Home Insurance Panel Shrinkage: Bidding on ZIPs Nobody Will Quote

Beacon & Bay is an illustrative Tampa agency, invented for this article, and their August looked like a win right up until it didn't. They put $41,000 into homeowners traffic across nine Gulf Coast ZIPs. Cheap clicks. Form-fill rate above eight percent. Then the quote rate came in at eleven percent, and the reason was not lead quality in any sense a scrubbing vendor would recognise. The people were real. The houses were real. Nobody on the panel would write them. That is panel shrinkage doing its quiet work, and it hides from every metric a media buyer normally watches.

What Panel Shrinkage Actually Does to Your Funnel

In auto, a bad lead is usually a bad lead: wrong number, bot fill, someone who filled a form for a gift card. In home, the lead can be immaculate and still worthless to you, because the constraint sits on the supply side. If four of your six carriers have pulled appetite in a coastal county, a homeowner in that county generates a real inquiry and zero quotes.

The NAIC-sourced state data reported by Insurance.com puts Florida's calendar-year 2024 homeowners non-renewal rate at 3.35%, the highest in the country, with California second at 3.18%. Non-renewals are the visible edge of a wider retreat. Behind every non-renewal sits a book the carrier stopped growing months earlier, and that quiet freeze is where panel shrinkage starts eating your economics first.

Panel Shrinkage: Scoring Eligibility Before the Click

Most teams handle appetite after the fact. Lead lands, rater runs, nothing comes back, everyone shrugs and calls it a soft month. The fix is to move that verdict upstream, so appetite becomes a targeting input rather than a post-mortem. That is the whole discipline of planning around panel shrinkage: you are not buying homeowners, you are buying quotable homeowners.

Beacon & Bay rebuilt their plan around one question asked before any money moved. For this ZIP, this roof, this build year, how many carriers on my panel would return a bindable number? Call the answer the panel depth. Three or more is a healthy market. One is a coin flip. Zero means you are paying to generate a phone call you will end by apologising.

The Three Signals You Can Get Pre-Click

You do not need a full underwriting stack to do this. You need three things, all of which exist before a form is submitted.

Stack those three and you get an eligibility tier per ZIP. Beacon & Bay ended up with four tiers and a rule that no tier-four ZIP gets a live bid without a written exception from the agency principal.

A Bidding Model Built on Quotability

Here is the illustrative model they ran for one week in September, using round numbers to make the arithmetic legible. Every figure below is an example, not a benchmark.

Look at the cheapest column and the most expensive outcome. They are the same row. The $14 lead in a tier-four ZIP costs twenty-five times more per quotable household than the $34 lead inland. Any bidding model that ignores panel shrinkage and optimises to cost per lead will walk straight into that bottom row and keep walking, because the platform reports it as the efficient one.

When the Right Bid Is Zero

This is where the meeting gets tense. Somebody has spent two years building volume in a market, and you are proposing to switch it off. The honest framing helps: you are not abandoning the ZIP, you are declining to pay for inquiries you cannot serve. When a carrier re-enters, the bid comes back on the same day.

California is the clearest live example of that dynamic. The state's Sustainable Insurance Strategy lets carriers use catastrophe modelling and net cost of reinsurance in rate filings in exchange for writing commitments in distressed areas. If that framework does what it is designed to do, appetite in specific ZIPs returns in steps rather than all at once, and the advertisers holding a current panel-depth sheet will see it before the ones running on last year's assumptions.

Reporting That Catches Appetite Loss in Days

Appetite does not vanish with an announcement. It thins. Your reporting has to be sensitive enough to catch thinning, which means weekly, by ZIP, on quotable rate rather than on close rate.

• Quotable rate by ZIP, seven-day rolling, flagged when it drops more than eight points week over week.
• Carriers-returning-a-number distribution, so you can see depth erode from three to two before it hits one.
• Declination reason codes captured as structured data, not free text in a CRM note.
• Non-renewal cohort split out from shopper cohort in every source report.
• Panel-depth sheet dated and re-confirmed monthly with each carrier rep.

Beacon & Bay caught their second appetite contraction in nine days instead of six weeks. Same media budget, roughly $19,000 of waste avoided in the illustrative model. The difference was entirely reporting granularity.

What Panel Shrinkage Looks Like When It Goes Wrong

The failure has a signature, and once you have seen it you can name it in a Monday call. Cost per lead falls. Volume rises. Form completion rate improves, because the cheap coastal inventory converts well on the front end. Every top-line number in the platform looks like a team doing good work. Meanwhile the quote rate drifts down two points a week and nobody connects the two, because the platform has no column for whether a carrier would write the house.

Beacon & Bay's first contraction ran for six weeks before anyone said the words out loud. The post-mortem arithmetic was ugly in a way that is worth reproducing, because it is how panel shrinkage actually bills you.

Spend climbed from roughly $12,200 to $12,800 a week while quotable households fell by a third. Read the middle two columns alone and the buyer looks like a hero. That gap between the reported metric and the delivered outcome is the entire cost of not instrumenting panel shrinkage.

Two Objections Worth Answering

"Our carriers tell us when they pull back." They tell you when they formally suspend a state. They do not send an email when an underwriter quietly tightens roof-age tolerance from twenty years to fifteen, and that tightening will show up in your quotable rate about five weeks before it shows up in a bulletin. Panel shrinkage arrives as a slope, not an announcement, which is why the measurement has to be yours.

"We buy shared leads, so the cheap ones are still worth testing." Test them by all means, but book the test against cost per quotable household rather than cost per lead, and cap it. The tier-four row in the table above is not a marginal loser you can optimise into shape; it is structurally unservable until a carrier comes back. A test budget with an exit rule is fine. A permanent allocation is a subsidy you are paying to a ZIP that cannot repay it.

The Trade-Off Nobody Advertises

Here is the part that does not work cleanly. A quotability layer shrinks your addressable market, sometimes by a third, and volume-based comp plans hate that. If your buyer pays you per lead delivered, tightening the geo filter cuts your revenue while improving their outcome. That misalignment is real and it is not solved by better analytics. It is solved by renegotiating to a quality-weighted rate, which is a conversation about contracts rather than panel shrinkage mechanics.

The other honest caveat: housing-stock age is a proxy, and proxies misfire. You will bid down ZIPs containing perfectly quotable homes. Accept a false-negative rate somewhere near fifteen percent and move on, because the alternative is paying full freight for the tier-four row above.

If you are building the auto side of the same operation, the mechanics differ enough to be worth reading separately in our US auto insurance lead generation guide, and if you are buying this traffic through a network, the margin structure in how networks, publishers, and buyers split the money explains who absorbs the cost of an unquotable lead.