See what each customer costs to serve

Annotate events with cost. Get profit and LTV per customer.

Polar already tracks revenue from orders and subscriptions. Cost Insights adds the spending side, so profit and lifetime value can be computed against actual cost of service rather than a guess.

How it works

The mechanism is intentionally small. When you ingest an event through the Polar API, you can attach a _cost property with an amount and a currency, and Polar will treat that as the cost of serving the event.

Those costs are aggregated alongside revenue on the same pipeline that already powers your usage meters, so there's no second source of truth to maintain and no separate ETL job to run.

The result is queryable through the Metrics API and visible in the dashboard with no additional setup. Profit and lifetime value land per customer, ready to plot or feed into your own analytics.

Cost events

Annotate any ingested event with a _cost amount and currency.

Cost traces

Drill into individual events to see which calls drove cost.

Profit metrics

Revenue minus cost, computed automatically and available through the Metrics API.

Customer LTV

Per-customer lifetime value, computed against the actual cost to serve.

What teams meter as cost

Anything you pay for that can be attributed to a customer can be tracked as cost.

LLM and AI APIsToken cost per call to OpenAI, Anthropic, or any provider you route through.
InfrastructureCompute, storage, bandwidth, and database load attributed to a customer.
Third-party servicesEmail, SMS, voice, search, and other vendor APIs you call on the customer's behalf.
Internal servicesAllocate the cost of an internal microservice, queue, or pipeline step.

What lands in metrics

The same Metrics API and dashboard you use for revenue, with cost as a first-class dimension.

Cost per customerSum of all cost events for a customer over any window.
Profit per customerRevenue minus cost, calculated the same way every time.
Lifetime valueCumulative profit since a customer signed up, ready to plot or pipe into a dashboard.

One pipeline, two sides

Cost Insights doesn't introduce a parallel system. The events you already ingest for usage billing get one extra property, and the rest of the work happens on the same backend.

That means a single source of truth across revenue, usage, and cost, and metrics that line up because they're computed from the same stream.

Track cost on existing events

Add a _cost property to your ingestion calls.