Ten years ago on our first anniversary, I related the results of an email and postal mail survey in a special issue. (That's right, postal mail: the Web wouldn't arrive for several years yet.) A recent trip back to that issue proved fascinating, and not just for the raw numbers (which revealed that we had a readership of between 5,000 and 20,000 people from 18 countries, as compared to today's numbers of about 60,000 readers hailing from more than 120 top-level domains that correspond roughly to countries). I was most intrigued to see how the comments on what people liked the most about TidBITS have changed little over the years. I suppose that's a testament to how consistent we've remained.
Why Do People Leave TidBITS? Nonetheless, things do change, and despite increases in Web traffic, we haven't seen mailing list subscription increases for some time now. The main reason is simple, but truly annoying - bounces. We run an extremely clean list, and after a certain number of bounces, we remove non-deliverable addresses to avoid wasting bandwidth. Bounces hit us harder than most lists, I suspect, because TidBITS has been around so long. People may have subscribed from one address years ago, set up forwarding when they moved to a different address. Years later, the first address may go defunct, or the user may move again, forgetting about the original forward. (Unfortunately, we have to spend a fair bit of time figuring out how some bounces correspond to subscriptions in our database - and there are a tiny handful we've never puzzled through.)
So, over the last 15 months, we've been asking people why they unsubscribe from TidBITS in order to learn if there are changes we could make that would prevent other people from unsubscribing. We've heard back from roughly 1,800 people; of those, nearly 1,100 don't really count, since they were just changing their addresses on our list or unsubscribing temporarily during a vacation. The people who truly stopped reading TidBITS did so for reasons mostly related to content and presentation.
As far as our content goes, I think we're on safe ground. About 23 percent of the people who unsubscribed did so because they don't use Macs any more, which makes our content less relevant. That's depressing, but not something over which we have much control (and for those switching to Windows, I'd recommend newsletters from Lockergnome, Woody's Watch, and The Naked PC). However, the other numbers tell me our content remains on the right track. Only 0.9 percent feel we need more breadth, 0.7 percent don't find TidBITS sufficiently technical, 2.3 percent find our articles too technical, 3 percent just aren't interested in our articles, and 0.6 percent think we're covering the wrong topics. Combined, about a third of the people who unsubscribe do so because of our content, and the lion's share of those people did so because of a platform change. I can live with that.
On the presentation side, the variables we control are fine. Only 1.4 percent of people said they unsubscribed because TidBITS was too long, and 0.6 percent unsubscribed because TidBITS wasn't timely. For some, email has ceased to be the medium of choice - 13.5 percent of those unsubscribing from our mailing list say they now read TidBITS on our Web site. Although we didn't break out the numbers, our handheld edition has also seduced a few subscribers away from the mailing list.
Here's the killer, though: 43 percent of the people who unsubscribed did so because they were receiving too much email. Ouch. I've certainly signed off mailing lists when the volume of messages became too high, so I can't argue, but TidBITS is only one message per week. Arguably, unsubscribing from TidBITS because of receiving too much email is a content issue, since if these people found TidBITS sufficiently interesting, they wouldn't stop reading.
A total of 11.2 percent of people leave either gave no feedback information or gave some other reason for unsubscribing. When these folks took the time to explain those other reasons, they were generally a combination of the issues above (and receiving too much email was almost always involved).
Addressing Signoffs -- Looking at these numbers, it's hard to see what content changes we could make that would prevent people from leaving. Obviously, if we were less Macintosh-specific, those people who were switching away from the Macintosh platform might stay, but many others might leave (and that's before considering the fact that Windows doesn't interest us particularly). Similarly, a few people complained that our content was less relevant to them because they were still using Macs from many years ago. But as much as we're strong proponents of using old Macs in appropriate ways as long as possible, too much historical content would simply turn off readers who have come to the Macintosh more recently.
On the presentation side, we can't reduce the frequency of TidBITS, but we can shrink the amount of time receiving an issue takes, which may keep people from considering us part of their email overload problem. Our approach here will likely be to create an announcement edition with URLs that point to articles on the Web. That would let people pick and choose what they wanted to read more easily - stay tuned for details.
Of course, the real question is if there's any way we could proactively reduce the number of bounces we get each week. We're noodling over some ideas that, if successful, may help our mailing list subscription numbers.
How Do People Find TidBITS? Shortly after setting up the auto-reply to ask people why they decided to unsubscribe, we did roughly the reverse, and created an auto-reply asking new subscribers how they found out about TidBITS. Here our goal was to try to see if there was something we could do to attract more new subscribers. After eliminating the people who were just changing their address (and five people who somehow had us confused with a non-technical paper publication called "Tidbits"), the results broke down as follows.
Of almost 1,000 legitimate responses, 29 percent heard of TidBITS by word of mouth, 20.9 percent followed a link to our Web site, 20.1 percent subscribed after seeing a mention of TidBITS in another publication, and 11.1 percent either gave no information or fit into some other category (the best was someone who subscribed after finding our Web site because his company makes "Tidbit Caramels"). Another 3.7 percent came to TidBITS after reading one of our articles reprinted in another publication and 2.9 percent found us after reading a book or article one of our editors wrote for another publisher.
Finally, 12.5 percent of subscription responses came from people who said they already read TidBITS on the Web site and wanted to receive it in email. That's mostly interesting because the raw numbers have more people moving from the Web to email (122) than vice versa (95).
Addressing Subscriptions -- The three main ways that people find TidBITS are word of mouth, following links to our Web site, and subscribing after seeing a mention of TidBITS in another publication. Unfortunately, there's no way for us to make any one of these happen short of asking you, our loyal readers, to help out with the first two.
If you know someone you think would find TidBITS useful, please let them know about us. You can either send them to our Web site or just give them the subscription address of <firstname.lastname@example.org>. Similarly, if you regularly forward our articles to someone, please encourage them to subscribe on their own. If you maintain Web pages relevant to the kinds of things we cover in TidBITS, we'd encourage you to link back to articles in our database using the permanent URLs at the bottom of each article. We also have a bunch of reader-specific badges, should you wish to add one to your Web pages. Although this might not seem all that important if your pages don't get much traffic, the more links to our site, the higher TidBITS will rank in some search engine results, including those of Google, our current favorite and primary referrer.
Being mentioned in other Internet publications is also clearly useful. The Kibbles & Bytes newsletter from TidBITS sponsor Small Dog Electronics and the Internet Tourbus newsletter from Bob Rankin and Patrick Crispen were by far the most commonly referenced in feedback messages. Given that their readers apparently like TidBITS, I'd encourage you to check them out as well for Macintosh industry commentary and hardware deals (Kibbles & Bytes) and general Internet explanations and site recommendations (Tourbus).
In the end, we publish TidBITS mostly because we want to provide useful information and analysis to as many people as we can. Anything you can do to encourage others to subscribe to our mailing list, read articles on the Web, or read our handheld edition is greatly appreciated!