Smart Bidding doesn't optimize on revenue. It optimizes on whatever you told it counts as a conversion. The Account Diagnosis checks whether that signal is telling the truth — and shows you the real ROAS next to the reported one.
Smart Bidding, Performance Max, AI Max — the machine makes almost every decision now. Exactly one lever is left in human hands: telling it what to count as success. Get that wrong, and the algorithm doesn't just waste a little budget on it — it scales the mistake, and it looks great while doing it.
ROAS looks healthy in Google Ads. The bank account doesn't reflect it.
The agency that configured your conversions is the same one reporting on them.
Since consent banners and iOS, a growing share of every reported number is modeled, not observed — and nobody told you how much.
That's almost never incremental. It's usually just measuring people who were buying anyway.
This isn't a "spend more" or "spend less" problem. It's a "do we even know what's true" problem.
We answer exactly one question: does the ROAS your account reports match what actually happened in your shop — and if not, what's the real number? Read-only account access, five checks, one page of answer plus the backup.
Every action, which one is marked "primary," what actually feeds the column Smart Bidding optimizes on.
Micro vs. macro conversions, values, double-counting, and Google Ads checked line-by-line against your shop backend for the same period.
The reported number next to the real one, what's measured versus modeled since Consent Mode, and whether brand-campaign performance is even incremental.
Every conversion action in the account, which is flagged "primary," and what actually flows into the column Smart Bidding optimizes on.
Micro- vs. macro-conversions, assigned values, double-counting, signal pollution across campaigns.
Google Ads numbers checked line-by-line against your shop backend for the identical period. The gap is the number that matters.
How much of the reported result is measured, and how much is estimated since consent banners and iOS signal loss.
Brand-campaign incrementality — a ROAS above 10 on brand terms is almost never incremental.
The real ROAS next to the reported one, the reasoning, and the exhibits behind it — not a 40-page audit.
The diagnosis ends with a precise implementation brief for whoever builds it — your developer, your agency, or a partner we can point you to.
Not the agency that set up your tracking, not a tool vendor whose dashboard shows the number in question — an outside, unbiased check.
Next to the reported one, for the same period, with the calculation shown — not a new dashboard.
The single measurement issue most responsible for the gap, not a 40-page list of minor findings.
The exact conversion actions, the account screenshots, the shop-backend comparison — every line defensible on a call.
A precise implementation instruction for whoever builds it. Not included: we build your tracking ourselves.
One question, one number, evidence you could defend to your own board. Not another strategy deck.
"We were just checking which conversion action the account treated as primary. Then we timed how long after a landing-page click a real purchase actually happened — and recalculated the ROAS."
The account reports a healthy ROAS. It's optimizing for something that isn't a sale.
The account's primary conversion is a prelander button click, not a completed purchase — every bid decision since launch has optimized for clicks, not revenue.
A conversion action gets marked "primary" during setup and never revisited — a pattern that recurs whenever the account was configured under time pressure or by whoever built the landing page, not whoever owns the revenue number.
Smart Bidding has spent months learning to produce more of the counted signal — clicks — which correlates with, but is not, a sale. Reported ROAS looks strong. It is measuring the wrong thing well.
Reported ROAS: 4.8. Recalculated against actual Shopify orders for the same 30 days, matched by order ID: 1.9.
The click-vs-purchase gap alone explains the ROAS discrepancy. Whether Consent Mode modeling adds a further gap needs a second pass most accounts also need — that's step 4 of the five checks, not skipped, just sequenced after the bigger fix.
"As someone who's run these accounts for nine years: I'd rather tell a client their ROAS is 1.9, not 4.8, than watch them scale a number that was never real."
| Maggie M. | Your current agency | |
|---|---|---|
| No conflict of interest — didn't build your tracking | ✓ | ✕ |
| Reviewed personally by the founder | ✓ | ? |
| One number, with the math shown — not a 40-page audit | ✓ | ✕ |
| Checked against your actual shop backend, not just the dashboard | ✓ | ? |
| Says what's measured vs. modeled, not just "it's working" | ✓ | ✕ |
| 48-hour delivery from read-only access | ✓ | ? |
| Fully credited toward a retainer if you continue | ✓ | ✕ |
Christian Lemke LinkedIn ↗
Founder of Maggie M., 9 years running Google Ads accounts. The Account Diagnosis exists because I've seen too many accounts scale a conversion signal nobody had checked in years.
Launch price for the first five clients (regular price €1,500). Fully credited toward the first month if you continue into a retainer afterward. 48-hour delivery once we have read-only access.
Tell us about the account. We'll confirm scope and send a payment link and access request within one business day — review starts once both are in.
We'll reply within one business day to confirm scope, send the payment link and request read-only access. The 48-hour delivery clock starts once both are in.
While you wait, answer one optional question — where do you currently feel the most pressure?
Noted — thank you.
Because it isn't guesswork — it's a line-by-line comparison of your Google Ads account against your shop backend, done personally, in 48 hours. It's priced at €999 for the first five clients (€1,500 after) and fully credited toward a retainer if you continue.
Read-only access to your Google Ads account and a way to compare against your shop's actual order data (Shopify admin access, an export, or an existing Windsor.ai/GA4 connection). No admin or edit rights required.
No. It's reviewed personally, account by account. Tools help pull and cross-check the data faster — the read and the judgment are not automated.
No, and that's deliberate — whoever repairs the tracking shouldn't be the one auditing it. You get a precise implementation brief for your developer, your agency, or a partner we can point you to.
48 hours from the point we have read-only access and the shop-side comparison data — not from when you submit the form.
A short scoping message, not a call — to confirm account size and send the payment link and access request. The diagnosis itself is delivered as a document, not a pitch meeting.
That's exactly the structural problem: whoever built or manages your tracking has an incentive not to find fault with it. This is an outside check by someone with nothing riding on the current setup.
You get the fix brief and can implement it with your own team. If you want us to run the account going forward, the €999/€1,500 is fully credited toward the first month of a retainer.
No. We guarantee an accurate answer to whether your reported ROAS is real, with the evidence shown — not a revenue outcome. What you do with the finding is up to you.
Find out before you scale a number that was never real.
€999 launch price · 48h from read access · Credited toward a retainer
When you request a diagnosis, we collect what you enter in the form: your shop URL, work email, role, the reason you're checking today, and anything you add in the optional notes field.
We use this only to scope, prepare and deliver your Account Diagnosis and to follow up about it. Form submissions are processed via Formspree and are not gated by the cookie banner, since they only happen when you actively submit the form.
If you accept the cookie banner, we also send a small set of events (that you visited this page and that a form was submitted) to Google Analytics and Google Ads, to measure how this page performs and, if you arrived via a Google ad, to measure and improve that ad's performance (including standard Google remarketing, if we run it). If you decline, no data is sent to Google Analytics or Google Ads, and Google's default privacy-preserving mode (Consent Mode) applies.
We do not sell your data or share it with anyone beyond the processors named above. To request access to or deletion of your data, email info@christian-lemke.com.
Maggie M. LLC
1209 Mountain Rd Pl NE, Ste N, Albuquerque, NM 87110, USA
EIN: 32-0780439
Website: maggie-m.com · Contact: info@christian-lemke.com
Christian Lemke
Founder, Maggie M.
€999 launch price, 48h from read access, fully credited if you continue into a retainer.
Check My Account ›