The conversions do not count what you sell
No purchase conversion, or one with no revenue attached, or the form that actually matters left untracked. Sometimes all of it exists and one campaign counts margin while the next counts order value, so the same month can be told two ways. Until that is fixed, nobody can say which campaigns are profitable, and every budget decision is a guess. This is the first week of work, always.
Google’s recommendations, applied as though they were a plan
They arrive with a score, they are easy to accept, and they are built to make you spend. A few of them are worth taking. Most raise budgets and loosen targeting on the campaigns that already spend the most, which moves money without moving revenue. An account run on auto-apply is an account run by someone whose interests are not yours.
The structure ignores your margins
Your best products and your best margins sit somewhere in the catalogue, and the campaigns have no label that tells them apart. So the budget spreads evenly over items that earn very differently, and a long tail of products that have never produced an order keeps taking spend. That is usually the fastest money on the table.
Your catalogue is invisible for the searches that buy
Titles that do not carry the words people type, product types missing, identifiers wrong, prices drifting against the street. The campaigns cannot bid on a query the catalogue does not qualify for, so a structure problem gets read as a bidding problem and the budget gets raised instead.
Nobody senior is actually in your account
An agency sells hours, so whether your revenue rises or falls changes very little for it. The senior who ran the pitch is not the person who opens the account afterwards, and new accounts go to whoever joined most recently. I spent eight years inside that and I am not guessing.
How the work runs
The diagnostic
Full history pulled, campaign by campaign and conversion action by conversion action, change history included. You get the ranked list of what is costing you money, with what each item is worth, and a 60-minute call to go through it.
Stop the losses, agree what a sale is worth
The priority fixes go live, one definition of a sale goes into the account, and the starting point is recorded. Every change arrives as a list with a reason per line and goes live once you say yes.
Move the budget to what sells
The catalogue, the labels your campaigns route on, the products that take budget and never convert, the prices against the market. This is where spend stops funding the tail.
Before and after, then your decision
At day 35, one report comparing the account before and after on the definition agreed in week one, with the data behind it. You have ten days to decide with all of it in hand.
What I do not do
No SEO, no social ads, no creative production, no full-funnel promises and no hourly billing. Google Ads, and Microsoft Ads when you run it, is the whole job.
Whether this is worth it for you
Built for e-commerce brands spending from about 5,000 EUR a month on Google Ads. Under 3,000 the fee becomes too large a share of the budget, and I will say so on the call rather than take it. No Google Ads yet? The setup is priced on its own, from 2,000 EUR.