The conversion gap: why law firm ad reports look better than the caseload

There is a conversation that happens in most law firms that advertise. The monthly marketing report arrives and it looks healthy — dozens of conversions, a respectable cost per lead, a chart pointing the right way. And somebody in the partners' meeting asks the awkward question: if the ads are converting this well, where are the clients?

Both things are usually true at once. The report is accurate on its own terms, and the caseload hasn't moved the way the report implies. That distance between them is the conversion gap, and it isn't caused by anyone lying. It's caused by what "conversion" is allowed to mean.

How the typical setup works

A firm hires an agency to run its ads. The agency — quite reasonably, given the tools it has — builds the machinery it can control: landing pages hosted on its own infrastructure where the ads point, a call-tracking layer that counts phone calls over a certain length, and form tracking on those pages. Each of those events gets reported to the ad platform as a conversion, and to the firm as a result.

Notice what that machinery can and cannot see. It can see that a phone rang and someone stayed on the line for forty seconds. It cannot see whether that was a new client, a recruiter, a courier, or an existing client chasing an update. It can see a form was submitted. It cannot see whether the enquiry was in scope, whether anyone called it back, or whether it became a matter. The tracking ends at the enquiry, because everything after the enquiry lives inside the firm — in the practice management system, the reception notes, the fee earner's inbox — where a marketing vendor has no access, and for confidentiality reasons rightly never will.

So "conversion", in the report, means an enquiry-shaped event occurred. Not a client. Not even necessarily a person with a legal problem.

Why the numbers inflate

It gets worse than loose definitions, because the inflation is mechanical. When we audited a live, professionally-managed account, the ad platform claimed roughly two and a half times as many conversions as there were real people who actually enquired. The excess wasn't fraud. It decomposed into three ordinary, almost universal causes:

Engagement counted as conversion. Ad platforms will happily count map-direction clicks, video views and page visits as conversions if the account is configured to — and accounts frequently are, sometimes on a single campaign, set years ago and never revisited. In the account we audited, a quarter of the claimed conversions were people clicking "directions" on a map listing. Nobody had noticed, because the number they inflated was the number everyone wanted to see rise.

The same caller counted twice. Layer a call-tracking service on top of the platform's own call reporting — a very common stack — and one genuine caller can register as two conversions, sometimes three. The report doubles; the phone rang once.

Modelled conversions. The platforms estimate conversions they couldn't directly observe and add them to the column. Reasonable statistics, but it means the headline number was never a count of people to begin with.

The part that actually costs money

If this were only a reporting problem, it would be an annoyance. It's worse than that, because modern ad platforms are optimisation machines: they spend your budget seeking more of whatever you tell them a conversion is. Feed the algorithm map-clicks and forty-second calls, and it will faithfully hunt down the cheapest map-clickers and brief callers it can find — with your money. The report improves. The caseload doesn't. The gap widens, and it compounds every month the campaign "learns".

And here's the structural trap: nobody in the chain can fix it alone. The agency can't — it has no visibility past the enquiry, and no firm should hand its client records to a marketing vendor. The firm can't — solicitors don't live inside ad accounts and shouldn't have to. The platform can't — it optimises whatever it's fed, garbage included. Everyone is doing their job. The loop simply never closes.

One more thing worth knowing before you change anything: those landing pages and tracking numbers usually live on the agency's infrastructure, not yours. Change agencies and the pages, the pixel history and the accumulated learning typically leave with them — the next agency starts your algorithm from zero. Whatever else you do, own your own funnel.

What closing the loop looks like

The fix isn't better dashboards on the same data. It's connecting the two halves that have never met: what the ad account claims, and what actually happened inside the firm.

That means capturing every enquiry at the firm — calls, forms, email, chat — with its source attached, then following each one through the firm's own systems to the only outcomes that matter: qualified, instructed, and what the work was worth. Once that record exists, two things become possible that the typical setup can never do.

First, the platforms can be fed the truth. When a verified enquiry becomes a real matter, that fact — and its value — goes back to the ad platform as the conversion signal. The algorithm stops optimising for phone rings and starts optimising for instructions. Same budget, different target.

Second, the claims become checkable. Put the platform's reported conversions next to the verified count of real people for the same period, and the gap stops being a suspicion in a partners' meeting and becomes a number on a page — one you can watch shrink as the configuration gets cleaned up.

This is what Kefilex does as a by-product of the firm simply answering its phones. It sits inside the firm, where the truth lives, which is exactly the seat a marketing vendor can never occupy. The enquiry record, the follow-through, the verified conversions fed back to Google and Meta, and the honest side-by-side of claimed versus real — none of it requires the firm to run its own ads, fire its agency, or learn ad jargon. It requires the loop to exist.

The old way isn't failing because agencies are bad at ads. It's failing because the definition of success was set by what the tracking could see — and the tracking has never been able to see a client. Measure instructions instead of enquiries, and everything downstream of that decision, including the ad spend, starts working toward the same goal the firm has.

To see what six months of closing the loop found inside a real firm, read Built inside a law firm.