Embedded insurance — the integration of coverage directly into a product or service at the point of sale — has grown rapidly in sectors such as travel, mobility and retail. For large enterprises, the path is often clear: acquire or build a dedicated insurance capability. For mid-market product firms, the decision is more nuanced. This article examines the threshold at which building an embedded insurance function becomes commercially sensible, and when buying from a specialist partner remains the better option.
The Embedded Insurance Landscape in 2025
Embedded insurance premiums globally are projected to exceed $100 billion by 2028, according to a 2024 report from McKinsey. The market is driven by consumer demand for frictionless purchasing and by product firms seeking new revenue streams and customer retention mechanisms. In the mid-market — firms with annual revenues between $50 million and $500 million — adoption has been slower but is accelerating.
Mid-market product firms typically sell physical goods, software or services where insurance can be a natural add-on: warranty extensions for electronics, liability cover for gig economy platforms, or theft protection for bicycle retailers. The question is not whether to offer insurance, but how.
The Build Decision: When Internal Development Makes Sense
Building an embedded insurance capability in-house requires significant investment in technology, regulatory compliance and underwriting expertise. A firm must either become a licensed insurer or partner with a carrier while building the front-end and data infrastructure itself.
Scale threshold. Industry practitioners suggest that a firm needs at least 100,000 policies per year to justify the fixed costs of a build. Below that volume, the per-unit cost of developing and maintaining a proprietary platform typically exceeds the margin available from insurance commissions. At higher volumes, the economics shift: the firm captures the full margin rather than sharing it with a partner.
Control and data. Firms that build retain full control over the customer experience, pricing and data. For a product company whose core offering generates rich behavioural data — usage patterns, claims history, customer lifetime value — owning the insurance layer can improve risk selection and pricing over time. This data advantage is difficult to replicate through a third-party partner.
Regulatory burden. In the UK, the Financial Conduct Authority (FCA) requires any firm arranging or advising on insurance to hold appropriate permissions. Building in-house means taking on compliance costs, capital requirements and ongoing regulatory reporting. For mid-market firms without existing financial services expertise, this can be a material distraction.
The Buy Decision: When Partnering Is Prudent
Buying embedded insurance capability from a specialist insurtech or broker allows a product firm to offer coverage quickly, with minimal upfront investment. Partners typically provide the policy wording, claims handling, regulatory compliance and technology integration.
Speed to market. A buy approach can reduce time-to-launch from 12–18 months to 8–12 weeks. For firms testing demand or entering a new market, this speed is valuable. It also allows for iterative experimentation without long-term commitment.
Variable cost structure. Partners charge a commission or fee per policy, converting a fixed cost into a variable one. This is attractive for firms with unpredictable or seasonal volumes. The trade-off is lower margin per policy, typically 10–30% of premium versus 40–60% for a build.
Risk transfer. The partner bears the underwriting risk and regulatory liability. For a product firm whose core competence is not insurance, this reduces exposure to claims volatility and compliance failures.
The Threshold: A Decision Framework
Based on available industry analysis and practitioner interviews, the build vs. buy threshold can be mapped across three dimensions:
- Volume. Above 100,000 policies annually, build economics become competitive. Below that, buy is usually cheaper.
- Data maturity. Firms with rich customer data and analytics capability benefit more from building, as they can optimise pricing and retention. Firms with limited data should buy.
- Strategic importance. If insurance is central to the product value proposition — for example, a home repair platform that includes cover — building may be justified even at lower volumes. If insurance is a secondary add-on, buying is safer.
Commercial Impact. The margin differential is significant. A mid-market firm selling 150,000 policies per year at an average premium of £200 could generate £30 million in gross written premium. Under a buy model at 20% commission, the firm earns £6 million. Under a build model at 50% margin, it earns £15 million — but must absorb £3–5 million in annual fixed costs for technology, compliance and claims handling. The net advantage of building is £4–6 million, assuming volumes are sustained.
Risks and Unknowns
Regulatory change. The FCA and European regulators are increasing scrutiny of embedded insurance, particularly around fair value and disclosure. A 2024 FCA review found that some embedded products offered poor value to consumers. Future rules could compress margins or require additional disclosures, affecting both build and buy models.
Claims volatility. Firms that build take on insurance risk directly. A single catastrophic event — a product recall, a natural disaster — could generate claims far exceeding premium income. Reinsurance can mitigate this but adds cost.
Partner dependency. Firms that buy face concentration risk. If a partner changes terms, exits the market or suffers a data breach, the product firm’s insurance offering may be disrupted. Contracts should include exit provisions and data portability.
Technology integration. Both models require integration with the firm’s existing checkout, CRM and customer service systems. Poor integration leads to friction, low uptake and customer complaints.
Why It Matters
Embedded insurance is becoming a standard expectation in many product categories. Mid-market firms that delay a strategic decision risk losing revenue and customer loyalty to competitors that offer seamless coverage. The build vs. buy choice directly affects margin, control and risk exposure. Getting it wrong can mean wasted investment or missed opportunity.
FY Outlook
Over the next 12–24 months, we expect the threshold to shift downward as insurtech platforms lower the fixed cost of building. Modular, API-first insurance infrastructure — sometimes called “insurance-as-a-service” — is reducing the minimum viable scale for a build. Firms with 50,000 policies may soon find building viable, particularly if they already have strong data and engineering teams.
Regulatory pressure will increase, especially in the UK and EU. Firms should monitor FCA guidance on fair value and ensure any embedded product delivers genuine benefit to customers. Those that treat insurance as a pure profit centre without considering customer outcomes may face enforcement action.
Partnership models will evolve. We expect more revenue-sharing and co-branded arrangements that give product firms greater control without full build risk. The line between build and buy will blur, with hybrid models becoming common.
Conclusion
For mid-market product firms, the embedded insurance decision is not binary. The threshold depends on policy volume, data capability and strategic intent. Most firms below 100,000 policies should buy from a specialist partner. Above that volume, building becomes viable and potentially more profitable, provided the firm can manage regulatory and underwriting risk. Firms should revisit the decision annually as volumes grow and technology evolves.
Source notes. This analysis draws on the McKinsey Global Insurance Report 2024, the FCA’s 2024 review of embedded insurance products, and practitioner insights from insurtech executives and mid-market product managers. No specific company names or quotes are attributed without direct verification. All volume and margin figures are illustrative ranges based on industry benchmarks, not precise data from any single firm.
