Let me tell you about a $50 million mistake.
An insurance brand scales aggressively after a few weeks of eye-poppingly low cost-per-quote numbers from their TikTok prospecting campaigns. Quotes start to spike, blended acquisition costs plummet, and everyone on the marketing team (and their higher-ups) are thrilled.
But three months later, the mood shifts when downstream data starts to arrive. Only 5% of TikTok-acquired users completed underwriting and even fewer ultimately bound a policy – well below the benchmarks for other paid channels.
What went wrong? Low initial CPAs looked like a win. However, once a more complete data picture emerged, the business realized that it had spent tens of millions acquiring consumers with little intent to bind coverage or stay insured long-term.
In insurtech, this mistake is all too common: teams optimize for top-of-funnel quote starts or app installs without understanding whether those conversions translate into bound policies, retention, or lifetime value. It's a costly trap, but one that performance marketers can avoid.
Why the insurance funnel is a trap
Unlike ecommerce, where a click often leads to a cart addition then to a purchase, insurance funnels are complex and multi-staged. A typical path might consist of six or more steps, including:
- Initial quote start or lead capture
- Contact or identity verification
- Eligibility or underwriting flow completion
- Quote review and policy binding
- Initial payment collection
- Renewal or multi-line conversion
Every one of these steps introduces friction. For instance, it's not uncommon for every 100 quotes to result in only three bound policies, and often even fewer that renew and expand coverage.
This leads to a bitter truth: the insurance marketing metrics that are easiest to track are also the ones furthest removed from profit, leaving many marketers flying blind when it comes to optimizing towards real business value.
What vanity metrics won't tell you
Let's explore what this trap looks like in practice. These are two example brands, but I've seen very similar anecdotes in my work with insurtechs on their incrementality measurement and forecasting capabilities.
The lead-gen paradox: An insurance company partnered with a lead aggregator offering low-cost quote leads, and saw a subsequent spike in volume with dramatically decreased cost-per-quote. But cohort analysis revealed that these leads had meaningfully lower quote-to-bind conversion rates vs. other traffic sources, and those who did bind went on to churn at 2x the rate.
The premium channel surprise: Another provider saw that podcast ads were acquiring users at 3x the CPA of paid social. However, further analysis showed that those users were more likely to complete the full quote-bind journey and stay active through renewal. In reality, podcasts were actually the cheaper channel on a cost-per-retained-policyholder basis.
Without a full-funnel view, the marketing teams for both companies would have made incorrect budget allocation decisions and the company would've been forced to foot the bill.
How to unlock full-funnel measurement and executive alignment in insurtech
Experiences like these have led us to develop four strategies for breaking out of the multi-stage acquisition funnel trap:
1. Define your true north star: Start with the business outcome that really matters: not applications or quotes, but profitable policyholders. That might mean a bound policy, reaching breakeven premium, or renewal over time.
2. Implement cohort-based channel analysis: Track cohorts of users by acquisition channel through every funnel stage. How many quotes from YouTube result in bound policies? What's the 90-day policy activation rate from CTV vs. paid social? The goal is to shift your team's thinking from cost-per-signup to cost-per-bound-policy.
3. Test for incrementality at multiple stages: Run geo-holdout tests or other structured experiments that are designed to measure downstream behaviors, not just upper-funnel or creative-level events. For example, does a regional OOH campaign drive more policy binding or premium growth in that market compared to a holdout region? For higher-level strategy, MMMs — especially those equipped to handle multi-stage funnels — can measure incrementality across channels and aid your team's ability to forecast policy-level outcomes.
4. Adjust for time lag: In insurtech, users don't always convert right away. Comparison shopping and underwriting take time, and so does trust. That means a slow-moving channel could actually deliver high-value policyholders if you wait long enough to measure it properly.
Shifts like these won't come from marketing teams alone. They also require alignment with leadership. To achieve this, start by asking:
- What's our true cost per bound and retained policy by channel?
- How long does the full quote-to-bind journey usually take?
- Which channels create policyholders who renew or expand coverage over time?
- Are we prioritizing quote volume over long-term premium and retention?
These questions may lead to uncomfortable answers. You might very well discover that your star channel is driving applications instead of policies. Or, you could discover that improving only conversion or retention is more urgent. In the end, though, these hard truths are the key to making better decisions as a scaling insurtech.
Measuring for the long haul
In insurance, the easiest, often fastest metrics to track (such as Clickthrough Rate, CP-Application, and Application Conversion Rate) are often the ones that will lead you astray.
If you want to build a sustainable growth engine, you need a better framework that accounts for the full funnel, the time it takes for your users to convert, and the true behaviors that drive value. This means auditing your key performance indicators, aligning around outcomes, and measuring incrementality across the conversion stages that matter for your bottom-line.
Because not all signups — be they applications or quote requests — are created equal. And in insurance, knowing the difference between a user and a customer is worth everything.
