Move small policies to lower-cost service
When it may fit: Cost per policy exceeds what premium supports.
Evidence needed: Cost per policy by segment.
Trade-off: Customers may resist.
Insurance Intelligence Brief
How to Evaluate Service Costs Across Customer Segments
Serving a customer costs about the same whether the premium is small or large. A segment view shows where service cost eats the margin.
1. The business problem
Which customers cost more to serve than they are worth?
Policy administration, billing, inquiries and service take effort that does not scale with premium. A small policy takes nearly as many calls as a large one. Segments with small average premium can have service costs that consume the margin.
Measuring service cost by segment shows where to redesign service, move to self-service or change pricing.
2. What the evidence can reveal
A service-cost analysis includes:
- Calculated Service and administration cost per policy and as a share of premium.
- Calculated Contacts and transactions per policy, by segment.
- Calculated Contribution after service cost by segment.
- Information gap Channel used, complexity of coverage and customer preferences, which affect cost to serve.
Cost per policy is the key measure, since it can be compared with premium per policy.
3. Investigate the problem
In the synthetic insurer, Calculated other underwriting expenses were $7.60M, up 7.0%, and acquisition costs were $10.5M. With roughly 43,000 policies in force, other expenses are about $176 per policy and acquisition about $243 per policy.
Illustrative A simple illustration, assuming a similar $176 service cost for every policy: on a $600 premium it is about 29% of premium, while on a $4,000 premium it is about 4%. A segment dominated by small policies could be unprofitable on service cost alone, even with a good loss ratio.
Calculated Lines differ in premium size. Specialty & Umbrella, with $6.50M of premium, and Small Commercial, with $15.3M, tend to carry larger policies than personal lines. Hypothesis Personal Auto, with a loss ratio of 74.3% and large volume, is where small-policy service cost and claims cost together would matter most. Information gap Policy-level premium and service cost are not in the demo. A first analysis would compute service cost per policy by line and compare it with average premium.
4. Explore the interactive experience
Investigate it yourself.
Insurance Portfolio Explorer: a portfolio dashboard with loss, expense and combined ratios, a scenario simulator for rate, volume, frequency and severity, evidence-linked answers from APIS, and a sample brief, all on synthetic data.
The experience below is built on a synthetic insurer, “Demo Mutual Insurance.” It uses no policyholder, producer or agency data of any kind, so it is used to show the portfolio economics that agency and brokerage decisions connect to. Every figure is illustrative.
- In the cost table, select Other expenses. How fast is it growing compared with premium?
- Select Personal Auto in the product table. Why might service cost matter most there?
- Ask APIS which expense categories explain the underwriting margin change.
5. Evaluate possible solutions
Responses depend on where service cost is highest relative to premium.
6. What leadership should do next
Questions to answer
- What is service cost per policy by segment?
- How does it compare with premium?
- What drives contacts?
- What can be automated?
Data to gather
- Expenses by activity
- Contacts per policy
- Premium per policy
- Customer preferences
Steps to take
- Compute service cost per policy.
- Rank segments.
- Redesign for the smallest.
- Re-measure.
7. Engage Two Pillars
Investigate Performance Across Your Insurance Business
Two Pillars can help carriers, agencies and brokerages investigate portfolio, claims operations, agency economics and process improvement using your own data. This demo uses no policyholder data. A good first step is a cost-to-serve analysis by customer segment that compares service cost per policy with premium.
Sources and data notes
- The organization, its figures, and every number in this article and its interactive experience are synthetic and illustrative. They are not real data, a benchmark, or a forecast.
- Reviewed October 2026. This article describes an approach to investigation. It does not replace review by qualified professionals.