up. Follow-up is where most brokers fall short, so put the next check-in in the calendar before you hang up.

If you need a structure for those conversations, our 20 to 30 minute one-on-one coaching scripts give you an agenda, and the quarterly one-on-one for top agents covers your highest producers. For agents who need practice between meetings, Soravel is our sister product: an AI coaching platform for real estate agents with voice or chat coaching sessions, daily check-ins, role-play practice and GCI goal tracking.

A weekly routine you can start this week

Monday, 15 minutes: Pull a list of agents with no closing in the last three months and no active listings. Add anyone you have noticed going quiet.
Pick three names. More than that and follow-through slips. Choose a mix of veteran and second- or third-year agents.
Reach out personally. A short call or a coffee invitation that is about them: “I was thinking about your business and wanted to hear how things are going.”
Ask one stay question and one friction question. Take notes.
Agree on one action with a date, and do your part first.
Friday, 10 minutes: Log who was flagged, who you spoke with, what you agreed, and what happened.

This takes about two hours a week. It is also the same discipline that makes recruiting work: agents who feel heard at your brokerage are the best advertisement when you are talking to agents elsewhere. Our guide to reducing agent turnover goes deeper on the surrounding practices.

Keep the process transparent and fair

Alerts only help if agents trust them. Some practices that keep the program healthy:

Tell agents what you look at and why. Say that you review production and activity patterns in order to offer support sooner.
Use the minimum data needed. Closings, listings and pipeline are enough. You do not need to read private conversations.
Separate evidence from interpretation. “No closing in five months” is a fact. “Planning to leave” is a guess. Say only the first one aloud.
Never attach penalties or favors to an alert. It exists to trigger support.
Check applicable state and local requirements with qualified counsel before you add new kinds of data to your program.

When the conversation feels like help, agents stay open. When it feels like surveillance, the alert becomes the reason they leave.

How to tell whether your alerts are working

Keep a simple log with five columns: agent, alert reason, date contacted, action agreed, and status twelve months later. Review it each quarter and look at:

Follow-through: Did you call every flagged agent within a week? Did you do what you promised?
Recurring friction: Do the same issues keep surfacing (lead flow, transaction support, marketing)? That is your next investment.
Voluntary departures by cohort and tenure: Are second- and third-year agents leaving less often than before?
Themes from agents who left: Remove names and look at the reasons.

Be careful with causation. If fewer agents leave, that does not prove the alerts did it. Use the log to learn which conversations help, and keep refining.

The stakes are real. According to a Courted report summarized by HousingWire, agents who left brokerages produced $2.11 million annually versus $1.44 million for the recruits who replaced them. Keeping an agent is often worth more than replacing one.

Perspective

Every broker we talk to knows what grows a brokerage: recruit steadily, keep good agents happy, and coach everyone up. The trouble is time. Nobody has an hour a day to watch for the agent who is slipping.

That is why we built retention alerts around the data a brokerage already has. Our retention prediction is designed to flag agents whose patterns are changing, then hand you a short list with the words already written. No new data entry, and no asking agents to fill anything out.

We believe agents stay where their goals are supported, not where the split is a point higher. So our alerts are designed to start a coaching conversation about the agent’s business, and the same data feeds the recruiting side: the support you give today is what you will talk about when you invite the next agent to a conversation.

Client

Put alerts to work with 3 Data Pulse

If you would rather not build this by hand every Monday, 3 Data Pulse retention software watches your brokerage data, predicts retention risk and gives you actionable alerts, alongside coaching materials and recruiting from MLS data. Pricing is usage-based: $1 per month for each agent you manage, $0.25 per month for each agent you are recruiting, with a $250 monthly minimum. See the details on the pricing page.

The same list that shows you who needs a conversation also shows which agents in your market are open to one. That is how retention and recruiting end up as one habit.

FAQ

What is an agent retention alert?

It is a notice that an agent’s pattern has changed in a way that suggests they may need support, such as months without a closing or an empty pipeline. It tells a broker or manager whom to contact this week. It does not predict a decision.

Which signal is most reliable?

In HousingWire’s study of more than 625,000 agents, time since last closing and the number of active listings were the clearest. Agents with six months without a closing and no listings switched at 15.6%, versus 2.7% for recently active agents with three or more listings. Failed listings were not a reliable signal.

What should I say when an alert fires?

Start with the agent’s business, not the number. Ask what they want their business to look like in a year, what would make their day easier, and where things have stalled. Agree on one action with a date, and follow up.

Should agents know they are being tracked?

Yes. Explain which activity data you review and that the purpose is to offer help sooner. Keep to the minimum data you need, and treat an alert as a reason for a conversation, never as a penalty.

How many alerts can a small brokerage handle?

Three conversations a week is realistic for most managers. Choose the agents whose pattern is clearest, and give each real follow-through rather than skimming a long list.