Paid media is breaking for law firms.
Local Services Ads, Performance Max, and broad-match Google Ads have all shifted under law firms in the past 18 months. The bidding playbook that worked in 2023 actively underperforms in 2026. Here's what changed — and what works now.
I've been managing paid media for law firms for fifteen years. The last eighteen months have been the strangest stretch I can remember. Campaigns that were stable and profitable for years have started to drift. Cost per signed retainer has crept up at almost every firm I audit. Conversion rates that hit benchmarks in 2023 are running 30 to 40 percent below those benchmarks now, on the same offers, the same landing pages, the same ad copy. Most firms have noticed but can't name what changed. Their agencies certainly haven't told them.
What changed is that the platforms quietly redesigned themselves around AI optimization while everyone was distracted by ChatGPT. The bidding system you think you're running probably isn't the bidding system that's actually executing. Three specific shifts are responsible for the bulk of the underperformance, and each one needs a different response.
Shift 1 — Performance Max is now the default, and it's optimizing for the wrong outcome.
Performance Max started as Google's experimental campaign type. It's now the default for new ad accounts and the recommended path for almost every existing account. The platform pitches it as "AI-powered, all-channel, optimized for conversions." For most law firms, that last word — conversions — is where the trap is.
Performance Max optimizes ruthlessly for whatever conversion event it's pointed at. If the firm's conversion event is "form submission," it will find every cheap form-submitter on the internet. The cost per form submission will look great. The cost per signed retainer will quietly double, because the people filling out forms are increasingly not real prospects. Click-and-fill bots, low-intent comparison shoppers, and people who fill out 12 firm forms in a row to "see who responds first" all submit forms. None of them sign retainers.
The fix: change the conversion event the campaign optimizes against. The right target for a law firm is "qualified consult booked" or, ideally, "signed retainer" — events that fire much further down the funnel and are 5 to 30 times rarer than form fills. Once Performance Max is optimizing against signed retainers (or the closest proxy your CRM can produce), it stops chasing form-fill volume and starts finding the people who actually convert. The CPA number will look worse. The cost per signed case will drop.
Shift 2 — Local Services Ads have become the highest-leverage channel, and most firms aren't running them right.
LSA — those map-pack ads with the green "Google Screened" badge that appear above organic and traditional Google Ads — have quietly become the highest-converting paid channel for most local-service legal practices. PI, DWI, family, immigration. The placement is enormous, the click-through rate is huge, and Google's algorithm specifically rewards firms with strong reviews, fast response times, and clean accreditation history.
Most firms that "run LSA" are running it badly. They set up the listing, get accredited, and then leave it on autopilot. The result: budget gets spent on weak inquiries that the firm doesn't respond to fast enough, the dispute system gets ignored when bad-fit leads come through, and Google's algorithm slowly downgrades the firm's prominence because the response time and review velocity aren't there.
The fix: treat LSA as an operational channel, not a marketing channel. The firm needs (a) a defined under-15-minute response process for LSA leads during business hours and a 24/7 AI capture for off-hours, (b) a weekly review of the dispute queue with bad-fit leads disputed against Google's criteria so the budget recycles, (c) a continuous review-collection process tied to closed matters, and (d) a monthly check on the listing's accreditation and prominence relative to direct competitors. Done well, LSA is the closest thing to "free leads" in current legal marketing — done badly, it's a slow budget drain.
Shift 3 — Broad match keywords are now matching against AI-generated queries that don't convert.
Google has been pushing broad match for years. In 2024, they doubled down — the ad platform actively recommends broad match in nearly every campaign setup, and the close variant matching now includes AI-inferred query intent rather than literal keyword overlap.
For most industries, broad match works fine. For legal, it's quietly disastrous. A broad-match campaign on "personal injury lawyer Syracuse" now matches against queries like "what is a personal injury lawyer," "do I need a personal injury lawyer," "personal injury lawyer salary," and dozens of other informational variants that Google's intent model thinks might be relevant. None of those queries convert. Most are people who just learned the term and are still in research mode three weeks before they would consider hiring anyone.
The fix: for any law firm running paid search, switch most campaigns back to phrase or exact match, build an aggressive negative keyword list including all the informational variants ("what is," "how to," "salary," "definition," "examples"), and run weekly search-term-report reviews. The first week after this change, most firms see a drop in impressions of 60 to 80 percent and a parallel rise in click-through rate, conversion rate, and cost per signed case improvement of 20 to 35 percent. The trade is volume for quality, and for legal — where one signed case can cover a quarter of paid spend — quality wins every time.
The platforms didn't break paid media for law firms. They optimized it for everyone except law firms.
The 90-day reset.
For firms that have been running paid for more than a year and noticed performance drift, the practical 90-day reset looks like this:
- Weeks 1–2 — audit and conversion event redesign. Get every campaign optimizing against the deepest funnel event your CRM can reliably fire (qualified consult, signed retainer, or both as separate events).
- Weeks 3–4 — match type and negative keyword cleanup. Move most campaigns to phrase/exact match, build out the negative keyword list against the search-term report, kill any keywords with high spend and zero signed cases over the past 90 days.
- Weeks 5–6 — Local Services Ads operational layer. Install the under-15-minute response process, the dispute review cadence, the review-collection workflow, and the listing health check.
- Weeks 7–8 — Performance Max recalibration. Either optimize PMax against deeper events, narrow asset groups by practice area, or — for many smaller firms — turn it off entirely in favor of search-only campaigns.
- Weeks 9–12 — measure, hold, scale. Let the new structure run for at least six weeks before judging it. Cost-per-signed-case is a slow-moving metric and short-term variance is noise. After six weeks, scale what's working and kill what isn't.
In the firms I've run this reset for, signed cases per dollar of paid spend typically improves 25 to 50 percent over the 90 days, on the same total budget. Sometimes the budget can come down 20 percent and still produce more signed cases than before. The platforms are genuinely capable of working for law firms — they just need to be told what success looks like in language they understand.
If your paid spend is up but signed cases aren't keeping pace, that's exactly the gap the diagnostic finds and quantifies. Start there →