The amount of unidentified traffic in your website analytics reports is higher than it needs to be. Almost everyone’s is. In Google Analytics, for example, we’re talking about the visits which are labelled as “(direct)/none” in the source report: the traffic which we have no idea where it came from.
Now, you may think that it doesn’t matter, as 75% of the traffic in your reports does have an identifiable source, and a 75% sample is pretty good. But it’s important to remember that the 25% which is labelled as “(direct)/none” isn’t the same stuff as the rest, but just unidentified. It’s something completely different.
Children at school are taught that if they go out and do an hour-long survey of the colour of cars passing by, it’s fair to assume that this is a fair sample and the results would be the same if they did the survey for 24 hours or a week.
We can’t do the same here. The 25% of the traffic in our reports which we don’t know about is not an extrapolation of the 75% which we can identify. It’s other – different – stuff.
By default, the identifiable traffic in your website analytics report is primarily from other websites and possibly mobile apps. The rest, in that “(direct)/none” category, is usually considered to be primarily from:
- Typed-in URLs; and
The traffic from these can be quite significant. As I’ve shown before, you can use Google’s Campaign URL Builder to identify traffic from your third-party advertising; you can also use this tagged traffic for your own emails, and you can set up domains which redirect to tagged links to use in print adverts and other sources. All these will cut down the amount of “(direct)/none” traffic, but more importantly, will show you what results other promotional initiatives are actually producing.
The main point of this article, however, is to highlight how in recent years, other sources have been adding to the unidentifiable traffic. Examples of potential issues include:
- Traffic from an HTTPS site if yours is an HTTP site;
- Traffic from many mobile apps;
- Traffic from messaging or chat services;
- Traffic from untagged links in PDF or other documents;
- Traffic from shortened URLs; and
- Traffic from installed software
Some marketing managers just consider that they’re not going to be measuring the website traffic generated from, say, a print advert, and that’s their decision. But if nothing else, they know that the maximum it generated was the traffic shown as “(direct)/none”. In reality, with so much other traffic gathered up under that header, it will be just a tiny fraction of that. That’s why identifying as much as you can through tagging is so important.
Further reading: Direct Traffic Is Dark Traffic, And That’s OK (Seer Interactive)