Testing ad blockers and their behvaiour • Adssets •

Document description

Show how the Ad Blockers behave in the presence of our ads when they are not coming through an ad server.

General context

AdBlockers have existed on desktops for many years but their actual reach, after a 40% increase from last year, is still around 25% on desktops in countries like Sweden according to recent reports.
With iOS9, Apple has included the ability to block content through the installation of Ad Blocking apps. But despite their initial popularity they seem to have dropped in the Top Apps-lists in App Store, giving them less visibility.

Test environment and context

The tests have been done on a Windows 8 in Chrome since it’s in Chrome and Firefox where most of the Ad Blocker plugins can be found and where they are more extensively used. It’s fair to consider mobile AdBlockers as a subset of desktop in the terms of techniques and blocking rules.
Current status of ad blocking per browser

Tested AdBlockers

AdBlock
ScriptSafe
Disconnect
Donottrackme
Ghostery

Results

The test involved loading our demo page containing just the ad and opening a publisher webpage to show the effect on the ads being shown there.

	Publisher Page	Demo Page
AdBlock	Blocked all ads in the publisher pager.	Ad not blocked. Impression counter blocked.
ScriptSafe	The whole site is blocked, as provider of unwanted content. Blocked all ads in the publisher pager when the whole page is trusted.	Block the ad. Detected apparently due to the presence of external impression counter.
Disconnet	25-30 connections detected and blocked.	1 connection blocked. Impression counter detected and blocked.
Donttrackme	25-30 connections detected and blocked.	1 connection blocked. Impression counter detected and blocked.
Ghostery	25-30 connections detected and blocked.	1 connection blocked. Impression counter detected and blocked.

From the results, it seems like our ads are not blocked by any major Ad Blockers, but by the presence of markers such as “impression counters” increases the chance to be noticed as ads and therefore be blocked.
Given that 25% of Swedish connections have some sort of Ad Blocking this would explain a notable deviation between the numbers generated with the impression counters and our statistics. Since users with Ad Blockers might be blocking impression counters.

What can be done?

The response to AdBlockers is more a commercial decision than a technical challenge.
The presence of AdBlockers can be indirectly detected in many cases and can for example prevent users with Ad Blockers to gain access to full content or the ads can be part of the news like in the case of native ads and avoid being blocked.
In both cases, AdBlockers can adapt its techniques to get access to full content (until the publisher develops a countermeasure) or detect the Ads embedded in the content to prevent their display.