In short: Clicks, impressions and cost per click (CPC) tell you how many people saw and clicked an ad, not whether it brings in money. That is shown by cost per enquiry or order (CPA), revenue against ad spend (ROAS) and the share of visitors who send an enquiry or place an order. All of these numbers only mean something if your website records every enquiry and every order exactly once.
Why don’t clicks tell you whether advertising brings in money?
Advertising platforms put first what they can count themselves: impressions, clicks, CPC and CTR. These numbers are useful to the person setting up the campaign, because they show whether the ad reaches anyone at all. For a business owner, though, the question is different: how much money comes in for the money that goes out.
A cheap click can turn out to be expensive. An ad with many low-cost clicks often attracts curious people who are not looking for what you sell. The opposite happens too: an ad with few, expensive clicks may bring exactly the people who buy. If you judge by CPC alone, you will stop the second ad and increase the budget of the first.
There is one more reason. Advertising platforms show ads in a way that achieves whatever result you set as the goal. If the goal is clicks, they look for people who click easily. If the goal is an enquiry or an order, they look for people who get that far. So your choice of metric affects not only the report but also who sees the ad.
Which metrics link advertising to money?
Short definitions are collected in the metric definitions. The table below shows at a glance when each metric helps and where it can mislead.
| Metric | How it is calculated | When it helps | Where it misleads |
|---|---|---|---|
| CPC and CTR | cost per click; clicks as a share of impressions | whether the ad attracts attention at all | says nothing about enquiries or revenue |
| Cost per enquiry (CPA) | ad spend / number of enquiries | services and B2B, where the sale comes later | counts weak enquiries the same as strong ones |
| Cost per order (CPA) | ad spend / number of orders | online stores | does not show how large the order is |
| ROAS | order revenue / ad spend | when orders vary in value | shows revenue, not profit |
| Conversion rate | enquiries or orders / visits | whether the problem lies in the ad or on the site | does not measure money directly |
Cost per enquiry
Ad spend divided by the number of enquiries it brought in over the same period. This is the main number for companies that sell services or work with other businesses: the customer first writes or calls, and the sale comes later. Its weak spot is that it counts all enquiries the same. Ten enquiries from people who want something you do not offer look exactly like ten enquiries from future customers in the report.
Cost per order
Ad spend divided by the number of orders it brought in. This is the main number for online stores. It tells you what one order costs, but not how large it is. Two campaigns with the same cost per order can bring in very different revenue if one sells inexpensive products and the other expensive ones.
ROAS: revenue against ad spend
ROAS makes up for the weak spot of cost per order, because it takes order value into account. It shows revenue, however, not profit. A campaign with a good ROAS can lose money if it sells low-margin products or if delivery and return costs are high. That is why the acceptable level is worked out from the margin on your products, not taken from market averages.
Share of visitors who send an enquiry
How many of the people who arrive from the advertising send an enquiry or place an order. It suggests where the problem is. If the advertising brings many visitors and the share is low, the cause is often on the site: slow loading on a phone, an unclear offer, a long form. If the share is good but enquiries are few, the advertising brings too few visitors.
What do the numbers only show together?
No single metric is enough on its own. Decisions become clear when you look at several numbers together:
- Many clicks, a low conversion rate and a high CPA: first check whether the ad promises the same thing people see on the landing page, then check the page itself.
- An acceptable CPA, but the sales team says enquiries are weak: the problem is probably the targeting or the message, and the advertising attracts people who are not your customers.
- Good ROAS, but not enough money at the end of the month: work out profit by product, as the advertising is probably selling mainly low-margin products.
- A high cost per order with customers who buy repeatedly: the first order may not cover the cost, but the profit from a customer over time may. You make this calculation with your own repeat order data.
ScaleLab tip: Once a month, take ten enquiries from your advertising and ask the sales team which of them were real opportunities. It shows the difference between a cheap CPA and a useful one, which the report alone cannot.
Why does your conversion data decide what the platform optimises for?
Google Ads is putting more and more decisions in the hands of automation. At Google Marketing Live on 20 May 2026, Google announced new AI Max capabilities for Search, Shopping and Travel. Dynamic Search Ads are also moving to AI Max: from September 2026, campaigns using automatically created assets or campaign-level broad match are upgraded automatically. The automatic upgrade of the remaining DSA campaigns begins in February 2027. Following Google’s June 2026 update, creating new DSA campaigns has been possible again since 15 June 2026, but that option ends in January 2027.
What does this mean for your business? When the system chooses the searches, the ad text and the pages itself, your job is to tell it correctly what success looks like. If every visit to the contact page counts as a conversion, the automation will look for people who open that page. If real enquiries and orders with their value are counted, it will look for people who get that far.
According to Google’s internal 2026 data for advertisers outside retail, AI Max delivers on average about 7% more conversions or conversion value at a similar CPA or ROAS when all of its features are used, compared with search term matching alone. That is Google’s own claim, not an independent measurement, and the outcome depends on the data the system works with. What to review during the switch is covered in the article on AI Max in Google Ads.
ScaleLab tip: Before you turn on AI Max or any automated bidding strategy, open the list of conversion actions in Google Ads and keep as primary only the actions that really bring in money. Set everything else as secondary: Google Ads reports it for observation but as a rule does not use it for bidding.
Which channel gets the credit?
Every advertising platform records results in its own way. A customer may see an ad on social media, search for the company on Google a few days later and send an enquiry. Both platforms may count that enquiry as their own. If you add up the numbers from the two reports, you will get more enquiries than you actually received.
That is why the number of enquiries and orders comes from your own system: the website, your customer system or even a spreadsheet in which the sales team notes where each customer came from. Advertising reports serve to compare campaigns within the same platform, and the overall picture is checked against the number the business actually received.
Keep the role of each channel in mind as well. The advertising through which a customer first hears of the company rarely gets the last click before the enquiry. If you stop everything that does not get the last click, searches for your company name may fall over time. So pause such a campaign temporarily first and watch the total number of enquiries in your own system.
How do you check the numbers before trusting them?
Every metric so far depends on one condition: the website has to record exactly what happens. The common problems are well known and easy to check:
- Double counting: an enquiry is recorded twice, for example when the thank-you page is reloaded.
- Missing calls: clicks on the phone button on the site are not counted at all.
- Wrong value: the store records the order but not its value, or it is unclear whether the value includes VAT and delivery.
- Returned orders: cancelled or returned orders stay in the report as revenue.
Send a test enquiry and place a test order, then check that each appears exactly once and with the correct value. Until this is right, every budget increase is made blind. If the numbers look fine but the business is not growing, see also the signs that your marketing strategy is not working.
How should you decide on the budget?
Start with the question of how much you can pay for one enquiry or one order. For a company that sells services, this depends on two things: what share of enquiries become customers and how much you earn from one customer. For an online store, it depends on the margin and on how often customers come back. This gives you a limit that comes from your business, not from a market average.
Then compare campaigns against that limit:
- Above the limit after several attempts to improve: the campaign stops, and the budget goes where the cost is below the limit.
- Little data: do not judge by a few days, because with a small number of enquiries chance carries a lot of weight.
- Changes: change one thing at a time so you know what caused the difference.
When advertising is managed this way, the monthly report starts from revenue and cost, and clicks and impressions stay in it as explanation. That is how we organise work on paid advertising: the website and tracking first, then the budget, and every month a decision on what continues and what stops.
Frequently asked questions
What is the difference between CPA and ROAS?
CPA shows what one enquiry or one order costs. ROAS shows how much revenue each euro spent on advertising brings in. CPA is more useful for services, where the value of an enquiry is not known straight away, and ROAS for online stores whose orders vary in value.
What is a good ROAS?
There is no single number that applies to everyone. An acceptable ROAS is worked out from the margin on your products and your delivery and return costs. A campaign with a high ROAS can still lose money if it sells low-margin products.
Why do Google Ads and my website show different numbers of conversions?
Each platform records results using its own model and may claim an enquiry that also came through another channel. The figure in the platform is its own estimate. For decisions, use the number of enquiries and orders from your own system.
What should I check before turning on AI Max?
Check that only real enquiries and orders are set as primary conversions and that their values are correct. Review the brand, URL and negative keyword controls as well. The automation optimises for the data you give it.
Sources
- Google: We’re upgrading Dynamic Search Ads to AI Max (15 April 2026, updated 11 June 2026)
- Search Engine Roundtable: Google Ads Extends Dynamic Search Ads (DSA) Timeline (12 June 2026)
- Google: Google Marketing Live 2026, News and announcements (20 May 2026)
- Google Ads Help: About primary and secondary conversion actions