The most important number for the invoicing of almost all of the consent manager packages are the PageViews. In order to understand what/how we count, here some details.
What is a “PageView” / “PageImpression” / “Seitenabruf” / …?
In all cases we mean the same: It is the event when a visitor loads a single page on which the CMP-code is installed.
Example:
Visitor A visits: Startpage, Article A, Article B, Article C and the Contact form (=5×1 pages = 5 pageviews)
Visitor B visits: Article C, Article E and Article F (=3×1 pages = 3 pageviews)
Visitor C visits: Startpage and then leaves the website and comes later again to the start page (=2x the same page = 2 pageviews)
Visitor D visits the start page and refreshes the page 3 times (=4x the same page = 4 pageviews)
In above example our system would have counted 5+4+2+4 = 15 PageViews.
More details
- A page view is not the same as a visitor. A visitor usually visits several pages on your website, hence there will be more pageviews than visitors.
- A page view is independent from whether the consent layer was shown or not.
- A page view is counted on all pages which have our CMP-code installed.
- A page view is also counted each time when the user clicks on refresh or the page reloads for different reasons.
- A page view from a search engine crawler is usually not counted, as long as the crawler follows robots.txt. See Bots, crawlers and AI crawlers below for what this does and does not cover.
Bots, crawlers and AI crawlers
We keep crawlers out of your PageViews in two ways. We publish a robots.txt file on the server that delivers the CMP, which tells crawlers not to request our counting endpoints, and we exclude a small number of common bots by their User Agent. A crawler that follows robots.txt never reaches the counter, so its visits never appear in your PageViews.
That covers the large search engines, which do follow robots.txt. It does not cover every bot:
- A crawler that ignores robots.txt and runs JavaScript loads our CMP code the same way a browser does, so we count the visit as a PageView.
- A crawler that sends an ordinary browser User Agent looks like a visitor to us. We have no reliable way to tell it apart from a real one.
- A crawler that only reads the HTML, without running JavaScript, never loads our CMP code at all. It is never counted, whatever it calls itself.
Several AI crawlers publish their User Agent and follow robots.txt, so they behave like a search engine crawler and stay out of your numbers. Others do not identify themselves, and some render pages with a full browser engine. Those we count as PageViews, because from our side they are indistinguishable from a visitor.
Please note
Our bot filtering is lighter than the filtering in analytics tools. An analytics tool can apply its own bot list and discard traffic after the fact, while we count the request our servers actually answered. If your analytics shows fewer page views than we do, this is one of the usual reasons.
How to spot bot-driven PageViews
No report shows the User Agent behind a PageView, so you cannot confirm bot traffic from inside consentmanager alone. You can look for the pattern:
- Compare the two figures. Put your analytics page views next to ours for the same period. A gap that widens suddenly is worth a look.
- Check the ignore rate. Bots do not click anything on the consent layer, so they push the Ignore rate up while Accept and Reject stay flat.
- Group by country. Look for a jump from a country or region that does not match your audience.
- Group by browser, OS and device. Automated traffic often clusters on one browser version, or appears on desktop only.
- Read your own server logs. This is the one place the actual User Agent appears. Look for named AI crawlers, or for requests arriving at machine-like intervals.
Detailed reports cover the last 45 days, so run these checks inside that window. Older traffic still counts toward your totals, but we can no longer break it down by country, browser or device.
What you can do about it
- Name unwanted crawlers in your own robots.txt. A crawler that honors it stops fetching your pages, and a page we never serve is a page we never count.
- Block the rest before they reach your pages. Crawlers that ignore robots.txt only stop at your CDN or firewall, through its bot rules.
- Get in touch if a bot spike pushed you over your page view allowance. We look at these case by case. If we confirm the extra traffic came from bots, we can correct the page views counted for that month and reissue the invoice. If your analytics gives a figure for the bot-driven page views, send it with your message, because that figure is what we use to work out the correction.
Why do we count the PageViews?
The main cost for our service are the network infrastructure and servers that we need to have in order to deliver the CMP-logik to the visitor. The more pageviews a website has, the more requests the visitors’ browsers will send to our servers, and the more server capacity we need.
Why not only count when the Consent Layer opened/showed?
The reason we are counting pageviews and not visitors or consent views is that our servers need to do the same amount of effort whether or not the consent layer is actually shown on the page or not. Even if the layer is not shown, in the background there will be our CMP script running, ensuring that all third party codes get blocked and enabling the visitor to change their choice.
How does a PageView count in an App?
In an app environment you will use our SDK in order to show the consent layer and perform the necessary tasks. The SDK will query our servers once a day when each user starts your app, in order to find out if settings changed or the user needs to see the consent layer again. This query will be counted as a PageView. Hence in an inApp environment you can see “PageViews” similar to “daily active users”.