You do read your site server logs, right? Not sure about your numbers. Ad blocking is about 20-25% world wide. It could be you are in a geo location where that number is higher, if your traffic is geo based. On sites where I have adsense (not many these days) between site logs, analytics I see 15-20% blocking. These are visitors which (I presume) have it turned on as default, rather than site by site.

Not sure I'd call this stealth traffic. You do have traffic, you just don't have adsense impressions.

How are you weeding out automated traffic?

How are you weeding out automated traffic?

What kind? If bots, then that's handled separately in .htaccess ... and has been for years.

Yes, bots. If you don't filter those, it's only natural that traffic will be up and bounce rates down, so I was wondering if eek2121 had thought of that. (Actually I'm not sure if page requests from, say, Googlebot would increase or decrease bounce rate. Do they maintain sessions? Probably not. I guess it would depend on how you're tracking visits.)

(Planting flag for 1000th post.)

Bot traffic is filtered out just like with normal Google Analytics. I am merely proxying the GA requests through my web server and using their API to log pageviews and events. I have double and triple checked the Analytics data. It is on the up-and-up.

As far as web server logs, they are monitored for security, but I don't run analytics software for stat tracking.

Finally, my niche is a technical nich where the core audience is on a desktop running chrome or Firefox.

I see the same ratio of visitors with adblock enabled as eek2121. My sites are tech informational with many visitors coming from all over the world. I used a simple script, based on BlockAdBlock script, and modified it to report back to my server. It also reports whether javascript runs or not. So far I ran it on two sites. One site is very small, around 1K PV per day. Adblock numbers for it range from 45% to 62% of the traffic that has javascript enabled. The second site is larger, and the numbers are more stable, 50% - 52% in the course of last 10 days. Here are some numbers. On May 6th the site had 78K pageviews according to server log files. My adblock checker reported 66K PVs with javascript enabled. 50% or 33K of those had also adblocker running, and the other 33K were supposed to show ads. On the same day, Google AdSense reported 43,595 pageviews, that had the Active View Visible number of 63.68%. It's curious that 43.5K * 0.6368, which equals to essentially 27.7K pageviews, and it's not far off from the 33K number. By the way, we block many bots, including many portions of AWS, XLHost, Digital Ocean, OVH and many others IP ranges.

I just implemented a beta solution to track ad/analytics blocking that I can't talk about yet cause I'm under an NDA; right now most of my event sites seem to be seeing 8-10% ad and/or analytics blocking software. That's actually not as bad as I was afraid it would be. My audience for those sites is probably not all that technical, and most of the traffic is mobile iPhone users. Theoretically the bots and crawlers are filtered out. I'm just going into my peak season, so it will be interesting to watch as the traffic increases. But at the moment, it doesn't seem to me to be anything I need to worry about.

I've been using AdReclaim to show ads to ad blockers. It's been making a couple bucks a day, which tells me that people with ad blockers on DO click on ads that they like.

azlinda: Of course they will. They are used to not seeing ads. When confronted with ads they don't have the 'ad blindness' that many people have. Netmeg: is this a big G solution? Or third party? Or is that part of your NDA as well?

Or is that part of your NDA as well?

Yep.

Bot traffic is filtered out just like with normal Google Analytics.

With "normal Google Analytics," Javascript is the primary gatekeeper. Most bots don't have JS enabled, so their traffic is not counted. How does that work in your server-side solution?

I've done something very similar- I've tested using the Google Analytics Measurement Protocol to record traffic server-side, as well as using the on-page tracker.

I also use custom bot detection which tags crawlers etc. with a custom dimension, which is then available as a filtered view (for bots or no bots).

There are also auto-generated AWStats reports, and another analytics system (Clicky) on some sites.

I've found that the javascript-based trackers are very close (with GA you have to be more careful to block your own visits and filter referrer spam, neither of which is done automatically). GA is blocked very slightly more than Clicky.

For server-side, filtering bots is a challenge. There are many with innocuous-looking user-agents that don't get caught by matching strings.

As far as those blocking tracking, I would need to revisit the numbers, but from memory I don't believe it will exceed 10%.

I'm not running ads,so can't comment on adblocking vs. blocking of tracking.

Given the chameleon nature of UAs, bots, and brokered traffic, and "how the web currently works" getting actual numbers is becoming more difficult. The whack-a-mole approach to bots can be a full time job which contains potential damage greater than versus ad blocking visitors.