System 4 of 6
Service & Persistency
Keep what the agency earns.
Service & Persistency is the operating system that keeps what the agency earns. It measures service demand, backlog, onboarding completion, client-fit problems, expectation gaps, retention, and persistency — with clearly defined denominators and time periods — and exposes when service capacity, not sales, is the real limit on growth.
Who feels it, and how
- New business is celebrated while lapses and cancellations leave quietly, uncounted.
- Service requests live in individual inboxes with no shared queue or backlog number.
- Onboarding is a welcome email at best; the first service contact is a problem call.
- Persistency is “probably fine” — no one has computed it with a real denominator.
- Growth pushes service response times up, which pushes retention down, invisibly.
The mechanism
Retention failures are usually created early — at fit, expectation-setting, and onboarding — and only harvested later as lapses, so the cause and the symptom are separated by months. Meanwhile service capacity is a hard constraint: when demand exceeds it, response times stretch, small issues escalate, and the team triages by loudest-voice instead of by impact. Because none of this appears on a production report, an agency can grow its way into shrinking.
What to measure
Persistency / retention rate. Policies (or clients) still active at the end of a period ÷ those active at the start plus additions, for a defined cohort and time window. Write the denominator down — most persistency arguments are secretly denominator arguments.
Service demand vs. capacity. Requests arriving per week vs. requests resolved per week, plus current backlog age. Three numbers, one glance: is the queue growing?
Onboarding completion. Share of new clients in the last 90 days who received every step of the defined onboarding sequence. If there is no defined sequence, that is the finding.
Common misreads
- Blaming lapses on price or competitors before checking onboarding and expectation gaps — the causes the agency controls.
- Treating a quiet client as a satisfied client. Silence is absence of evidence, not evidence of retention.
- Averaging response times across easy and hard requests, which hides the backlog where relationships are lost.
Countermeasures to test
- Compute persistency once, honestly, with a written denominator — then track the same formula every period.
- Create a single shared service queue so demand, capacity, and backlog age become visible numbers.
- Define a minimum onboarding sequence and measure its completion rate like a production number.
- Interview or review a sample of recent lapses for cause before buying any retention tool.
A worked example
An agency computes twelve-month persistency for the first time with an explicit denominator and finds it lower than assumed. A review of recent lapses shows a cluster in the first policy year, concentrated among clients who never completed onboarding. The countermeasure is one defined onboarding sequence with a completion metric and a named owner — reviewed monthly — rather than a discount program aimed at the wrong cause.
If the data is missing
If persistency has never been computed with a real denominator, label it Unknown rather than assuming it is fine. The first countermeasure is the calculation itself — it takes an afternoon and frequently changes the agency's priorities.
Frequently asked
Is persistency the same as retention?
They are related but not interchangeable. Persistency usually tracks policies remaining in force over a period; retention usually tracks client relationships. An agency can retain the client and still lose the policy, or vice versa. Define both, with denominators, and track them separately.
What if service is drowning right now?
Measure demand versus capacity for two weeks before hiring. Some backlogs are capacity problems; many are focus problems (poor-fit segments generating outsized load) or leverage problems (no procedures, so everything is custom). The fix differs by cause.