The Dashboard Says You're Winning. Your Bank Account Disagrees.

You open the monthly report. The numbers are green. Clicks up. Conversion rate up. Quality Score up. There’s a chart with a pleasant upward slope. The narrative paragraph at the bottom uses the word “momentum.”

Then you open Shopify. Revenue is flat. In-store foot traffic hasn’t moved. Your accountant asks, politely, what the Google Ads line item is actually doing.

If this has happened to you more than once, your agency isn’t reporting on the thing that matters. They’re reporting on the thing that makes the report look good.

The tracking sleight of hand

There is a short, well-worn list of metrics that agencies use to tell a good story regardless of whether the account is actually producing:

“Conversions” — often a catch-all that includes add-to-cart, newsletter sign-ups, “store visit” estimates, and soft micro-conversions. None of those pay your suppliers.

“Engaged sessions” — Google Analytics 4’s default. A session qualifies as engaged if the user stays on the site for ten seconds, fires two page views, or triggers a conversion event. Ten seconds.

“View-through conversions” — attributing a sale to an ad the person never clicked on, because the ad was rendered on a page they scrolled past at some point in the last month.

“Brand search lift” — presented as a headline campaign win when brand keywords are sitting inside the same campaign as acquisition keywords (see: branded-blending). A branded campaign has a real defensive job; the sleight of hand is letting brand revenue mask a weak non-brand acquisition number and letting Google’s optimiser siphon acquisition budget onto brand clicks.

“Impression share” — a ceiling metric. Irrelevant unless paired with what those impressions cost and what they produced.

Any of these, in isolation, can be a legitimate diagnostic. Stacked into a monthly report with no attributable revenue figure alongside them, they are a conjuring trick.

What you should be measuring

There is really only one number that matters for a performance-driven Google Ads account: attributable revenue from every paid Google click, divided by total Google Ads spend. That is ad-attributable ROAS. That is the number that tells you whether the account is paying for itself.

Attributable revenue means: orders that originated from a paid-search click, measured through server-side or well-configured client-side tracking that survives ad blockers, cookie restrictions, and cross-device journeys. Not sessions. Not estimates. Not view-throughs. Actual purchases.

The total number has to be the number you judge the ad programme on — because it’s the number your bank account cares about. Brand and non-brand clicks both cost real money and both produce real revenue. Both belong in the fee calculation. What doesn’t belong is mixing them inside a single campaign, which is a structural problem that hides a weak acquisition number and lets Google’s optimiser burn acquisition budget on brand clicks (more on this on our branded-blending page). The fix is structural separation at the campaign level, with unified reporting on top.

Why agencies avoid this number

Because it’s hard. Reporting on total ad-attributable revenue with structurally separated campaigns requires:

1. A tracking setup that actually fires correctly — server-side enhanced conversions, Google Ads + GA4 reconciled, offline conversion uploads wired in for in-store purchases where possible.
2. Ruthless campaign segmentation so brand and non-brand spend are separable — two campaigns, hard negative-keyword walls between them, not one blended pool.
3. A willingness to report numbers that sometimes look bad, because you’ve stripped out the vanity layer.

Each of those takes work. Each of those exposes the account to being judged on its real performance. Most agency reporting is optimised to avoid that judgement.

How kPixies reports

Every report you receive from us has, at the top, three numbers: total Google Ads spend, total attributable revenue, and total ad-attributable ROAS. Not buried on slide seven. Not sandwiched between engagement charts. Right at the top. Underneath, you’ll see the same three numbers split by brand and non-brand campaigns, so you can see where the return is coming from — the diagnostic layer sits below the number the fee is calculated on.

If the total is under 5X, the report explains exactly why and what we’re doing about it. If the total is over 5X, we show you which campaigns pulled the weight. Either way, you can audit the math against your own Shopify backend in under five minutes.

Before we take anything on

The first thing we do after an application is an honest tracking audit. Most of the accounts we see have some form of tracking breakage — a Google Tag Manager container firing twice, an enhanced conversions setup that was never actually deployed, a GA4 event definition that’s silently double-counting.

We fix that before any optimisation work begins. It’s the cheapest, highest-leverage thing a new account can do, and it’s the thing most agencies skip because it doesn’t look like “strategy.”

If this sounds like your situation

The form below takes two minutes. It’s a call request, not a commitment. We reply within one business day. If the tracking is fixable and the product economics are real, we’ll tell you.

A clean number is a better conversation than a green chart.