How I work
Most accounts I inherit aren't broken. They're just busy — years of overlapping campaigns, half-finished tests and conversion actions nobody trusts. This is the order I work through it in, and why that order matters.
01
Two weeks of reading before a single bid changes. Search terms, wasted spend, what's been tested already and what was concluded from it. Most accounts have a decision buried in the change history that nobody remembers making.
You get the audit as a document either way. If we don't end up working together, it's still yours.
02
Optimising against numbers you can't trust is worse than not optimising at all, because it's confident. Before anything else I check what's actually being counted: duplicate tags, conversions firing on page load, cross-domain journeys losing their source, offline sales never making it back in.
This is the least glamorous step and the one that changes results most.
03
If the account structure can't be explained in one sentence, it's wrong. Usually that means fewer campaigns, cleaner separation of brand and non-brand, and enough volume per campaign that automated bidding has something to learn from.
04
Changing five things in a week means learning nothing. I run tests with a stated hypothesis and a stop date, and I leave them alone until that date arrives — including the ones that look like they're failing in week one.
The record of what was tried and what it proved gets kept. That's what stops the next person repeating your last two years.
05
A monthly write-up in plain English: what changed, what it cost, what I'd do next, and what I got wrong. No dashboard tours, no metrics chosen because they happened to go up.
Accounts rarely fail because nobody touched them. They fail because someone touched everything.
Questions people ask
A flat monthly fee based on the size and complexity of the account, agreed before anything starts. Not a percentage of spend — that pays me more for spending more of your money, which is the wrong incentive to build a relationship on.
No hard floor, but below a certain volume there isn't enough data for anyone to make good decisions, and you'd be paying me to guess. If that's where you are, I'll tell you at the audit stage rather than take the retainer.
Tracking fixes and obvious waste, a few weeks. Structural changes need a full learning period plus your sales cycle before they mean anything — so six to eight weeks on e-commerce, longer on B2B.
Anyone promising results in the first fortnight is either cutting spend and calling the efficiency gain a win, or hasn't checked.
You do, always. Ads, Analytics, Tag Manager and Merchant Center stay in your ownership with me added as a user. If we stop working together you keep everything, including the history. Agencies that hold accounts hostage are relying on the switching cost rather than the work.
Yes — Meta and LinkedIn alongside Google and Microsoft. Search is the core strength and where most of the eighteen years sits, but the two are usually worth running by the same person so the budget can move between them.
For the analysis, heavily — search term clustering, anomaly detection, first-pass copy variants. It handles the volume of reading that used to eat a day a week.
Deciding what to do with any of it is still a judgement call, and that part is mine. I'll tell you which is which.
User access to the ad and analytics accounts, a realistic view of your margins, and someone who can answer questions about the business within a day or two. That last one matters more than it sounds — most bad campaign decisions come from a marketer guessing at something the client already knew.
Next step
If you want to know what shape your account is actually in, that's the place to begin — and it costs you nothing to find out.
Get in touch