Measuring Readership

A common dilemma for writers of all kinds, how do I know how many readers I have? If you run a site that gets thousands of unique visitors a day this question might not be as serious than if you run a site where you get less than 100 uniques.

The Fundamental Problem

First you have basic techniques to determine readership: number of people coming to the site -vs- number of people commenting or otherwise active on the site. However not every visitor to your site is going to read your work. Some may click the back button within seconds. Obviously this junk traffic shouldn’t be counted as part of your readership. Yet counting readership by activity is troublesome too because the vast majority of internet users (Wikipedia estimates 90%) prefer to lurk, a small portion will maybe comment once or twice (9%) and an even smaller portion will be active (1%). (Within the business this is sometimes referred to as Wikipedia’s 90-9-1 rule, it’s not an exact science obviously … but it is, in general, pretty accurate) Counting readers by activity is basically only counting 1% of your probably readership.

Beyond the Basics

Thankfully, technology exists to save us. Here are some slightly more accurate ways to measure readership and their vulnerabilities

• Measure by RSS subscription/blog watchers …. some flaws to this in that not everyone who comes to your site regularly is going to be interested in subscribing. Maybe they don’t understand or use RSS, or they don’t have a blogger (or whatever) account. Plus just because they click “subscribe” doesn’t not mean they ever really come back after that.
• Measure by sales/downloads … only relevant if you offer sales in the first place One could also argue that just because someone buys your book, does not mean they READ it … but I think most of us would make off with our $2 sale and not give a damn about that ^O^
• Measure by analytics … complicated and involves some careful analysis, but using a sophisticated analytics tool like Google Analytics you can separate your quality traffic out from your junk traffic and look at your update days to average how many readers you have.
• Measure by membership … This is the way we do it, can’t get access to the new/fresh content until you’re a member. Has some of the flaws of RSS subscriptions, tends to underestimate true readership but suits our purposes.

Personally I find the membership method to be most accurate … but still, not every website can implement it. So let’s look at the next best option: Analytics. Many of you have Google Analytics, many of you probably only use Google Analytics to a fraction of its true power. Today we’re going to play with some advanced functions to get a better idea of true readership.

Step 1: Open up your Analytics account and click on the Advanced Segment link. Select to create a new segment if it does not bring up that option immediately.

Step 2: Who do we consider a “reader” really? Obviously it’s someone who comes to the site regularly and takes the time to actually read the content. So Advanced Segments allows us to instruct Google Analytics to filter out all traffic that doesn’t fit a certain profile and then redo all it’s pretty calculations based on the specific segment of traffic remaining.

So here’s how I set up the segment for Split-Self:

I want it to count ONLY people who have visited Split-Self’s section AND have been to the site before AND spent more than two minutes on the site. Now, depending on the structure of your site you may need to set this up differently. For example if you’re running a WordPress blog as your principal method of distribution and only have one serial, you obviously do not need to restrict the segment by “Page”

Name it, save it and this is what you get back:

WOW! 97 readers!!! That’s way more than the 12 listed on fluffy-seme!!!— er except this is in no way accurate. Google Analytics, as much as I adore it, is not very good with displaying information on regular visitors. There are special sections where you can look at visitor loyalty and such, but that data doesn’t carry over to other portions of the site. In other words if you have 10 readers who come back every week, once a week to read your serial, Google will tell you have “40 visits” because … well you did! 10 people x 4 weeks …. If you want to distill down to how many people come back once a week, Google can DO IT but has trouble displaying it.

Step 3: So we’ll distill these numbers ourselves. Note that if you have a serial that updates more than once a week this can get a bit complicated and estimates therefore a little rougher. But Split-Self only updates once a week, so all we have to do is click on the date range and tell Google to restrict the data for just one week.

After we’ve done that we have a much better number: 19 readers. And as we look through the segment week by week this number stays more or less consistant, 19 … 19 …. 19 …. A month ago it’s 15-16~ on average, so we also see how readership has grown thanks to whatever cultivation activities we’ve used.

Interesting to note … this pretty much fits with projections based on 90-9-1

Lurkers: 17.1 people

Semi-Active Readers: 1.71 people

Active Readers: .19 people (AmHarte you are officially .19ths of a person ^o^)

Readership -vs- Draw

No doubt you’ve noticed that our segment only looks at RETURNING visitors. It does not account for people who’ve just found your content recently, fallen in love with it and haven’t had the opportunity to come back yet. Shouldn’t we also count those people as readers?

I like to keep these numbers separate to look at how fluffy-seme retains and converts readership. Sometimes you read a few chapters of a serial, like … LOVE IT even, but just never get back to it. I don’t really consider those people “readers” per se, but it’s important to look at how the setup of your site is helping people come back to it or losing people.

So I’ve setup a new segment called “Split-Self Draw”. It looks like this:

Since these people are new, they don’t have access to the member’s only section so we count only the portions of the site open to them in the first place: the archive where they presumably start reading at Chapter 1. We only want to look at people who are actually READING so I up the time filtering to 5 mins and of course we specify that these people must be new visitors to be counted.

Comparing Serials

Now the real fun begins, since fluffy-seme hosts multiple serials we can actually compare draws and see which serials attract new potential readers best. We just go back to the Dashboard and click on the Advanced Segments tab above our traffic chart. You’ll see something like this:

So I want to compare Split-Self to Martin Ostrowski’s Season in the Red, which one are new visitors more interested in? Again I restrict this to one week, not so much because of double counting but because the lines are going to be pretty close together and the longer your range the harder it becomes to read the output in graph form.

Split-Self draws more people, but snarky little Martin Ostrowski gets them more active in the site (10.41 pages viewed per visit compared with 3.61). Here’s another week:

Obviously there are lots of options to play with here. You can look at how spikes in the Draw effect patterns of Readership (in other words, do new readers become regular readers?) you can even segment out traffic from out particular source and look how those visitors behave on your site (I’ll be doing that soon … as a follow up to my early thoughts on MCM’s 3D1D project).

All and all Google Analytics continues to be the most mind blowing awesome thing ever

Bookmark It

Hide Sites