Blogging for Writers: Looking for Love in All the Wrong Bounce Rates


Photography: tc7

Just about any blogger eventually discovers the alluring world of site traffic tracking, whether it’s through Google Analytics, Site Meter, or lovely fresh Mint.

We wonder: what impact are we making out there? Who reads us?

Who likes us?

Now that’s one complicated question, and not always so easily answered. There are no little “like” graphs on your WordPress.com stats dashboard. We can only infer—outside of honest confirmation from every single one of your visitors—whether people like your site or not.

At first, you might wonder if mere visit count is a direct-enough indicator of love. But these days, social media sites like Digg and StumbleUpon can send you hundreds, even thousands, of hits—and they can all evaporate over a matter of days. So visit count isn’t sufficient enough of an answer.

Another measurement, frequently proposed as an answer to a blogger’s most troubling question, is what’s known as bounce rate.

Bounce rate is the percentage of people who visit one page of your site, and then disappear. The lower your bounce rate, the more people are sticking around to read more of your stuff during just a single visit. The higher your bounce rate, the less people like you—or so the wisdom runs.

The popularity of bounce rate as the answer stems in large part from the ease of its measurement, and the cleanness of its oracular nature: high bad, low good.

So is bounce rate the final answer? Many people act as though it is.

I’m here to tell you that it isn’t.

The Two Most Important Questions

Question #1: Who’s Visiting You?

When playing pin-the-tail-on-the-causes-of-increased-bounce-rate, who visits you is an important predictor of love. So let’s start out by factoring out the traffic whose bounce rate you have the least amount of control over.

One important source previously mentioned is traffic from social media sites. Your site gets Dugg, or Stumbled, or Slashdotted if we’re talking old school—just on one page, usually. Extremely high traffic results, with an extremely high bounce rate; we’re talking upper 90s or so. And like the crash of a wave on the shore, it’s soon gone (although some sites, like StumbleUpon, are structured such that a trickle of traffic continues to come through).

Search traffic, on the other hand, is more a constant river of people googling and hitting some page or other of your site. Sometimes it’s raging white water (like Halloween LOLcats on All Hallow’s Eve), but most times it’s a lazy brook (students searching for critical analysis of the Speckled Band). In any case, the bounce rate here is also high.

We’re not going to cover search bot traffic, for spiders do not care over whom they lavish their indexing ways.

Now we get to the more normal traffic: direct/referred visitors.1 You’re more likely to draw out a low bounce rate from this portion of your audience.

Of course, that’s no guarantee.

But if your bounce rate is high, does it mean that this most stable part of your audience loves you not? For that, we need to consider the next question.

Question #2: What’s Your Purpose?

Or rather: what’s the purpose of your site? Purpose, more than anything else, determines bounce rate for the “normal” part of your audience—and whether a high or low bounce rate is actually meaningful.

So let’s examine some site types and look at their resulting bounce rate patterns—once we’ve factored out social media and search traffic.

Reference Sites

Simply put: a site that acts as a reference; wikis, for instance. And the most popular reference site is, of course, Wikipedia; but reference sites also include All Recipes, Godchecker.com, and DVD Verdict.

Should a reference site have a low bounce rate? That depends on the nature of the information and its presentation. Most good reference sites tend to have a lot of internally related links on every page, such that a curious reader could begin a long walk through multiple pages, wandering far from their original point of inquiry. This is a commonly known problem, and lowers bounce rates phenomenally.

News Sites

People like news. And news sites, from Boing Boing to CNN.com, always feature new stuff, interesting and relevant to their target audience, frequently updated multiple times of the day, every day.

These kinds of sites don’t have to be limited to what we may think of as regular old news, either: io9 and Lifehacker also count as “new shinies every hour” sites.

If you have a front page with a ton of intriguing headlines updated every hour, you’re going to see two behaviors: people visiting, finding nothing interesting, and going away (for a little while); and people who click through. The more stuff that’s posted and cycled through, the more likely there’ll be something interesting for everyone.

A lower bounce rate is a good indication that people are interested in your news items—but at the same time, you don’t always cry over a high bounce rate, because you know it’s likely that former bouncers will return and click through on another visit.

In other words, a high bounce rate is not as much of a concern. In fact, the bounce rate could always be 80%—it’s just that in any one interval, a different 20% of the audience found something to click through.

Community Sites

Anywhere folks gather in large numbers to converse upon certain topics. These of course include forums like Whitechapel and the Authority Blogger Forum, but also blogs that have managed to build up a community, including Making Light and Whatever.

Because discussion between multiple users is a highly important part of these sites, rather than just off-hand comments, the bounce rate of a healthy community site will also be low, as people surf through threads and write posts.

You could say that the low bounce rate comes at a high cost, however: the same human foibles that drive people to lower your bounce rate can also drive flamewars, so you’re either policing—or letting flamewars happen. Owners of successful community sites are far more worried about tending to the community—the low bounce rate is a given and, all things considered, not hugely important.

Serial Stories

With the onset of the web, suddenly the business of getting work out in front of people you don’t know became much, much easier—and not just occasionally, when you’ve got enough money to print up some copies, but you could do it every day, even multiple times per day, for a fraction of the cost of print.

And so with the web was reborn the serial work. We most often see the serial art in the form of webcomics, such as Sluggy Freelance, Girl Genius, and FreakAngels.

People are enthralled by good stories, so you see two types of bounce rates. The first are the newbies to your story—the ones who dig back into your archives to find out where the hell all this goodness is coming from. They’re the ones with the low bounce rate.

The second are the veterans, who’ve been following your story for some time now, and thus are less likely to go digging into the archives. They’re the ones with the high bounce rates—and yet they are obviously loyal and this is obviously good traffic.

You welcome both kinds of bounce rates. Which means that bounce rate doesn’t really matter to you as a metric of love.

Journals

Not all blogs are journals, and not all journals are blogs. In fact, blog is a rather looser term than most people think it is, since a blog can be just about anything these days—it can be a community, it can be a serial, it can be a reference, it can be a wide smattering of news.

Very often these days a blog is a journal.

There are two definitions of journal, however: either a diary (most writers keep this kind, like Neil Gaiman and Elizabeth Bear), or a very focused magazine (e.g., a periodical on Sherlock Holmes like The Baker Street Blog, or one on the unholy marriage of writing and blogging). On the web, both tend to be a list of articles ordered by date, added to from time to time.

This tends to mean:

  • Articles have short memories; things often get lost in archives, and there often isn’t enough backlinking to keep them active.
  • Motivation for reading back articles is usually much lower than motivation for reading current articles.
  • Articles tend to be dated.
  • Updates aren’t frequent enough to yield news-sites-insanity-level stature.

All of these add up to: high bounce rate. Sometimes nearly 100%.

However, this does not mean FAIL. Neil Gaiman’s journal likely has an obscenely high bounce rate at this point, and that’s not a failure by any meaning of the word. People simply visit everyday.

This also does not mean instant success either. It just means that bounce rate tells you little.

Bounce Rate is Not Always the Answer

Bounce rate yields some information, but not a whole lot of it. In fact, depending on the purpose of your site or blog, bounce rate may mean nothing to you as a health indicator.

However, there is one metric that always indicates a healthy site with lots of love waiting in the silent wings of the web.

A Truthier Answer: Returning Visitors

Returning visitors is a different measurement from bounce rate. Bounce rate doesn’t cover an absence of hours or days; repeat does. Returning visitors, on the other hand, will catch visitors who come once a day to read the newest thing that got pulled up on your RSS feed, whereas bounce rate would, if you were being rather strict about the interpretation of the percentages, declare them to be of fleeting affection at best.

For every type of site listed above, repeat visitors is a definite, positive indicator of health—no wangling about with percentages and use cases here.

So why aren’t returning visitors the prime metric over bounce rate when it comes to affairs of the blogosphere’s heart?

The first strike against returning visitors is that it’s much more difficult to track and calculate. It’s one thing for a traffic tracker to remember if someone visited a page twice in a row in the last 7 minutes; it’s quite another for it to remember if someone has visited 24 hours before. Many site traffic trackers don’t even have a way to present returning visitors as a metric to you.

The second strike against returning visitors is that it’s not a metric that can turn up results in a few minutes; it may take days or at least hours to settle down to an exact number. And as we’re all rather impatient with the near-instantenous gratitude the web can provide, this is annoying to some.

And the final strike against returning visitors is that the data is simply hard to collect and collate in the first place. Someone who’s visiting the next day with the same IP may be an entirely different person—and person who visited previously is coming from a different IP. You’d almost certainly have to set cookies on their browsers to identify them properly; and that means, with browsers that block cookies and the use of different browsers, the data that results has to be conservative—far more than bounce rate.

Who Measures Returning Visitors?

Google Analytics, that’s who. As this is Google, they have the data mining power to bring you quite a few statistics—though they don’t provide live updates (and returning visitors isn’t a live update metric anyways).

Under the Visitors tab on the left, you’ll see “New vs. Returning” as well as “Visitor Trending” (here you can see bounce rate, among other things) and “Visitor Loyalty”. The latter section fascinates me a bit, since “Loyalty” will show you the percentage of people who visit once, twice, etc. times; and “Recency” will show you the frequency that people visit you.

Those are the numbers I work to improve.

‘Cause bounce ain’t got nothin’ to do with love.

For More Information

If you’re new to Google Analytics, I suggest you check out Ben Barden’s totally awesome 4-part series, “A beginner’s guide to Google Analytics”.

1 I’m not going to distinguish between direct (might be visiting your site under their own whim) and referred (visiting your site due to clicking a link from somewhere else). This does run against traditional wisdom; however, I do so mostly because some referral traffic may be thought of as direct traffic (for instance, clicks from online feed readers like Google Reader or Bloglines) and some direct traffic is actually referral traffic (not everybody’s browser lets your software snoop at their previously visited site).
Advertisements

15 thoughts on “Blogging for Writers: Looking for Love in All the Wrong Bounce Rates

  1. O Hi Thar,

    I’m impressed by this whole presentation. For starters, I wasn’t so familiar with all the various genres of sites out there, one being “Girl Genius.” I will subscribe, and see how deep this rabbit hole goes. Cheers!

  2. Thanks, Guiconner! Now I have your full name in my comments in all its glory… :)

    Bamboo, thanks! I’m always glad to introduce more people to the somewhat weird sites I love, and definitely to Girl Genius, which I think is brilliant.

    Hmmm. I’ve sort of fallen off the Sluggy Freelance wagon (not sure why) where webcomics are concerned. But I’ve been with Girl Genius since it started.

  3. I would like to have a Community Site, but I’m still a long ways from that yet.

    I need to check into Google Analytics, I haven’t done that yet.

  4. One of the best explanations I’ve seen (or at least one that even I can understand). I’m glad I found you on the AB forum…I can tell you’re going to be very helpful. Thanks for posting such important information in a clear, concise way.

  5. Trisha, Google Analytics is cool and free. Ben Barden’s got a super-cool 4-part tutorial for beginners on his site; I’ve updated the article to point to his tutorial. (Should have done that earlier!).

    Anita, you’re welcome. :) The Blogging for Writers series is aimed at regular writers, fiction or non-fiction, who may not be all that familiar with the ins and outs of some of the aspects of blogging. Probably it should be named “Blogging for Beginners”, but I still aim certain aspects towards writers (most notably in sites I select for examples, for instance) although even that is not so strong.

  6. This post is excellent! :D

    One thing that really rings true for me is that stats are not much use when used in isolation from other figures. When used with a combination of other measures, bounce rate can be more useful.

    I particularly enjoyed reading your analysis of how bounce rate varies depending on the type of site you run. Things like people visiting everyday, or the bounce rate always being 80% but the 20% who don’t bounce aren’t always the same people, these are very, very good points that few people seem to be aware of.

    The next time someone moans about Entrecard generating a high bounce rate, I’ll try and remember to link to your post.

    And thanks for the link. :)

  7. Hey Ben! Good to see you here! And thanks. :) Your tutorials are, as I said, awesome, especially since Google Analytics can be a bit overwhelming to use at first.

    This article was in fact inspired by the bemoaning of bounce rates in various forums and blogs, including the EntreCard world but also outside of it. You have to consider multiple metrics as well as what those metrics mean to you. Metrics without meaning is just statistics without context—and we know how the latter goes.

    It’s also harder and thus less attractive. I come from a world where combining metrics and context to get a good picture is very important, so I know how difficult it can be. Spreadsheets FTW, man.

  8. I look at the numbers from Alexa and Technorati but they are pretty meaningless. What counts for me are the comments and those who come back again and again and find something to say. That’s how I know I’m being heard/ read. I also keep a hit counter on my blog cause I’m old fashioned and that used to be the only way of measuring whether or not anyone was out there but yourself.

  9. Hi Laura! I also treasure my commenters. They’re more involved with the site.

    But for myself, I know that there’s a lot of quiet people on the web. And that there’s nothing wrong with that. If they have something to say, that’s cool. If they just want to read, that’s cool, too. I like them both coming back. It’s more difficult to track, but I guess I’m just a crowd-pleaser.

    Before the advent of Google Analytics and friends, I used to use server-side perl scripts to compile the statistics for various websites I tended to. They needed no cookies or javascript, because this was just the webserver logging when someone (naturally) requested a page or image or whatever. AWStats was my favorite.

    Alexa tells me not that much right now because I don’t have enough traffic for it. Technorati has been a little more informative, because there are a few blogs linking back to me now, or mentioning me, or whatever. Whether or not people should worry about either is up to them; I don’t really worry much, but I know when my rankings go up it’ll be a good sign—and at that point they’ll be more informative.

  10. I have never cared about the bounce rate maybe because I really didn’t understand it, but since reading your post I now know I was right to not let it bother me even though I heard so many people fussing about it. I just appreciated the fact that I was getting traffic that was real and they even leave a comment every now and then. Thanks for making it simple to understand…Jude

  11. I find that the people who moan about a high bounce rate are the people who understand it the least. Blogs inherently have a high bounce rate. People pop in to read the latest post, then leave. That induces a bounce, which in this case isn’t a bad thing.

Comments are closed.