What bad PPC looks like on a chart (and how we fixed it)

Numbers are easy to present. What's harder to find is an agency willing to show you the numbers from before they took over — and explain exactly what was wrong.

This is one of those posts.

In September 2025 we took over the PPC account for a professional services firm in the insolvency sector. We're keeping the client anonymous, but everything in this post is real data from a real account.

What we inherited

The previous agency had been running the account for some time. On the surface it looked functional — ads were running, clicks were coming in. But when we audited the account we found three significant problems.

First, they were bidding on brand terms. The client's own company name. They were paying around £5 per click for searches that would have landed on the site for free through organic results. This is one of the most common ways agency spend gets quietly inflated — brand clicks are cheap to generate, they convert well because the user already knows who they're looking for, and they make ROAS figures look impressive in monthly reports. But you're paying for something you were already getting for nothing.

Second, all search keywords were lumped together with no structural logic. Insolvency services split cleanly into two very different audiences: individuals facing personal bankruptcy, and businesses dealing with corporate insolvency. These users have different search terms, different intent, different landing page needs — and critically, different values to the client. Corporate insolvency cases generate significantly more revenue, which means they justify a higher Target CPA. Running both audiences in the same ad groups meant a single blunt bidding strategy couldn't reflect that difference. You end up either overpaying for personal leads or leaving corporate leads on the table.

Third, there was no display activity at all — a channel that, used correctly, drives qualified reach at a fraction of search cost.

What we did

Three changes, in order of impact.

We turned off brand bidding immediately. That spend was recovered and redirected to high-intent non-brand terms where the client wasn't already winning organically.

We restructured the account into two clear streams — personal insolvency and corporate insolvency — with separate keyword sets, separate ads, and separate Target CPA targets for each audience. Corporate gets a higher CPA ceiling because the case value justifies it. Personal gets a tighter target. The bidding strategy can now actually do its job.

We introduced a display campaign to build reach at low cost, targeting users showing intent signals around financial difficulty and business restructuring.

The result

The chart tells the story. Hover over any point for the exact figures.

CPC fell from a peak of £8.30 to under £2 within two months of takeover. Qualified leads, which had dropped to 5 in the months before handover, were at 30 by January 2026 — a six-fold increase.

The CPC drop alone tells you what was happening before. When you stop paying £5 a click for your own brand name and redirect that budget to structured, high-intent non-brand terms, costs fall and quality rises. It's not complicated. It just requires someone who isn't incentivised to keep the old structure in place.

The wider point

Every agency inherits accounts that could be better. What separates a good audit from a bad one is whether the agency is willing to tell you the truth about what they find — even when that truth reflects badly on whoever came before them, or on the platform reps who signed off on the previous setup.

We show clients this kind of data because transparency is the job, not a bonus feature. If your current agency can't show you a before-and-after like this, it's worth asking why.