Cost Per Signed Case Is Not Cost Per Click: A PI Law Media Model

A PI firm partner asked me last quarter why his cost-per-lead was $180 but his marketing felt like it was bleeding money. The answer was simple: 96% of those leads were never going to sign. The 4% that did were expensive enough that the real cost-per-signed-case was over $4,500 against a docket of mostly soft-tissue cases averaging far less in attorney fee.

The math is not the problem. The wrong math is the problem. Most PI firms are running the wrong model.

The metric stack, in order of seriousness

There are five layers in the funnel. Each is useful. Only one decides whether the program is profitable.

• Cost per click (CPC). Tells you what the auction costs. Useful for diagnosis, not for decisions.
• Cost per lead (CPL). Sometimes called cost per form fill or cost per call. Hides the qualification problem completely.
• Cost per qualified lead (CPQL). Calls and forms that match your practice area, geography, and severity. This is the first metric the agency and the firm should agree on.
• Cost per signed case (CPSC). The actual operating metric. Includes everything above plus the conversion rate from qualified lead to signed retainer.
• Cost per dollar of attorney fee (CPDAF). The metric your CFO and your equity partners actually care about. Different case types have different average attorney fees, so a $4,500 CPSC means very different things on a soft-tissue case versus a commercial trucking case.

If your weekly performance review still anchors on the first two metrics, you are flying with the gauges turned off.

Why CPL is structurally misleading in PI

Three reasons the cost-per-lead number lies to you:

• Practice-area mismatch. Your "free case review" form pulls in workers' comp, medical malpractice, slip-and-fall, and auto, but you only handle two of those. The leads count, the cases do not.
• Geographic spillover. You buy "car accident lawyer" in your city, and Google Ads quietly serves it 80 miles outside your service area on broad match. The form fill is real, the case is unsignable.
• Quality decay over time. As an account scales, Smart Bidding pushes deeper into the audience, and lead quality drops on the margin. CPL stays flat. CPSC quietly doubles.

The model your CFO will actually approve

Here is the spreadsheet structure we install. One row per campaign per month. Columns:

• Spend
• Clicks
• Form fills + qualified phone calls
• Qualified leads (after CallRail tagging and intake review)
• Signed retainers (pulled from the CRM)
• Average attorney fee by case type for that campaign
• Total expected attorney fee from those signed cases
• ROAS expressed as expected fee divided by spend

Every campaign gets a verdict against a target ROAS, not against a CPL benchmark.

This is not a complicated spreadsheet. The reason most firms do not have it is the underlying data wiring is broken (see our post on CallRail-to-CRM attribution), not because the math is hard.

The realistic ROAS targets

We do not publish specific client numbers, but the working ranges across PI accounts we have built or audited:

• Soft-tissue auto: target ROAS in the 3x to 5x range against expected fee. Below 3x, the program is barely covering case acquisition cost plus operations.
• Commercial vehicle and trucking: target ROAS in the 8x to 15x range. Higher case values justify higher CPCs and higher CPSCs.
• Medical malpractice and product liability: the math gets weird because case settlement timelines stretch past the offline conversion window. We model these on signed cases per quarter rather than ROAS.

These are patterns, not promises. Your local CPC environment, intake conversion rate, and case mix all move the numbers.

The two adjustments that compound

Two changes inside this model produce most of the gain:

• Move the bidding signal from form fill to signed case. Once Google Ads is bidding on signed retainers, not form fills, the algorithm rewires itself toward the patterns that produce actual revenue. This requires the offline conversion wiring to be clean.
• Segment campaigns by case type, not by keyword theme. A "car accident" campaign that mixes commercial trucking, soft-tissue, and rideshare is averaging three completely different ROAS profiles into one number. Split them, fund the winners separately, starve the losers.

What this looks like in a Monday meeting

The conversation moves from "our cost-per-lead is up 12%" to "our soft-tissue ROAS dropped from 4.1 to 3.6 because the intake conversion rate fell, not because acquisition costs rose." That is the conversation that produces decisions. The first one produces hand-wringing.

If you want a media model built specifically for your firm's case mix, we can stand one up in 2 to 3 weeks against your CRM data. Book a free audit.