SiteSonar: Measuring the Impact of Ads

A while ago I wrote about about web performance and ad blockers. In that blog post I explained that I block ads because I can’t take the performance hit, that running an ad blocker or using Firefox’s tracking protection makes the web responsive again and a real pleasure to use. That blog post lead to discussions with a few people, including Mozilla intern Francesco Polizzi. Francesco and I discussed a study Tracking Protection in Firefox for Privacy and Performance (pdf) and how we could build on that. That study measured a 44% increase in page load performance and a 39% reduction in bandwidth usage across the Alexa top 200 news sites when tracking protection was enabled

So, ad networks, on average, are dramatically burdening page load times and drastically increasing data usage. This makes people sad and makes the Web less competitive with mobile. But publishers depend on ad networks for their livelihoods and surely some ad networks are better than others, right? Francesco and I wanted to quantify this and because Francesco is awesome, he created a browser extension for Firefox and Chrome, called SiteSonar, that identifies and benchmarks performance information about ad-related assets on the web as you browse.

(The extension uses Disconnect‘s list of ad domains to identify ad-related assets and then uses the WebRequest API to determine network response time for individual ad assets. That benchmark, along with information received like file size, status codes, and a timestamp are recorded locally. Finally, every few minutes the anonymous benchmarking stats are sent to a server where they’re parsed and logged into a database. The data powers a dashboard that displays aggregated information about ad network performance that the extension collected. To learn more, check it out on github.)

The early results from project SiteSonar are available at this dashboard. These are preliminary results from just a couple of us running the extension. We’d like to improve the results by broadening the base of people using this extension and submitting data. If you browse without an ad-blocker and would like to share some of your browsing history with us to improve the results of this experiment, you can find and install the Firefox extension at addons.mozilla.org and the Chrome extension at the Chrome web store.

4 thoughts on “SiteSonar: Measuring the Impact of Ads

  1. I know you’re mainly sticking to the performance, load time and data usage angle, but ad networks have become a significant, if not the primary, vector for malware (malvertising).

    Here is a recent & good article about the significance and severeness of malvertising:
    https://www.rjionline.org/stories/breaking-news-1-how-monetizing-became-malvertising
    (Interesting to me is that the audience for that article seems more towards News site purveyors rather than consumers.)

    A friend recently tried an experiment where he went without addons (including ad blockers) for ~6 months. In that time he hit the same “Firefox urgent update” malware serving site 4 times. He was at innocuous sites like walmart.com and was forwarded to the official looking update site through the ads served from it (in this case a Kovter variant).
    One of those times he was testing an older version of the browser (with security vulnerabilities) and became infected. He has since stopped his experiment.

    I, too, am not against sites paying the bills, etc, but I don’t think we should ignore the direct threat currently presented by ad networks.

    1. Great point. There is surely risk involved with exposing yourself to “malvertising.” Just a few days ago, one of my roomates stumbled upon the exact “urgent Firefox update” you mentioned. It is the harsh reality of the digital ad landscape.

      As Asa pointed out, it is hard to imagine a world where blocking gets at the source of the problem. I am hoping that with enough data, this project can provide an objective look at digital advertising for both websites and ad networks. If that can be accomplished, users and content creators alike can make smarter decisions about where they browse and who they advertise with.

      Knowing that problem ads exist and can serve as a threat to users, how do you think the internet can best address their existence?

  2. Asa — Are all the ad networks owned by a company being collapsed and represented under a single entry for the parent? For example, while I noticed that there is an entry for Google, I didn’t see Doubleclick on the list.

Comments are closed.