The First 90 Days Decide Whether a Medicare Advantage Sale Sticks

Search “why did my client cancel their Medicare Advantage plan” and you get a lot of correct but shallow answers: the beneficiary moved, a doctor left the network, a family member intervened, the plan changed formularies. All true. All mechanics of how a cancellation happens.

None of it answers the question an agency owner actually has: why do my lapses cluster the way they do? If nine of your last twelve early cancellations trace back to two lead sources and one agent, that is not bad luck. That is a pattern with a cause, and the cause is fixable in a way that “the client’s doctor retired” is not.

• •Bad-fit plan selection.The plan sold matched the commission schedule better than it matched the client’s doctors, pharmacy, or drug formulary. The mismatch usually surfaces the first time the beneficiary tries to use the plan — a rejected prescription, a doctor visit that turns out to be out-of-network. That is squarely inside the first 90 days.
• •Competitor churn-and-switch. A rival agency or call center specifically targets recently-enrolled beneficiaries during their own outbound windows, re-selling them into a different plan before the ink is dry on the first one. This is a volume game for the competitor and a chargeback for you.
• •A lead source that was never a fit.Some lead sources — certain aggregator feeds, certain call center hand-offs — produce volume but not qualified volume. The prospects convert on the call because the pitch is good, then reconsider once the enrollment kit arrives and the plan doesn’t match what they expected.

The mechanics-focused content treats these three as interchangeable “reasons clients leave.” They are not interchangeable — they point at three different fixes (retrain on plan fit, tighten your own outbound cadence, or cut a lead source), and you cannot tell which one you have without looking at where the lapses come from.

A cancellation at month eleven is a persistency statistic. A cancellation at month two is a check you have to write back. Under the rapid disenrollment rule, a plan change inside the first three months of enrollment triggers full recovery of the compensation paid for that enrollment — not prorated, not partial.

That is the practical reason 90-day churn deserves more attention than a standard persistency dashboard gives it: every early lapse is a guaranteed clawback, and a cluster of them from one lead source or one agent is a recurring hit to cash flow, not a rounding error in your book of business. For the fuller mechanics of how chargebacks connect back to source, see the chargeback breakdown.

You do not need attribution software to start seeing the pattern. A spreadsheet with five columns, updated weekly, gets you most of the way:

1. 1Enrollment date and disenrollment date (or "still active" if none) for every policy written in the last two quarters.
2. 2Days between the two dates, flagged red if under 90.
3. 3Lead source for that enrollment — the actual source, not a generic "referral" catch-all.
4. 4Writing agent.
5. 5Plan type sold vs. the beneficiary’s stated network/formularly needs at time of sale, if you captured it.

Sort by days-to-lapse under 90, then group by lead source and by agent. Two columns of a spreadsheet turn a vague sense that “something feels off with that lead vendor” into a number you can act on.

Once you have the tracker, the useful comparison is not plan-type-to-plan-type. It is lead-source-to-lead-source, holding agent and plan type as close to constant as you can.

Most agencies only have the tools to see plan-type churn, which flattens all three of these into one undifferentiated number. Splitting by source and agent is what turns “our persistency is soft this quarter” into “drop Vendor B, retrain Agent C.”

• If the same lead source produces early lapses under multiple agents, it is a targeting problem with the source.
• If lapses are spread evenly across sources but concentrated under one agent, it is a plan-fit or sales-practice problem with that agent.
• If early lapses spike right after a known competitor recruiting period (often right after AEP-adjacent enrollments), it is a churn-and-switch problem, not a quality problem on your end.
• If lapses are evenly spread across both sources and agents, the plan mix itself — not the funnel — is the thing to review.

A 90-day lapse tracker is a start, but it only covers three months of a policy that might earn commission for years. The same discipline — tying every lead source through to what happened downstream — applies across the full life of a policy, not just the early window. That is the difference between tracking cost-per-lead and tracking which source actually produces policies that persist and renew; see the full persistency-by-source guide for how to extend the same logic across the whole policy lifetime.

ClaimFlow does this by tying every recorded, SOA-timestamped lead through policy to renewal and chargeback automatically, using multi-touch attribution instead of a single last-touch guess — so a 90-day lapse and the lead source behind it show up in the same view instead of two different systems.