Sam Dean over at Five Thirty Eight pointed out a curious aspect of modern computing, It’s 2015 — You’d Think We’d Have Figured Out How to Measure Web Traffic By Now. It deals with a strange issue in website metrics: the unique visitor. In the old days, people talked about “hits” and then later “total visitors.” But today, the metric that everyone cares about is “unique visitors” — it is seen as more authentic. The problem is that the companies who determine how many unique people visit a website always indicate that there are a lot fewer than the websites’ own access logs indicate. So what’s up with that and should we care?
I must admit that I’ve always focused on unique visitors — even when people were talking about “hits.” The problem with hits was always obvious. If a website has a page with 9 images in it, that represents 10 hits. Thus, a website with a lot of images would have more hits than one that didn’t — even if they had the same number of visitors. But total visitors can be a problem because the same person can come back again and again. Take, for example, a visitor who has left a comment and refreshes it throughout the day to see if anyone has responded. In my case, I’m more worried about my own use. I used Frankly Curious far more than anyone and I don’t want that to skew my statistics.
There are various ways that a website can seem to get more unique visitors than it actually does. The most obvious is the time period. Frankly Curious gets about 1,500 unique visitors per day; but only about 25,000 per month — not the 45,000 the daily number would imply. But the bigger issue is that the whole counting thing is done with cookies. So if you visit the same site from a different browser or computer (or phone), the web server counts you twice. And people delete cookies. And so on.
But I’m not certain that unique visitors really mean that much. A whole lot of visitors come to a page, and back right out a few seconds later. I suppose in an industry that sees advertising revenue as its product, maybe that’s fine: people came onto the page long enough to get an opportunity to click on your Ben & Jerry’s ad — mission accomplished! But for personal websites, people normally think of others actually reading their content as an important aspect of what they do. So if one person comes onto the site, gets involved, and reads thirty articles, that’s a big deal that is clearly not represented with the unique visitors metric.
Just the same, big websites need some way to compare themselves, and unique visitors is as good as any. Personally, I don’t much care about absolute numbers. Even Google Analytics messes up from time to time. What I care about is the trend. Roughly twice as many people visit Frankly Curious today than they did last year. But interestingly, according to my access logs, my unique monthly visitors hasn’t really changed since early 2013. That’s because it is swamped by spam. I wonder if that isn’t a problem at Huffington Post as well? Actually, I don’t care.