FIELD GUIDE
Why people give up.
What visitors do on almost every site, how to measure it, and the change to try first. Every number links to where it came from.
10 seconds
is when a visitor is most likely to leave
Why people leave in ten seconds
Most people who open a page are gone before they have read a line. Here is what the research measured, which part of that is normal, and what to change on the first screen.
4 reasons, 5 sources
14 in 100
shoppers will never give a store their phone number
Why people abandon forms
A form is rarely abandoned as a whole. People get most of the way, reach one field, and stop. The research below says which fields those tend to be, and what to change about each.
6 reasons, 4 sources
7 in 10
filled carts are never paid for
Why people abandon checkout
About 7 in 10 carts that get filled are never paid for. A good share of that is people browsing, and nothing you do will change it. The rest comes down to a short list of reasons that shoppers name themselves.
6 reasons, 3 sources
57 in 100
seconds of looking go to the first screen
How far people really scroll
People do scroll. They just spend most of their looking at the top, and each screen further down gets less. Where you put a thing decides how many people ever see it.
4 reasons, 2 sources
3 clicks, 1 second
is all it takes to count as one
What rage clicks really mean
Someone pressed the same thing over and over, fast. Usually that means they expected something to happen and saw nothing. It is one of the few signals that points at a single button.
6 reasons, 3 sources
1 line
is all a site built by an agent needs to start counting visits
Analytics for vibe coded apps
You described a site, an agent built it, and it is live. Now you want to know if anyone uses it. Here is how to add analytics to an app built with Cursor or Claude Code, what agents tend to get wrong, and the shortest prompt that gets it right. It is a setup that suits indie hackers and solo founders: one line, no team needed.
6 reasons, 4 sources
10 seconds
is the line between a bounce and an engaged visit in Google Analytics
Your website stats in plain English
Bounce rate, sessions, users, engagement rate. Each one has a plain meaning, and most are a count of how many people did one simple thing. Here is what they mean for a site you run alone, which ones mislead, and what to read instead. It is also how Moxel, a simple website analytics tool, words its numbers.
5 reasons, 6 sources
Not only cookies
are covered by the EU consent rule
Analytics without a cookie banner
Plenty of sites show a cookie banner because of one tool they never thought about. This guide is about when an analytics tool needs one and when it does not. It is a plain-English reading of the rules, not legal advice.
6 reasons, 7 sources
1 field
is the answer you are after, not the whole form
How to track form abandonment
You want to know which field people quit on. Google Analytics can tell you a form was started and sent. It cannot tell you where it stopped. Here are six ways to find out, what each one shows, and how to do it without recording what anyone types.
6 reasons, 7 sources
3 steps
Shopify already counts for you
Shopify traffic but no sales
Your Shopify store has visitors and no orders. Before you change a theme or buy ads, find out where the visitors stop. Shopify already counts three steps for you, and the step that leaks the most points at the fix.
5 reasons, 4 sources
5.1 steps
in the average checkout for a new customer
Where people drop off in checkout
Checkout has steps, and people leave at one of them more than the others. Finding which one is a funnel: count how many people reach each step, then look for the step that loses the biggest share. Here is how to build one with free tools, and how to read it.
5 reasons, 4 sources
14 in 100
shoppers say they would never give a store their number
Should the phone field be required?
A required phone number is one of the fields people stop at most. Some will not hand it over, some type a fake one, and some get told a real number is wrong. Here is when you need it, and how to ask without losing people.
5 reasons, 2 sources
How these are written.
- Every number was read at its source in October 2026, not copied from someone quoting it.
- Each guide says how old a study is and who ran it. When the source sells a tool, it says that too.
- Every reason ends in a change you could make today.
- Studies are averages over other people's sites. Each guide has a section on what it cannot tell you about yours.