Subscriber analytics

The subscribers analytics page answers three questions: how fast the list is moving, which channels brought people in, and who is still reading. It sits beside your subscriber list, one page per newsletter, and it needs nothing set up.

Growth over time

Growth plots active subscribers at the end of each day, over 30, 90 or 365 days, and opens on 90. Hover any day for the list size, how many joined and how many unsubscribed.

It is the same chart as the one on your dashboard, for the same reason: one line for the list, with joiners and leavers in the tooltip rather than crammed onto a second axis where they would distort the shape of it. Days are counted in your newsletter’s own timezone, today’s figure refreshes hourly, and earlier days are finalised overnight.

One honest limitation. Each day’s figure is a snapshot of the list as it stood, and there is no membership history behind it, so a day cannot be recalculated exactly once you have deleted a subscriber. That is a deliberate trade rather than an oversight: the alternative is a permanent record of everybody who ever gave you their address, which is more data about more people than a newsletter has any business keeping.

Where your readers came from

Five sources are counted: your embedded form, your newsletter website, an import, addresses you added by hand, and test addresses from the browser extension. Each shows its total and its share, and clicking one filters your subscriber list to the people who arrived that way.

The count is everybody who ever joined through that channel, not only the people still subscribed. That is the point of the panel: it is asking which channel brings people in, and a channel that brings in people who then leave should still be visible. The signup forms and archive pages behind the first two sources are the pair worth comparing, because they reach different people: one catches visitors to your own archive, the other catches readers of whatever site you put it on.

How many each source kept

Under each source is the number still subscribed and the share of that channel’s arrivals it represents. Two channels of the same size are not the same channel if one has kept most of its readers and the other has kept a third of them.

Kept means active right now. An address that never confirmed its double opt-in, one that unsubscribed, one that bounced and one that complained all read the same way in that figure, which is deliberately unforgiving: a channel is only worth what it left you with. It is a figure worth looking at again a month after any import, rather than on the day the file went in.

Churn, over two windows

Churn is the share of the list that left, over the last 30 days and the last 90 days, measured against the list as it stood the day each window opened.

Measuring against the opening size rather than today’s is what keeps the figure honest. Leavers are drawn from the people who were already there, so dividing by a larger present-day list would make a growing newsletter look healthier than it is. Each window also shows how many people left, out of how many, and the net change once joiners are counted, so a healthy month and a stagnant one do not read alike. A window with no history before it shows nothing rather than a made-up percentage. What happens when somebody unsubscribes covers the mechanics on the other side of that number.

Who is still reading

Three segments are counted over your last 5 sent issues: opened or clicked, received and did neither, and not sent to yet.

Opened or clicked means the reader was sent at least one of those issues and opened or clicked at least one of them. Received and did neither means they were sent at least one and did nothing with any of them. Not sent to yet is everybody who has received none of them, which is where a subscriber who joined this week sits, and keeping them out of the middle group is the whole reason there are three: somebody who has had nothing from you is not disengaged.

A first open or any click counts, so a reader with images blocked who follows a link is engaged. Open tracking explains why the click half of that test carries more weight than the open half.

Using a segment as a filter

Each segment links straight into your subscriber list filtered the same way, and the count on the page is produced by the same definition the filter uses, so the number and the rows always agree.

That is what makes the segments useful rather than decorative. Received and did neither is the group to write to before you consider removing anybody, and it is the group worth pruning if your list has quietly filled up with addresses that stopped reading years ago. The filter carries into the CSV export as well, so you can take the segment away and work on it in a spreadsheet.

What the figures leave out

Addresses waiting on an import review are excluded from every figure on this page until the import is released. Nothing has been sent to them, they have not agreed to anything, and refusing the import deletes them, so counting them would put a spike in your growth chart attributable to nobody. A held import explains what that review is and how long it takes.

There is also nothing here about your website’s visitors. Issue Mint counts subscribers and what they do with your issues, not page views, so an archive page that brings you readers shows up on this page only as the people it signed up.