Conversions
How many sales Google says it drove — counted Google's way, not your shop's.
What it means
Conversions is the count of tracked conversion actions Google credits to your ads in the period. It counts whatever you marked as a conversion in the Google Ads account — usually purchases, sometimes sign-ups and calls as well — and it arrives as a decimal, because credit for one order is split across the clicks that led to it. Expect it to disagree with your Shopify order count; the two measure different things.
Show the math
Formula and a worked example
Tracked Google conversions are the actions marked as conversions in the account. Add a newsletter sign-up to that list and this number jumps overnight.
Fractions are normal. Data-driven attribution hands one order to several clicks in slices, so 240.6 isn't a rounding fault. Google records it against the click date, not the day the order landed, so the last few days keep filling in for a week.
Worked example. March: €12,000 of spend and 240 conversions, so each one cost €50. Shopify shows 265 Google-attributed orders for the same month — Google credits the click, Shopify the last session.
It answers the question
Did the money do anything? A rising count on flat CostCostRead against conversions and value-per-cost, never alone. means the account is getting better; a flat count on rising Cost means you're paying more for the same result.
Why it matters
Every efficiency figure in Google Ads stands on this one. Cost/Conv.Cost/Conv.What each Google sale costs you. divides by it, so adding or removing a conversion action moves your cost per sale without a single bid changing.
It's also the number Google optimises towards. Smart Bidding chases whatever you've marked as a conversion, so a cheap, low-value action on that list will be bought for you in volume.
What good looks like
There's no universal target — a count only means something against the size of the shop that produced it. Judge it against your own trend at a steady Cost, and against your own category, since a considered €600 purchase converts at a fraction of the volume a €30 impulse buy does. The question is never whether 240 is good, but whether 240 beats last month for the same money.
How to improve it
| Lever | What you do | Expect | How long | Watch out for |
|---|---|---|---|---|
| Fast Count only what you actually sell | Take secondary actions out of the conversion column so purchases are all that's counted | Conversions fall on paper, Cost/Conv. rises to the truth | 1–2 days | Smart Bidding loses signal volume and re-learns for a week or two, and your history stops comparing like for like. |
| Fast Fix what happens after the click | Send each ad group to the product page it promised, not the category page or the homepage | Conversions up 5–15% on the same Cost | 1 week | Deep links break when products go out of stock, so somebody has to watch the URLs. Narrower ads also match fewer queries, so Search IS can slip. |
| Slow Widen what people can buy | Add the sizes, bundles and payment options that people ask for and then abandon over | More conversions from the same clicks | 4–8 weeks | Merchandising work, and bundle pricing dilutes Gross Margin — the count can rise while each order is worth less. |
| Slow Buy demand you're not showing for | Raise budget where Search IS says you're missing auctions you can afford | Conversions up, Cost up with them | 1 quarter | The new auctions are colder and dearer, so Cost/Conv. normally gets worse for a few weeks before the extra volume pays for it. |
Every lever costs something somewhere. The last column is the one to read twice.
Read it with
The count says how much you sold. The price beside it says whether you can do it again.
Scaling cleanly
More sales at a lower price each. Move in steps rather than all at once, so you can see the point where it stops working.
Growth you're paying up for
Normal when you scale into new auctions. It's fine while each sale still covers its costs and a problem the moment it doesn't.
Smaller and sharper
You cut the waste and kept the efficiency. There's room here; the account isn't broken, it's throttled.
Losing both ways
Fewer sales at a higher price usually means something broke rather than drifted — a tag, a page, or a product that went out of stock.
A count with no price is a vanity number and a price with no count is a rounding exercise. Together they say whether growth was bought or earned: 300 sales at €40 and 300 sales at €70 look identical on this card and are completely different businesses.
Together they tell you whether flat sales are a demand problem or a coverage problem. Flat conversions with Search IS at 80% means you're already in nearly every auction, so the next gain has to come from the shop. Flat conversions at 30% means the demand is running and you're not turning up for it.
Common misreads
Expect a gap. Google records against the click date and splits credit across touchpoints; Shopify records the order date and the last session. Some difference is ordinary; a large one is worth chasing.
Check how recent that week is. Conversions backfill for days after the click, so the most recent stretch always reads low and always recovers. Judge a fortnight, not a Tuesday.
Only when they're the same kind of conversion. A count padded with newsletter sign-ups climbs while revenue falls, and Smart Bidding will happily keep buying the cheap action.
Also called
Google conversions · tracked conversions · conv.
See yoursYour conversions for the period, next to the spend and the cost per sale that produced them.
Open Google Ads →