Friday, May 1, 2020

The Numbers Game

I'll try and keep this short because if blogging about blogging can seem a little self-indulgent then blogging about one's own blogging stats verges on the self-obsessed. In my defense, I can only say that I didn't start it.

There was a bit of a discussion in the Blaugust/Blapril Discord yesterday about viewing figures. SDWeasel followed it up with a list of his top five most-viewed posts. It made me realize I hadn't looked at my own long-term stats in years.

I used to pay very close attention to them. Blogger provided a handy graph on the front page and I would watch the peaks and troughs, enjoying the way the average always trended slowly but steadily upwards as Inventory Full grew and developed what seemed to be a consistent audience.

That lasted for several years. Then everything went haywire. The blog started receiving huge spikes of attention, page hits jumping by an order of magnitude. The line on the graph veered crazily skyward, peaking at many times the previous zenith.

It would have been great if those were genuine readers but of course they weren't. Something had changed. Bots had either found my blog or found a way through Blogger's blocks. I spent a while investigating and I did discover where most of them were coming from but there didn't seem to be much I could do about it.

Old school Blogger. For now you can toggle between this and the new style. Each has its merits. And why do I only have seventeen followers? Up from sixteen, where it stood for many years. Once again, figures elsewhere in Blogger tell me I have a couple of hundred.
The bot invasion had an unexpectedly benign outcome; I stopped paying attention to my stats. They didn't mean much any more so why bother? Instead of caring whether my numbers were trending in the right direction I began to focus on the responses I was getting.

Comments have always been the most important form of feedback. There's plenty of bot action there, too, but fortunately Blogger's spam filter is very effective so few see daylight. The ones that appear represent genuine readers and that confirms there's someone out there reading this stuff, at least.

I stopped watching the big picture but I still look at my dailies with some interest. I have no way of knowing how many of the page hits are real readers but when I have issues with bot traffic it almost always goes to much older posts, as do automated spam comments. No clue why. You'd have to ask a botmaster.

As with Wilhelm's monthly analysis of SuperData's reports, it's not necessarily the intrinsic accuracy of the data that matters so much as what it shows by comparison. If the data is consistent in its shortcomings it still provides useful information on trends.

After I'd read SDWeasel's post I took a look at the lifetime stats for the blog for the first time in what must be at least three years. At first I thought they were hopelessly borked, as I'd expected they would be. At the very top, with way more views than anything else, were two posts that I couldn't imagine ever having mattered much to anybody.


According to Blogger, my most-viewed post in the entire time I've been at this game is the New Year's Eve post from 2017. A decent comment thread shows it did get some traction and someone even left a comment as late as September 2018 but I very much doubt it was seen by just under thirty thousand real, live human beings. More like three hundred.

Second on the all time list, with a mere seventeen and a half thousand views, is my first-ever post on Project:Gorgon from December 2013. At the time it got a couple of hundred hits, which was very good back then and I've linked to it in many of the posts I've done about the game since, which probably accounts for a few more.

With all of that, I would guess the genuine views for that post over six and a half years might come to five hundred or so. Then, on February 1 2018, more than 17,000 curious P:G fans just happened to drop by to see what I'd thought of the game five years earlier. Yeah, right.

As was mentioned in the Discord discussion yesterday, having a post linked in other media can bring a huge spike in interest. Posting about popular games also helps. I thought at first that might have accounted for the third entry on the all-time list but it's another fake. It may be a post about World of Warcraft but it's just those bots again.There's a huge spike in early 2018, a month which seems to have been a bot hot spot, which accounts for almost the entirety of the five thousand views.

The fourth entry, though, that's different. It's a post I wrote about Black Desert when that game was flavor of the month. I also mentioned the word "boat" in the title. Everyone and their dog was trying to build boats in BDO at the time and I got a lot of traffic just from google searches.


The graph for that one shows what the numbers for a post that purports to be a guide to a common problem in a popular game can look like. You can see how BDO players were still wondering how to build a boat for several years and some of them were misguided enough to come to me for advice.

That post also tells you in a nutshell why my page views are never quite what they could be. That post sounds from the title as though it might have some practical advice (it doesn't) and it mentions the name of the game, too. That's a rare combination. I don't often write guides and even when I do I tend not to let anyone know I've done it.

Following some advice I read when I started blogging, I used always to include the name of the game I was writing about after the title. Now I hardly ever do, partly because I stopped caring about artificially boosting page views a long time ago but mostly because I've come to care more and  more about how the blog looks and adding the name of a game to the title is aesthetically displeasing. These days I pride myself on being as obscure as possible in my titles, which exist primarily to feed my increasingly frequent musical collations rather than to give anyone a clue what the post might be about.

At this point the eagle-eyed will have spotted some very strong evidence as to why Blogger stats are not to be taken literally. In the full list above, the boat post has 4.9k views. In the detailed pull-out it has 7.62k. All of the results vary like that. I don't know why.

As well as the internal Blogger stats not agreeing with themselves I get a monthly report from Google Analytics which doesn't agree with either of them. The GA figures are always much, much lower. Often by a full order of magnitude.

I think the analytics are probably quite representative of what you might call the real audience for the blog. In April, for example, it seems to show a fairly consistent pattern of fifty or sixty unique visitors per day, which sounds realistic. Blogger routinely shows many times that, usually a couple of thousand. I really have no way of knowing which is closer to the truth but sanity leads me towards the lower figure.

The fifth post on the all-time list is also an all-time puzzle. It's my first post about Neverwinter Online, published back when the game was new. As the graph shows it was consistently in demand all through a long period when, as far as I know, blog-botting wasn't an issue here.

I never understood why people were interested in that post. I tried to work it out but I came up blank. Maybe it's nothing to do with the game. The title is vaguely suggestive. Perhaps it's that.

That's my top five. I won't go into the rest of the list in detail but I will say that I can account for the numbers on at least half a dozen of the top twenty by their having been linked on reddit or Massively or the official forums of the game in question. That always brings in the crowds.

Looking at this stuff is endlessly fascinating but it doesn't mean much. Self-evidently the figures are horribly compromised. And it's not as if they're even coming from different sources - all of them come directly from Google!

In the end I work by feel more than these supposed facts. Some posts get traction, some don't. How many comments they get, how many other bloggers mention them or link to them or spin posts off the ideas they contain, that tells me far more about how "successful" posts are than anything Google sends me.

And the truth is, the audience that ultimately matters is myself. Am I pleased with what I've made? Does it give me pleasure when I read it back? Am I satisfied that my time making it has been well spent?

Smug, self-satisfied git that I am, I can almost always answer "yes" to all those questions. Anything more is sugar.

3 comments:

  1. I have my annual anniversary stats wallow where I go over top posts and all of that, mostly because I enjoy numbers in that way. Since Google pretty much dictates which posts are popular and which are not, I can usually figure out the rhyme and reason. But they are never the posts that I feel strongly about. Fluffy short posts seem much more popular than my walls of text.

    I have said in the past that if I wanted traffic I would write a lot more about WoW, but in looking at my stats, I think Pokemon would be the way to go. My fluff Pokemon posts seemed to do very well over the years.

    ReplyDelete
    Replies
    1. Or pictures of your cats. I bet those score well.

      I definitely think short posts, preferably in bullet point form or lists, do significantly better than enormous walls of text. Speaking of which, I have such a wall lined up for tomorrow...

      Delete
  2. fortunately Blogger's spam filter is very effective

    I sure wish! Just the other day I had to manually bin about 50 identical comments from someone called "Lord Mark" offering to turn people into vampires if they email him...

    ReplyDelete

Wider Two Column Modification courtesy of The Blogger Guide