# How a program administrator tracks loss ratio and premium against plan (https://riska.io/blog/program-administrator-portfolio-metrics-against-plan)

_2026-10-12 · Riska · use-case_

A program administrator writing commercial property, crime and auto runs its annual plan as portfolio metrics in Riska: written premium, loss ratio, bind ratio and mix, each split by line and program and read against the target in the plan. Every number opens onto the policies and claims behind it, so a ratio that drifts in week three becomes a list of accounts to review, not a surprise at the carrier meeting.

## Key takeaways

- A program administrator's quarterly plan is a handful of numbers: written premium, loss ratio, bind ratio and mix, each set per line and per program.
- In Riska each of those numbers is a portfolio metric, defined once on the policies, premium and claims the team already works, and measured against the target in the plan.
- A ratio is shown with its numerator and denominator side by side, so a loss ratio that moves tells you whether claims rose or earned premium fell.
- Every bar opens onto the records behind it, so drift in week three becomes a list of accounts to review instead of a surprise at the carrier meeting.
- Metric definitions are drafted and released like the rest of the configuration, so the number on the screen always has one definition behind it.

> **NOTE: An illustrative use case**
> This post describes how a team of this shape runs the workflow on Riska. It is a composite example, not a report from a named customer, and it quotes no results. Where a number appears, it describes the example plan, not a measured outcome.

## What a plan review looks like at a program administrator

Picture a program administrator writing three lines for a single carrier: small commercial property, commercial crime and commercial auto. Each line runs as a program with its own rate book, its own producers and its own targets in the annual plan.

The plan the carrier signed off on is short. Property written premium on pace for the year. Auto loss ratio held under 65%. Crime bind ratio above a floor, because a crime program that stops binding has usually stopped being priced right. No single line above half the book.

Checking those numbers is where the week goes. Someone exports policies from one system and claims from another, joins them in a spreadsheet on the insured name, and rebuilds the same pivot every month. By the time the auto loss ratio looks wrong, it has looked wrong for a quarter.

The review itself takes an hour. Getting the numbers ready for it takes three days.

> **DEFINITION: Portfolio metric**
> A portfolio metric is a named measure of the book, either a flat total such as written premium or a ratio such as incurred claims over earned premium. Each side is a filtered view of records, and one column splits the result, so the same metric breaks out by line, program or jurisdiction. It reads from posted premium, closed periods and claims, never from draft quotes.

### The plan, written as metrics

Four numbers from the example plan, and how each one is defined in Riska.

| Plan line | How the metric is built | Split by | What the team watches |
| --- | --- | --- | --- |
| Written premium | Sum of posted premium lines for the period | Program, jurisdiction | Pace against the annual plan for property |
| Loss ratio | Incurred claims over earned premium | Line of business | Auto held under 65%; which leg moved when it rises |
| Bind ratio | New business binds over new business terms opened | Program | Crime above its floor; a sudden jump as well as a fall |
| Mix | Annualized premium in force by product | Line of business | No single line above half the book |

_Targets are from the illustrative plan described above, not from any customer._

## A Monday review, from headline to policy

Following the auto loss ratio from the number on the screen to the claims that moved it.

1. **Read the headline** Portfolio opens on annualized premium in force, with written premium for the period and earned premium for closed periods beside it, split by line of business. Property is on pace. Auto is a little ahead of plan on premium.
2. **Open the metric against its target** The Metrics tab carries one tab per released metric. Loss ratio shows a bar per line of business next to the target from the plan. Property and crime sit inside it. Auto has crossed 65% for the first time this year.
3. **Split the ratio into its two legs** A ratio cannot be stacked, but its parts can. The incurred claims and the earned premium behind the auto bar are shown as separate totals, with a count of the records under each. Earned premium is on plan. Incurred claims are not, so the question is now about losses rather than pricing.
4. **Narrow it to the records that moved** Filter by product, by policy status or by the split value, and every chart on the page redraws for that slice. On the Overview tab, clicking a bucket opens the underlying records. The auto slice comes down to a few fleet accounts with claims reported this quarter.
5. **Act on the account, not the spreadsheet** Each record opens on the policy, its terms and its claims, so the underwriter reviewing renewal terms on those fleets is reading the same claims the ratio counted. The finding goes to the carrier meeting with the accounts attached.

## Where drift shows up before the quarter closes

Drift rarely announces itself in the headline. A book can hit its written premium target while one program quietly fills with the wrong risks, and the total hides it until the losses arrive.

So the team reads the splits, not the totals. Mix shows auto creeping toward half the book while property holds flat. Bind ratio by program shows crime binding more often than it did last quarter, which is a pricing question as much as a sales win. Time in state and backdated changes show where policies stall or are edited after the fact.

None of these needs a new report. They are the same handful of metrics, read weekly, each one a click from the records that produced it.

## Portfolio metrics on the records the book runs on

Most program administrators track their plan in a spreadsheet built from extracts, which means the number the carrier sees and the policies an underwriter edits are two copies that drift apart.

Riska is built the other way round. [Portfolio metrics](/platform/reporting) are defined on the same policies, posted premium and [claims](/platform/claims) the team works every day, and the charts and dashboards in [reporting](/platform/reporting) read the same records. A metric is drafted, released and pinned like the rest of the configuration, so when the carrier asks how the loss ratio was computed, there is one definition to show them and a history of every change to it.

## Questions we get asked

### Is this a real customer?

No. This is an illustrative use case describing how a program administrator writing commercial property, crime and auto would track its plan in Riska. It names no customer and reports no results.

### What portfolio metrics can a program administrator track in Riska?

Written premium, earned premium, annualized premium in force, average premium, loss ratio, bind ratio and premium mix come defined, each split by line of business, program or jurisdiction. Teams can define their own flat totals or ratios from any filtered view of their records.

### How is loss ratio calculated?

Loss ratio in Riska is incurred claims divided by earned premium, split by line of business. Both legs are shown alongside the ratio, so a change can be traced to losses or to premium.

### Can I see the policies behind a metric?

Yes. Every chart and metric can be filtered by product, status and split value, and clicking a chart bucket opens the underlying records, so each number leads to the policies and claims that produced it.

### Does changing a metric definition change past reports?

A metric is drafted and released like other configuration, and the editor previews a draft before it ships. The portfolio screen reads the released definition, and every earlier version is kept.