Blog
Writing
What we have had to work out to build this, written down properly. Mostly the law, occasionally the arithmetic.
What this blog covers
Two subjects keep coming up when you build an analytics tool that refuses to identify anybody. The first is legal: what consent actually requires, and which of the scripts on an ordinary web page trigger it. The second is arithmetic: what the numbers on a dashboard mean, which ones are lying to you, and what you can honestly conclude from a measurement that counts visits rather than people.
Everything here is written for somebody who has to act on it — and, where it matters, with the article, the recital or the regulator decision named so you can go and check rather than take our word.
4 pieces
Privacy and the law
Consent banners, the GDPR and the ePrivacy rules, and what each of them actually asks of a website that wants to measure its own traffic. Written for the person who has to explain the decision to somebody else, with the articles and the regulator decisions named so they can be checked rather than taken on trust.
-
Do you actually need a cookie banner?
The rule is simpler than the advice around it, and it is not the rule most people think it is. Here it is, with a flowchart.
Read it -
Removing the cookie banner, and what is actually keeping it there
Swapping your analytics and leaving the rest in place is how people do a week of work and still have a banner. Start with an inventory.
Read it -
Why everybody looks new, and what you still get inside a day
Somebody who reads your site every morning appears as a new visitor every morning. That is the mechanism working, not a gap in the data.
Read it -
What makes web analytics GDPR-compliant
Two separate laws apply to analytics, and most write-ups conflate them. Here is what each one asks, in the order a regulator asks it.
Read it
10 pieces
Making sense of numbers
What the figures in an analytics dashboard mean, where they mislead, and which ones are worth acting on. Bounce rate and why Google changed its definition, how to read a click or scroll map without fooling yourself, what a session recording really stores, and how far you can trust an estimate of somebody else’s traffic.
-
Dead clicks and rage clicks, and how to fix them
Two signals hiding in an ordinary click map. Both point at something specific and broken, and both are usually a quick fix.
Read it -
Why two analytics tools never show the same number
Install two analytics tools and they will disagree, every time. Here is exactly where the gap comes from.
Read it -
How to check website traffic, yours or anyone else's
Checking your own traffic and checking a competitor’s are completely different problems. Only one of them has a right answer.
Read it -
What session replay actually records
Everyone sells this as watching a video of your visitor. It is nothing like a video, and the difference is where the risk lives.
Read it -
The first week after switching, and what to expect in it
A drop in a dashboard on migration day looks exactly like a drop in business. Almost everything that worries people in week one is expected.
Read it -
Pageviews, visits, visitors and events, defined exactly
Two dashboards disagree before either has counted anything wrong, because the words underneath them mean different things.
Read it -
What a scroll map is counting, and the two readings worth acting on
The most quoted number on a heatmap and the most often misread. Two of the three obvious readings are wrong, and both of them flatter the page.
Read it -
Which form field people give up on, and how to tell why
Most forms are abandoned at one field, and it is rarely the field anybody suspects. The counters say what kind of trouble it is causing.
Read it -
How to read a heatmap, and what to look at first
Three kinds of map, three different questions. Here is what each one is good at, and the handful of checks that turn a nice picture into something you can act on.
Read it -
What is a good bounce rate, really?
Google changed what this number means and most of the advice online never caught up. Here is what it measures now, and when it is lying to you.
Read it
Elsewhere
If you want the reference material instead
The writing argues; the documentation specifies. If you want what is collected field by field, how erasure requests are handled or the full Stats API, start there. How cookieless analytics works is the mechanism in one page, and how to judge an alternative is eight questions that name nobody.
Posts get written when there is something worth writing down rather than to a schedule, and nothing here is published to fill a slot.
Ready
The short version: no banner, no cookies, no IP addresses.
Everything the writing argues for is how the product already works. Paste one script tag and see for yourself.