Why cost per lead is the wrong number
Cost per lead is the easiest auto-insurance metric to move and the easiest one to fool yourself with. A lead is a form fill. It costs the same whether the person behind it is a clean 45-year-old with two cars and a mortgage or someone with a suspended licence that no carrier will write. Meta's optimiser does not know the difference, and if you tell it to find cheap leads, it will find you the people most likely to tap a form, which is not the same population as the people most likely to bind a policy.
The thing a campaign is supposed to produce is a bound policy: a quote accepted, underwriting passed, first premium paid. That is the only event in the chain that turns into revenue. Cost per lead sits four steps upstream of it. When the metric you report and the metric that pays the bills are four steps apart, you optimise the account into a place that looks efficient on the dashboard and bleeds money on the loss run.
This is not a Meta-specific failing. It is a measurement failing that Meta happens to expose ruthlessly, because the platform optimises toward whatever event you feed it. Search captures people typing "cheap car insurance near me," which carries real intent. Aggregator leads carry intent too, then dilute it by selling the same lead to several agents at once. Social captures attention earlier in the decision. That is fine, and often cheaper per impression, but it means a Meta lead is a colder object than a search click, and pretending the two convert at the same rate is how budgets get misallocated.
What published Meta insurance benchmarks actually say
Start with the numbers that do exist, because they are useful as long as you read them for what they measure. The most-cited public source is the WordStream and LocaliQ Facebook Ads Benchmarks 2024, published 8 September 2025 on 2,946 campaigns running between February 2023 and April 2024. Here is the Finance and Insurance row against the all-industry baseline.
| Metric | Finance and Insurance | All industries |
|---|---|---|
| Cost per lead | $38.09 | $21.98 |
| Cost per click | $4.57 | varies by vertical |
| Click-through rate | 1.84 percent | 2.53 percent |
| Conversion rate | 5.98 percent | varies by vertical |
Read the table carefully. Finance and Insurance pays nearly double the average cost per lead, earns a lower click-through rate, and still posts a healthy conversion rate, because a regulated, high-consideration purchase attracts fewer idle clicks but more serious ones once someone does click. That $38.09 is a defensible planning anchor for a Meta lead-gen form fill in this vertical. It is not, and was never meant to be, a cost per customer.
The other public benchmark sources behave the same way. Databox's Facebook lead-gen benchmark group reports median insurance cost per lead from contributor data. Gupta Media tracks Meta cost per thousand impressions with vertical premiums. Every one of them stops at the lead or the on-platform action. None of them publishes a cost per bound policy, because none of them can see your underwriting. That absence is the whole reason this post exists.
The lead to quote to bind funnel
To get from a lead to a policy you pass through a funnel that the insurance industry names consistently: lead, then quote, then bind. A lead is captured contact information. A quote is a price offered after the underwriting criteria run. A bind is a policy issued and paid. MediaAlpha's 2020 S-1 draws the same shape as click, then quote, then bound policy; EverQuote frames it as consumer request, then match, then bind. The vocabulary is stable across the sector.
- Lead to quote measures how many of your form fills are real, reachable people with insurable risk that your team or system actually prices.
- Quote to bind measures how many priced prospects accept and pay. This is the close.
- Lead to bind is the product of the two, and it is the rate that turns a cost per lead into a cost per policy.
One honest caveat: the funnel is tidier on a slide than in a call centre. The "quote" stage usually hides a messy re-engagement layer, a string of calls, texts, and emails where most of the drop-off actually happens. A Meta lead that never picks up the phone never becomes a quote, which is why speed-to-lead and follow-up discipline move the bind rate as much as creative does.
What does a realistic bind rate look like? EverQuote, an auto-insurance lead marketplace, publishes its own figures: agents see bind rates around 20 to 30 percent on consumer-initiated inbound calls, against a best case near 10 percent on other lead types, and agency close rates that commonly sit at 25 to 33 percent. Shared aggregator leads, sold to several agents at once, bind in the low single digits. A first-party Meta lead-gen fill generally lands between those poles, cooler than an inbound call, warmer than a thrice-sold aggregator lead. The number that matters is your own backend close rate, not any public average, because lead quality swings enormously by creative, audience, and follow-up.
Deriving cost per bound policy
Because no public source publishes it, cost per bound policy is something you build. The formula is short and it is division, which trips up more people than it should.
cost per bound policy = cost per lead / lead-to-bind rate
And since lead-to-bind is itself the product of the two funnel stages:
cost per bound policy = cost per lead / (lead-to-quote rate x quote-to-bind rate)
Plug the benchmark in. At the $38.09 Finance and Insurance cost per lead and an 8 percent lead-to-bind rate, cost per bound policy is $38.09 divided by 0.08, which is about $476. At a 4 percent lead-to-bind rate it doubles to roughly $952. The bind rate, not the lead price, is the lever with the most travel. The table below shows the same cost per lead resolving to wildly different cost per policy as the bind rate moves.
| Lead-to-bind rate | Bound policies per 100 leads | Cost per bound policy |
|---|---|---|
| 2 percent | 2 | about $1,900 |
| 4 percent | 4 | about $950 |
| 8 percent | 8 | about $476 |
| 12 percent | 12 | about $317 |
| 20 percent | 20 | about $190 |
One warning the math itself will not give you. Multiplying cost per lead by the bind rate, rather than dividing by it, is a common slip that produces a number so small it flatters the account into disaster. If your derived cost per bound policy comes out lower than your cost per lead, you have inverted the formula. The policy always costs more than the lead, because most leads never become policies.
The second warning is about inputs. A public cost per lead married to a generic public close rate from a different universe gives you a fantasy figure, useful only for a first sanity check. The real version uses your own cost per lead from the Meta account and your own lead-to-bind rate from the CRM and policy system. Joining those two datasets is the analytics-and-insights work that makes the whole exercise honest, and it is the part most teams skip.
Wiring the bind event back with the Conversions API
Deriving cost per bound policy in a spreadsheet is the reporting half. The optimisation half is teaching Meta to bid toward bound policies instead of leads, and that requires sending the bind event back to the platform. The Conversions API is the pipe: a server-to-server connection from your server, CRM, or policy system into Meta's optimisation engine, independent of the browser pixel.
Six steps. None of them exotic.
- Capture the lead on Meta, with the pixel and a click identifier stored against the record.
- Quote and bind offline, in the CRM or call centre, the way the business already works.
- Fire the bind event server-side through the Conversions API when the policy is issued, mapped to Meta's standard
Purchaseevent. - Attach the value: pass the policy premium as the
valueparameter, with a currency, so Meta learns which leads turn into bigger policies. - Raise the match quality by sending hashed email, phone, name, and ZIP alongside the browser identifier
fbpand the click identifierfbc. Meta scores this as Event Match Quality on a 1 to 10 scale, and a higher score means the value reaches the right ad. - Deduplicate a quote-online, bind-by-phone journey by passing a shared
event_id, so Meta counts one conversion, not two.
With bind events flowing, you switch the campaign from optimising for leads to the Maximize value of conversions goal, Meta's value optimisation strategy. Now the algorithm bids toward people likely to bind a higher-premium policy, not people likely to tap a form. Meta's separate value rules feature, renamed in 2025, sits on top of this and lets you adjust how value is interpreted by segment, for example weighting higher-premium ZIP codes, and in 2025 you can attach or detach those rule sets without resetting the learning phase. The mechanics are documented in Meta's Conversions API reference.
This is the heart of the performance marketing work: senior judgement on what to measure and which event to optimise, with the machine doing the bidding underneath. Getting the server-side plumbing right, the CRM webhook, the event mapping, the consent handling, is also where website-and-systems development earns its place, even though it is rarely the part anyone budgets for. If you want that conversation, it starts at the contact form, not a media plan.
What this changes in how you run the account
Shifting the target metric from cost per lead to cost per bound policy changes the daily account work, well beyond the quarterly slide. The creative and audience decisions you make start serving a different master.
- You stop chasing the cheapest lead. A higher cost per lead is acceptable, and often correct, when it buys a population that binds. The cost-per-lead column becomes a diagnostic, not a goal.
- You report cost per bound policy to the people who hold the budget. A CFO recognises cost per policy against premium and lifetime value. Cost per lead means nothing to that audience, and reporting it trains everyone to optimise the wrong thing.
- You invest in the follow-up layer. Speed-to-lead, call discipline, and re-engagement move the bind rate, which moves cost per bound policy more than any bid tweak. The media and the operations stop being separate problems.
- You hold Meta accountable to the right event. Once value-based bidding is live, the platform's interests align with yours, because both of you are now optimising toward a paid policy.
This is the same logic we apply across regulated paid social accounts: pick the metric that maps to money, instrument the funnel so the platform can see it, then let value-based bidding do the heavy lifting. For the video and disclosure side of insurance marketing, the companion read is our piece on compliant YouTube pre-roll for life insurance; for the broader Meta account craft, see Meta Ads mastery in the Advantage+ era. Cost per bound policy is not an exotic metric. It is just the honest one, and the only reason it feels exotic is that the public benchmarks never had the data to publish it.