You have the Google Ads account open. There are tabs, columns, and a lot of numbers, and it is not obvious which ones decide whether the money is working. Most beginners end up staring at impressions and clicks because those are the biggest numbers, when the metrics that decide success are further down the row. This guide explains how an account is laid out and the handful of metrics that actually tell you what is going on. Each metric comes with its formula, so you can read a row without guessing, and you can apply the same reading whether the account spends fifty dollars a month or fifty thousand.
How an account is structured
Google Ads nests four levels, from broadest to narrowest. Reading top to bottom keeps you from drowning in detail before you know where to look.
- Account: the whole advertiser. Everything below rolls up here, and it is where billing and conversion tracking live.
- Campaign: a budget and a goal. Each campaign has its own daily budget, target locations, and type (more on types below).
- Ad group: a theme inside a campaign. It bundles closely related keywords with the ads that should serve against them.
- Keyword / Ad: the keyword is the search term you bid on; the ad is the creative that shows. This is the lowest level, where you see exactly what people searched and clicked.
When a number looks off at the account level, you drill down: account to campaign to ad group to keyword, narrowing until you find the part doing the damage (or the good work).
The core metrics, with formulas
Almost every report is built from the same short list. Learn these and most dashboards stop being mysterious.
- Impressions: how many times your ad was shown. A raw reach count, not a measure of quality.
- Clicks: how many times someone clicked the ad.
- CTR (click-through rate) = clicks ÷ impressions. The share of people who clicked after seeing the ad. It tells you whether the ad and keyword are relevant to the search.
- Cost: total spend over the period. The denominator in most of the ratios that matter.
- CPC (cost per click) = cost ÷ clicks. The average price you paid for a click. Useful for spotting expensive keywords.
- Conversions: the valuable actions you count: a purchase, a lead form, a call. You define what counts in conversion tracking.
- Conversion value: the revenue (or assigned worth) of those conversions. For lead gen you may set a fixed value per lead.
- ROAS (return on ad spend) = conversion value ÷ cost. Read it as "for every $1 you spend, you get $X back." A ROAS of 4 means $4 returned per $1 spent.
- CPA (cost per acquisition, also cost per conversion) = cost ÷ conversions. What each conversion costs you. The number to compare against your target.
Ratio totals are weighted, not averaged. The overall CTR is total clicks ÷ total impressions, not the average of each row's CTR. Averaging percentages lets a tiny, high-CTR keyword distort the whole picture. The same goes for blended CPC, ROAS, and CPA: always divide the totals.
Campaign types in one line each
- Search: text ads on the Google results page, triggered by keywords. The most transparent type, since you can see the exact queries.
- Performance Max: one campaign running across all Google inventory (Search, Display, YouTube, Gmail, Maps), largely automated, with limited visibility into individual placements.
- Shopping: product listings with image and price, driven by your Merchant Center product feed rather than keywords.
Spend tells you what you paid. ROAS and CPA tell you whether it was worth it.
What to look at first
When you open an account, resist the urge to scan every column. Work in this order:
- Cost: are you spending roughly what you expect for the period?
- Conversions: are the valuable actions actually happening?
- ROAS or CPA against target: for an ecommerce account, is ROAS above your break-even? For lead gen, is CPA at or below what a lead is worth?
- Drill by campaign: once the headline looks right or wrong, sort campaigns by cost and by ROAS to see which ones carry the result.
That sequence answers the only question a client really asks: did the money do its job, and where? A campaign can have a high CTR and a healthy click count and still lose money if the conversions are not there, so resist judging an account on traffic alone. Conversions and conversion value also depend on tracking being set up correctly, which is far easier to trust once Ads and your analytics agree on the same numbers. Our guide on the Google Ads and GA4 connection walks through that, and it is worth doing before you start drawing conclusions from conversion data.
Cheat sheet
| Metric | Formula | Why it matters |
|---|---|---|
| CTR | clicks ÷ impressions | Whether the ad and keyword match the search |
| CPC | cost ÷ clicks | The average price of a click; flags expensive terms |
| Conversions | counted actions | Whether the valuable actions are happening |
| ROAS | conversion value ÷ cost | Return per $1 spent; the ecommerce headline |
| CPA | cost ÷ conversions | Cost of each conversion; the lead-gen headline |
In Clearly
Connect Google Ads once and your campaigns flow into a report without exports or copy-paste. The headline scorecards put cost, clicks, CTR, conversions, ROAS, and CPA in a single row, each with a period-over-period delta coloured the right way (cost down is good, ROAS up is good), so a glance tells you whether things moved in your favour.
Because Ads and GA4 live in the same report, you can put spend next to the outcomes it drove: sessions, conversions, and revenue side by side, in one client-ready page you can share with a link. No screenshots, no spreadsheet stitching, and the weighted totals are handled for you.
The short version
Read an account top to bottom (account, campaign, ad group, keyword) and judge it on a few metrics, not the whole column wall. Start with cost, then conversions, then ROAS or CPA against your target, and drill by campaign to find the cause. Keep the four formulas (CTR = clicks ÷ impressions, CPC = cost ÷ clicks, ROAS = conversion value ÷ cost, CPA = cost ÷ conversions) within reach, remember that totals are weighted, and the numbers stop being noise.