I recently sat down to audit a Google Ads account for a speciality food and ingredients store in the US. They import lovely products, they have a brilliant organic search presence, and they run on roughly 75% gross profit margins, which is the kind of margin most store owners would happily trade a kidney for.
They had been running Google Ads on and off for years, mostly Performance Max, and yet they never fully trusted the numbers staring back at them. When I lifted the bonnet, I understood exactly why. The account had three problems quietly tangled together, and once I pulled them apart, the path forward was obvious. So let me walk you through exactly what I found, and what I told them to do, because I would bet good money your account has at least one of these same issues.
The Conversion Tracking Was Lying
The first thing I always check is whether the numbers are even real, and in this case they were not.
The account was using New Customer Acquisition bidding, which artificially pads your conversion values by adding extra revenue, around $46 in their case, for every new customer Google brings in. That is a perfectly valid setting if you know it is switched on, but it means the Conversion Value column you have been staring at is not your actual revenue.
When I added the Original Conversion Value column next to the standard Conversion Value, the truth came out. The actual revenue was only around 63 to 64% of what was being reported. The reported number was nearly double the real one. On top of that, the historic setup had at some point been counting non-purchase events like add-to-cart as if they were purchases, which muddied years of data.
So here is my first and most important bit of advice. Before you trust a single ROAS figure, add the Original Conversion Value column. It strips out the padding and shows you the actual revenue from your conversions. If that number is wildly different from your headline Conversion Value, you have found your first problem. And when you do flip off a setting like New Customer Acquisition padding, write down the date you did it. I cannot stress this enough, because in two months when you are comparing performance, you will thank yourself for knowing exactly when the numbers changed.
One quick aside: I would not bother trying to verify any of this against the new Revenue metric in Google Ads. It is still in beta, it does some murky calculation involving revenue minus cost of goods sold, and it is simply not reliable enough to check your working against. Use Original Conversion Value instead.
The ROAS Target Was Strangling the Account
Here is where it gets interesting, and where most store owners have their lightbulb moment.
This account was targeting an 800% ROAS, and it was smashing it, consistently pulling in 10 to 17 times return. On paper that sounds fantastic. In reality it was a flashing red warning light. When you are hitting 10x to 17x against an 800% target, it means Google is being held on such a tight leash that it can only chase the cheapest, easiest, most obvious sales. Everything slightly more expensive to win is left on the table for your competitors.
That is the bit so many people miss, and I want to be really clear about it. High ROAS is not the same as high profit. An eye-watering ROAS usually means you are underspending and missing a pile of profitable sales you could be making.
Think about the profit curve for a second. As you lower your ROAS target, you give Google permission to bid higher, capture more traffic, and generate more sales. Yes, each sale comes back at a lower efficiency, but the extra volume more than makes up for it. The relationship between ROAS and profit is not a straight line. The sweet spot, the point of maximum profit, sits somewhere between maximum efficiency and maximum volume, and it is almost never up at 800%.
Profit Is the Metric That Actually Matters
The root cause of all of this was a simple misunderstanding. The client was laser focused on protecting that high ROAS without realising that ROAS is just an efficiency metric, not a profit metric. There was no profit tracking in the account at all.
So I set one up, and you can do exactly the same in about two minutes. Create a custom column called Gross Profit After Ad Spend. The formula is dead simple:
(Conversion Value x Gross Profit Margin %) - Cost
With their 75% margin, that becomes (Conversion Value x 0.75) - Ad Spend. Now, instead of guessing, you can see the actual money left in your pocket after Google has taken its cut and the goods have been paid for. This single column changes how you look at the whole account.
From there I gave them what I call the Core 4, the only four metrics you really need to watch together: Ad Spend, ROAS, Revenue, and Profit. The trick is to never judge any one of them in isolation. A lower ROAS looks scary on its own, but if Ad Spend, Revenue, and Profit are all climbing alongside it, that is exactly what winning looks like.
How To Find Your Own Sweet Spot
So what did I actually tell them to do with that 800% target? You have two ways to play it.
The conservative approach is to hold at 800% for a month or two until everything stabilises after the tracking fix, then gradually walk it down towards 400% over the following two to three months. Less volatility, lower risk, slower payoff.
The aggressive approach is to drop straight to 400% now. You reach the optimal spend much faster, but you live through more volatility during the adjustment, and you need the nerve to ride out the fluctuations.
Whichever route you pick, the real answer comes from testing systematically. Run a target for a month, note the profit, then step it down: 800%, then 650%, then 500%, then 400%, giving each level a month or two to settle. Compare the profit at every step and you will literally watch the profit-optimal ROAS reveal itself.
And if you want a floor to never drop below, calculate your breakeven ROAS. The formula is 1 divided by your gross profit margin. For this store, 1 divided by 0.75 is 1.33, so a 133% ROAS is the point where they make nothing. You always want to be comfortably above that, which is why I suggested a sensible starting floor of around 300%, or 3x, while testing.
A Few Things To Stop Worrying About
Two final points that will save you a lot of fretting.
First, do not micromanage where Performance Max spends its money. In a healthy account your Shopping ads, the ones using your product data, should be doing the bulk of the spend at the highest ROAS. Your Search text ads usually take 15 to 20% at a slightly lower but still profitable return. Display, Gmail, YouTube, and Discover mop up retargeting with whatever is left. As you lower your ROAS target, expect more budget to naturally flow into that retargeting. Trust Google to allocate it; just steer with your ROAS target.
Second, stop trying to reconcile everything against Shopify. Its channel attribution is genuinely unreliable for working out which marketing source drove a sale. Use Google Ads and GA4 for that. Shopify is only good for confirming your total revenue, nothing more granular.
And one gentle warning before you go changing everything at once. Any big change to your ROAS target or bidding settings will trigger a learning period, and performance will wobble before it settles at its new level. That is normal. Do not panic and yank it all back two days in.
The headline lesson from this whole audit is one I want you to tattoo somewhere visible: a high ROAS is not a trophy. Fix your tracking, track your actual profit, and find the ROAS target that fills your bank account, not the one that looks impressive in a screenshot.