In a direct response to Vice and PCMag’s article on Avast Antivirus, Avast’s Corporate Board and CEO Ondrej Vlcek ceased operation of their data subsidiary Jumpshot. It seems that core mission of Avast cannot be uncorrupted by the sale of their user’s data.
“I realize the recent news about Jumpshot has hurt the feelings of many of you, and rightfully raised a number of questions – including the fundamental question of trust.
https://blog.avast.com/a-message-from-ceo-ondrej-vlcek
As CEO of Avast, I feel personally responsible and I would like to apologize to all concerned.”
And once you have that it’s a simple database query to ID users.
Jumpshot came to my attention this August of 2019 shortly after Rand Fishkin released a click-thru study for Google Search Results. I think it is fair to say that it raised many eyebrows in the SEO community. Not just for the primary claim of the study, approximately 50% of searches didn’t resolve in a click, but also for the source of the data.
So where was Jumpshot data coming from?
Several sources as it would turn out. Avast and AVG anti-virus are the two most popular products, CCleaner, PiriForm, are some of their lesser known products. Avast also offers: Virtual Private Network (VPN), Password Manager, Browser Security Extensions, and has even expanded into Mac, iOS, and Android.
From PCMag who partnered with Vice on this article:
Device ID: abc123x Date: 2019/12/01 Hour Minute Second: 12:03:05 Domain: Amazon.com Product: Apple iPad Pro 10.5 – 2017 Model – 256GB, Rose Gold Behavior: Add to Cart
At first glance, the click looks harmless. You can’t pin it to an exact user. That is, unless you’re Amazon.com, which could easily figure out which Amazon user bought an iPad Pro at 12:03:05 on Dec. 1, 2019. Suddenly, device ID: 123abcx is a known user. And whatever else Jumpshot has on 123abcx’s activity—from other e-commerce purchases to Google searches—is no longer anonymous.
https://www.pcmag.com/news/the-cost-of-avasts-free-antivirus-companies-can-spy-on-your-clicks
Even if you own a website that kind of user behavior is difficult to measure, let alone find. Google Search Console gives you an idea of what your search impressions are like and you can determine a rough click-thru rate from that, but it is not a complete picture by any stretch. If you are running Google Analytics that isn’t going to help much either. This is concerning. Even if it is in the Terms of Service, it isn’t defensible. The CEO of Avast and Board see the conflict that it creates within their company and the potential to corrupt their values.
Fishkin has been sounding an alarm about Featured Snippets affecting the click-thru rate and one can only assume that the purchase of this data was related to that concern.
My academic background is Political Science, and despite what people think about that, it is actually rigorous study and requires some basic statistical analysis. If you have ever seen the picture of Harry S. Truman holding “Dewey Beats Truman” Newspaper Headline and know the story behind, it you know that you can’t get over a bad statistical sample; and it doesn’t matter how large that sample gets.
If your data source isn’t representative of the actual population, it isn’t going to yield useful data about the larger population. The polling data that showed Dewey beating Truman was from households who had installed a telephone and were more likely to be Republican. Polls also include other data to show you that the sample is accurate. Are you getting numbers that are proportionate to reality? Or at least other data? For Jumpshot it’s pretty simple, you could use other user data may include browser, OS, and that can be compared to see if the sample is representative. Google Analytics will show you at least that much data so I would assume Jumpshot could show you lots more.
The data I have collected form myself and years of consulting for clients don’t back up this zero-click claim. And even if I did believe the study was from an authoritative source, and found the conclusions compelling for my clients it wouldn’t mean I would stop doing SEO and pivot to influencer marketing or something equally crazy. Google is still going to be the largest source of traffic for my clients, and for me frankly.
Science is about method, experimentation, and scrutiny from your peers. It fails on those grounds. I’d like to state loudly and publicly that this “study” doesn’t stand up to any kind of academic scrutiny and shouldn’t be taken seriously aside from the privacy concerns raised. We should strive for better as a community. If we are going to claim to be scientific at all our mindset should be more critical and skeptical.
I doubt Rand Fishkin is going to apologize. While some people realized their family’s entire web history was for sale by the anti-virus program and VPN they installed, Fishkin was enjoying Pasta Condescendo with a nice Chianti without a moment of self-reflection.
It is not clear at the time of this publication what will happen to the data at Jumpshot or if we will get a list of who else has that data. Avast has been breached before, and this treasure trove of data is a massive target for hackers.
What is clear is that Rand has doubled-down on Jumpshot.
