I recently came across a fascinating piece by Paramal Satyal . He is Nepalese although these days he seems to live in France. His piece is largely autobiographical, but what caught my eye was a bit in the middle where he looks at one of the most ordinary web pages he can find. The Le Monde Experiment He took a look at a very unexceptional web page on the Le Monde website: Astronomie : la sonde Juno s’est mise en orbite autour de Jupiter . Yes, it's in French, but you don't need to know any French to read this post (unless you decide you want to read the piece itself, which is about the Juno space probe). This is an article of 1,500 words, with three images and a video embedded. It should load instantly on a modern connection and a modern computer (or phone). But: the page is 3.1 MB in size, makes about 460 HTTP requests of which 430 are third-party requests (outside of Le Monde), and takes 20 seconds to fully load on a fast 3G connection (from Paris, France). It also stores 100 cookies and contacts 118 third-parties Ignoring the cookies, and contacting tracker sites, that means that only about 6% of the data delivered is the content that you wanted to look at. Most of the remaining 94% is to do with advertising (some is to do with optimizing content delivery) where you are literally the product—these companies make money by selling information about you. There's a video in there though, perhaps that is what makes the volumes big. As it happens, Paramal looked at that, too. That single video adds over 100 requests, 60 of which are to an additional 15 third parties, and adds 16 third-party (not Le Monde) cookies. That's before you decide to play the video. But wait, there's more. The page continues to transfer data and make web requests even after it has fully loaded, as you scroll around (and probably if you don't). In just a few minutes, over 30MB of data will be transferred between your browser and the over 100 third parties. This will continue as long as you leave your browser open on the page. No wonder our data caps seem to get used up so fast. Just for comparison with the tests below, I re-ran Webbkoll on the Jupiter probe page just now. It is even worse than when he did the analysis. 100 cookies is now up to 142, 118 third parties is now up to 131. You Are the Product Even if you don't play poker, you've probably heard the adage that " If you have been in a poker game for a while, and you still don’t know who the patsy is, you’re the patsy." In the same way, a TV sitcom (on regular network TV, not some special Netflix special) is a way of delivering an audience to advertisers. You are the patsy, although in this case it is hardly a secret. In the same way, Google makes its money not by delivering search that is so valuable that you pay lots for it, but by delivering you to advertisers. In their case, search makes you even more valuable since, presumably, if you just searched for "leather sofas" then you are probably interested in buying one. However, increasingly, many websites are in the business of delivering you to advertisers. A couple of years ago when I was at Semiwiki, I interviewed Peter Theunis, the CTO of Methodics. He'd been something senior at Yahoo working directly for Marissa Meyer, and joined Methodics (based in San Francisco) since he got fed up of the three hours on buses to and from the valley. But also because he got disillusioned that he was making a difference: The purpose of every major site on the internet, such as Yahoo, is to get you to click on ads. That is the metric. How many ads got clicked on. A lot more tracking goes on than you might realize. You see those little Facebook buttons so you can "like" a piece? They call home and so Facebook knows you've been there even if you never click on them. Almost everyone runs Google Analytics so Google knows everywhere you have been. I wrote the first half of this post as an implied criticism of Le Monde, which you think of as a newspaper but appears to be, as Paramal says, " an invasive, insecure advertising platform with good content, in that order". Cadence.com Next, I thought I should see how we stack up. Cadence does not run its website to make money from the website. We run it to provide information to our customers and potential customers. I ran Webbkoll on Cadence.com to find out: So just going to the home page of our website installs 66 cookies and contacts third parties.This actually is a similar count to many other web sites for high-tech companies. I'm not involved directly with the web team but I would assume that when you visit our website that we try and notice if you've been before, so we can measure engagement. That requires cookies. Of course, if you are registered on our website, presumably we know who you are, and you probably don't want to have to type your name and password every time, which is also done using cookies. We use a couple of analytics sites to see how many people visit and how long they stay - those require cookies. So I looked at the list of third-party sites contacted and checked with our web team. Cadence's Web Team As it happens, the head of Cadence's web team sits on the other side of one of the walls of my office. The web marketing team was as surprised as I was at some of these results, in particular, the number of 3rd-party requests and cookies, given the low number of first-party cookies (a first party cookie is one placed directly by the Cadence website, a 3rd-party cookie is one placed by someone else, such as Google Analytics). Some of the cookies come from normal business needs such as web tracking using Google and Adobe Analytics, Google AdWords, email marketing (we use Marketo), and so on. One tool we use allows website visitors to share our web pages and we are now investigating whether this sharing tool is where these 3rd party requests are coming from. Obviously, we are not trying to monetize our traffic with ads, sell user data, or anything like that. We don't want anyone else doing this on our website either. If we find out anything really interesting, then I'll write another post about it. Try It Yourself You don't need any technical knowledge to try this yourself. Go to Webbkoll and type the name of any website (actually, any URL you want, it doesn't have to be top level). Bruce Schneier Time to try one more. Let's see if we can catch out security guru Bruce Schneier, who has his own website and blog that I visit regularly. He practices what he preaches. More to the point, he shows that you can have a website with a lot of content without having every move tracked by the advertising networks. Sign up for Sunday Brunch, the weekly Breakfast Bytes email.
↧