Returning visitors

The Returning visitors page answers one honest question: when someone visits your site for the first time, do they come back? It is easy to bring new people in once. The real sign of a healthy site is whether those people return on their own. This page measures exactly that, and it shows it as a simple grid so you can read it at a glance.

Think of it like planting seeds. Every week you plant a fresh batch (the new visitors who came that week). The grid then shows you how many of each batch were still growing one week later, two weeks later, and so on. Some batches thrive and keep coming back. Some fade fast. The grid lets you compare batch to batch and see which weeks planted the strongest roots.

What you see

From top to bottom, the page is built from a few clear parts.

A page title that reads “Returning visitors”, with the date range you are looking at shown nearby, for example “Last 90 days” with the two calendar dates spelled out.
The grid. This is the heart of the page. Each row is a group of people who first visited in the same week. The top row is the most recent group, and older groups sit below. So one row is “everyone who first came in the week of March 3”, the next is “everyone who first came in the week of March 10”, and so on.
The columns read across as time passes. The first column is the size of the group, meaning how many new visitors first arrived that week. The next columns show how many of that same group came back 1 week later, 2 weeks later, 3 weeks later, and onward. Each cell is usually shown as a percent of the starting group.
Color shading on the cells. A stronger, darker shade means a higher share of people came back. A pale cell means few returned. So a healthy site shows color holding across the row, and a leaky site fades to pale quickly as you read right.

Each row tells the story of one batch of new visitors over time. Reading down a single column instead lets you compare the same moment (say, one week later) across many different starting weeks.

How to read it

Read one row across to follow a single group, and read one column down to compare groups.

Here is a worked example. Say you group by week and look at the last several weeks. One row, for the group of people who first visited in the week of March 3, might read:

Starting size: 1,000 new visitors that week.
1 week later: 28 percent came back (280 people).
2 weeks later: 19 percent came back.
3 weeks later: 15 percent came back.
4 weeks later: 13 percent came back.

What does this tell you? Out of 1,000 brand new visitors, a bit more than a quarter returned the very next week, and the share settles down to around one in eight a month later. The first drop, from 100 percent to 28 percent, is the biggest and is normal, because most people who try a site once never come back. What matters more is the shape after that. A row that holds steady around 13 to 15 percent for weeks is a sign your site has real staying power. A row that fades all the way to near zero means people are not finding a reason to return.

Now read down a column instead. Look at the “1 week later” column across several starting weeks. If older weeks returned around 22 percent but your most recent weeks are returning 30 percent, something you changed lately is making new visitors come back more often, and that is a win worth understanding and repeating.

Customize it

You shape the page with the controls at the top. Here is each one and when to use it.

Group by day or week. Choose whether each row is a group of people who first visited in the same day or in the same week. Use weekly groups for most sites, because they are steadier and easier to read, and because daily numbers can be noisy on smaller sites. Use daily groups when you have a lot of traffic and you want a finer, closer look, for example right around a launch when a single day matters.
Set the date range. Use the calendar to pick the period the grid covers, such as the last 30, 90, or more days, or a custom start and end day. A longer range gives older groups more time to show whether people kept coming back, so the grid has more columns to read. Use a longer range when you want to judge real staying power, and a shorter range when you only care about very recent groups.
Scope to one website. If you track more than one website, narrow the grid to a single site so its returning visitors are not mixed with the others. Use this when one site has a very different audience, or when you simply want one site’s grid on its own.

Use cases

See if a product change made people come back more. Pick weekly groups and a date range that covers several weeks before and after the change. Read the “1 week later” and “2 weeks later” columns down the grid, comparing groups from before the change to groups from after. Action: if the groups that first visited after your change return at a higher share than the groups before it, the change is helping people come back, so keep it and build on it. If the share did not move, the change may look nice but is not yet giving people a reason to return.
Measure how sticky your site is. Look at where each row settles after the first few weeks. A site that holds a steady share of returning visitors week after week is sticky. One that fades to near zero is not. Action: if your rows fade fast, people are not finding a lasting reason to return, so add something that pulls them back, like fresh content, saved progress, or a reminder, then watch later groups for a steadier shape.
Compare two launch weeks. Find the two rows for the weeks you launched something. Read them side by side across the columns. Action: if the visitors from one launch week keep coming back far more than the other, look at what was different about that launch (the message, the audience, the offer) and lean on what worked next time.

Tips

The first drop is always the biggest, and that is normal. Almost no site keeps most of its first time visitors. Judge your site by the shape after that first week, not by the fact that most people did not return at all.
Read down a column to compare, read across a row to follow. Reading across one row tells the story of a single group over time. Reading down one column lines up the same moment across many groups, which is how you spot whether things are getting better or worse.
Weekly groups are steadier than daily ones. On most sites daily numbers bounce around too much to read clearly. Start with weekly groups and only switch to daily when you have the traffic and a reason to look that closely.
A flat line low down is still a good sign. A row that holds even a modest share of returning visitors for many weeks means you have a loyal core. Flat and holding beats high then crashing to zero.
A brand new site has almost nothing to show here yet, and that is expected. People need time to come back before any return can be measured, so the newest groups will look empty in their later columns simply because not enough time has passed. Give it a few weeks of real traffic.