How much direct traffic is really direct?

[Dennis-52961]

Does anyone else think that a large chunk of traffic labelled as "Direct" in your analytics isn't direct at all. When you analyse traffic trends it seems that a large percentage could just be browsers with their referring URL hidden so it only appears direct. Here's the evidence:

1. When we've been affected by major search algorithm changes, we've seen big changes in direct traffic as well as organic, but not in referral traffic.

2. If direct traffic is just bookmarks, typed-in URLs, and people clicking through from emails why is direct traffic 85% new visitors? We don't do any offline advertising, so you'd expect genuine direct traffic to be returning visitors -- either our brand loyalists or subscribers to our email newsletters.

3. If you segment direct traffic into new and returning visitors and look at a major algo update as discussed in 1), you find all the drop in direct traffic is from New Direct visitors, with no drop at all in Returning Direct visitors.

Can anyone explain who these New, Direct visitors are if not simply mislabelled new, search visitors. Cookie deletion can't be the problem (ie: they can't be Returning, Direct really) because the traffic doesn't behave like returning, direct (that is, it varies too much).

I'd be really interest to hear theories, and whether anyone has any figures on the extent of HTTP referrer blocking.

[Dennis-52961]

We get a lot of brand searches, but even if they come from unrecognised search engines you'd expect them to be registered as referrals not direct, wouldn't you?

[JohnParker2792]

I have thought the same thing before but, then again it could be true. Do you have high number of searches for your business name? I've also thought that maybe Google can't track it because its coming from another search engine. I also get a pretty high number of not provided keyword searches. Hopefully someone will answer this question for the both of us!