Google Ads for law firms: paying for enquiries, not clicks

Almost every underperforming legal account I have opened has the same root cause, and it isn't bidding. It is that the account cannot tell the difference between a form submission and a signed client — so Google's automated bidding optimises enthusiastically toward the wrong thing.

The situation

A firm running search campaigns across several practice areas, spending a serious monthly budget, and unable to say what it cost to acquire a client. The reporting showed a healthy number of "conversions" and a cost per conversion that looked acceptable. The partners' actual experience was a phone that rang with people asking for free advice, wrong practice areas, and the occasional job applicant.

Both things were true at once. The account was performing well against what it was told to measure. It was told to measure the wrong thing.

What the audit actually found

• Every page view counted as a conversion. The thank-you page tag was firing on page load across the site, so ordinary browsing was recorded as a lead.
• Calls weren't tracked at all. In legal, phone calls are usually the majority of real enquiries, and none of them existed in the data.
• Broad match with no negatives. The search terms report was a museum: "free legal advice", "law degree requirements", "how to become a solicitor", "citizens advice", competitor names, and long tails from entirely different practice areas.
• One campaign for everything. Family, conveyancing and employment shared a budget, so the cheapest clicks quietly ate the budget of the highest-value matter type.
• Traffic sent to the homepage. Every ad group landed on the same generic page, so a person searching for one specific matter had to navigate to find the relevant information.
• No geographic discipline. Location targeting was set to "presence or interest", pulling in searches from far outside the firm's service area.

If the conversion definition is wrong, everything downstream is wrong with confidence. Smart bidding will find you more of exactly the thing you told it to find.

The approach, in order

1. Rebuild conversion tracking first

   Nothing gets optimised until the account can count a real enquiry. That means form submissions tracked on actual submit events rather than page loads, call tracking with a minimum duration threshold so a ten-second wrong number doesn't count, and — where the firm's intake process allows — offline conversion imports so that a qualified matter, not just a phone call, is what the account learns from.

   Conversion values matter here too. A conveyancing enquiry and an employment tribunal enquiry are not worth the same amount, and telling the account they are guarantees budget flows to the wrong one.

2. Then split by practice area

   One campaign per matter type, each with its own budget, its own target, and its own definition of success. This is partly about control and partly about honesty: it makes it immediately visible that one practice area is subsidising another, which is a conversation worth having with the partners.

3. Then the negative keyword programme

   Not a one-off list — a routine. I start with a foundation negative list covering the categories legal accounts always attract: free/pro-bono seekers, students and careers queries, DIY and template searches, government and charity services, and (usually) competitor brand names. Then the search terms report gets reviewed weekly at first, monthly once stable.

   Match types get tightened at the same time. Exact and phrase carry the core commercial terms; broad match, if used at all, runs only where smart bidding has clean conversion data to work with.

4. Then landing pages that match the search

   One landing page per matter type, each answering the three questions a person in a legal problem actually has: can you help with my specific situation, what will this cost me, and what happens if I contact you. A visible phone number, a short form, genuine credentials, and no navigation to wander off into.

   This is usually where the biggest single improvement in cost per enquiry comes from, and it is the part firms most often skip.

5. Then bid strategy and reporting

   Once thirty days of clean conversion data exist, the account can move to a value-aware smart bidding strategy with a target that reflects the firm's real economics. Reporting shifts from clicks and impressions to cost per qualified enquiry by practice area — the only number the partners were ever asking for.

What changed

The shape of the improvement is predictable even when the size isn't. Recorded conversions usually fall in week one, sometimes sharply, because the account stops counting things that were never enquiries. That is the point at which a firm needs to have been warned in advance, or it looks like a disaster. Cost per genuine enquiry then typically improves over the following six to ten weeks as negatives accumulate and the bidding strategy learns from clean data.

The compounding gain is quieter: fee earners stop spending time on unqualified calls. That time saving rarely appears in an ads report and is often the largest return in the project.

What I would do differently

I would agree the qualification criteria with the intake team before touching the account, not after. "Qualified enquiry" has to be defined by the people answering the phone — matter type, jurisdiction, and whether the person can actually instruct — otherwise offline conversion imports just move the argument to a new place.

Common questions