Stop Tracking Cost-Per-Lead. Start Tracking Persistency-Per-Source.

Every lead vendor pitch leads with cost-per-lead. It is easy to measure, easy to compare across vendors, and easy to put in a spreadsheet column. It is also the wrong number, because it stops counting the moment the sale closes.

A Medicare Advantage policy pays initial commission at issue, then renewal commission every year the policy stays active — assuming it does not lapse early and trigger a chargeback. Two leads with identical cost-per-lead can produce wildly different lifetime value if one channel systematically attracts clients who disenroll in month four and the other attracts clients who stay for five years. Cost-per-lead cannot see that difference. Only persistency can.

Search "lead source ROI tracking" and you get generic B2B marketing-attribution content — first-touch versus last-touch, multi-touch models, dashboards built for SaaS demo requests. Search "chargeback" or "persistency" and you land on FMO (field marketing organization) blogs written for compliance and back-office teams, full of carrier-specific chargeback schedules and none of it connected to where the lead came from in the first place.

Neither world addresses the actual question an agency owner has: which lead source, by name, produces the policies that stick — and which one is quietly funding a chargeback problem three months from now. That gap is the reason agents end up building this tracking by hand.

Persistency-by-source is a simple idea executed rarely: tag every policy with the lead source that produced it, then follow that policy's status over time — active, lapsed, chargeback — segmented by that same source. The output is not a lead-quality score. It is a renewal-commission forecast, broken out by channel.

• •Source tag — internet aggregator, T65 direct mail, referral, inbound call center, social lead form, and so on, attached at the point of first contact.
• •Policy status over time — active, lapsed, or charged back, checked at regular intervals (90 days, 1 year, each renewal date), not just at the point of sale.
• •Renewal commission compounded — the commission a policy pays in year one is a fraction of what it pays if it renews for three or five years. Persistency is what decides which fraction you actually collect.

This is a different question than which touchpoint gets credit for the sale. It is asking what happens after credit is assigned — whether the source that got credit for the sale is also the source quietly producing your worst retention.

The numbers below are a simplified illustration to show the mechanic, not a published industry benchmark. Use your own historical lapse rates and commission schedule to run this for real.

On cost-per-lead alone, the internet channel looks like the better buy — a fraction of the acquisition cost per policy. But if that channel's policies lapse at nearly double the rate of the direct-mail channel, the direct-mail leads are paying renewal commission on twice as many policies three years out. The cheaper lead source can be the more expensive channel once you price in what it does to persistency. You will not see that in a cost-per-lead report. You only see it once persistency is tracked by source and carried forward across renewal years.

Search Medicare agent forums and Facebook groups and you will find agents doing exactly this analysis manually — pulling a carrier commission statement, cross-referencing it against a CRM export tagged by lead source, and building a pivot table to see which vendor's leads are still paying a year later. It works, until the agency scales past a few hundred policies or adds a second lead vendor and the manual match breaks down.

The reason this gets built by hand instead of bought off the shelf is that no Medicare-specific tool connects lead source to commission and renewal status in one place. General attribution software has no concept of a chargeback. FMO chargeback reports have no concept of a lead source. The spreadsheet is what fills the gap — right up until it can't keep up with the volume.

You do not need a platform to start. You need three fields captured consistently, and a recurring check-in against them.

1. 1Tag every lead with its source at intake — vendor name, not just "internet" or "referral." Specificity here is what makes the analysis usable later.
2. 2At policy issue, carry that source tag forward onto the policy record, not just the lead record. Most CRMs drop the source tag once a lead converts — force it to persist.
3. 3On a monthly cadence, pull active/lapsed/chargeback status by policy and roll it up by source. Even a manual VLOOKUP against a carrier statement gets you a real answer.
4. 4Once you have three to six months of this, compare persistency curves by source against cost-per-lead. That comparison is the actual ROI number — not the acquisition cost alone.

This is manageable at low volume. It is also exactly the process that breaks down first — the source tag gets lost in the handoff from lead to policy, long before anyone realizes the persistency numbers being reported are wrong.

A lead source with poor persistency is not just a retention problem — it is a chargeback problem with a delay built in. Carriers claw back commission on policies that lapse early, and if one lead source is disproportionately represented in your early-lapse policies, that source is quietly generating the chargebacks that show up on next quarter's statement. The chargeback connection is the other half of this picture: once you can see persistency by source, you can also see which source to renegotiate, cap spend on, or drop before the next AEP cycle — rather than discovering it after the chargebacks land.