Web sites still suck - why business models hurt user experience
Update time: last week, I did an enterprisey riff on a piece by The Verge Editor in Chief Nilay Patel, The Mobile Web Sucks. In a laudably open style, Patel made a sensible argument (the mobile web experience isn't so great) but with a simplistic conclusion (it's Apple Safari's fault).
My piece attracted the ire of an ad tech professional, who felt I had made a "blatant assumption" that ads equal slow loading web sites. Here's what I should have said: most media business models result in a poor browsing experience. Ads and affiliate plug-ins lead the way, but full screen pop-ups and auto-play videos take their toll.
That said, the most absurd web sites I visit every month are not behemoths but local news sites (have a gander at FOX Atlanta or CBS Dallas - two random picks - if you want a super fun browsing experience).
Supporting evidence comes from an excellent piece by developer Les Orchard, The Verge's web sucks (Orchard is a Verge fan who is not enamored with their browsing experience). Orchard also took issue with Patel fingering Safari, so he did his own research on what causes The Verge to load like a cement truck.
Orchard's punch line:
Did you know that The Verge delivers you to around 20 companies for advertising & tracking purposes? I didn't. That might foul up your web experience a little bit. Maybe we should try something different.
To get there, Orchard used Mozilla's devtools to determine that a sample Verge page (of a typical 1,600 word article) loaded 7MB of Javascript (Orchard says this is a lot, I'll take his word for it). But what was loading?
Orchard used a nifty Firefox add-on called Lightbeam that provides a visualization of "who's tracking you online." Orchard used the tool on The Verge, generating a visual look at all the third parties the site loads with each page. The results showed 47 sites, which reduced down to 22 after eliminating subdomains, here in list form:
Now, Orchard calls this "spyware," which may be a bit unfair given that when we visit web sites with Javascript on, we are essentially giving tacit permission for the likes of Google and Amazon to track us. But whether or not it's spyware, there's a TON of marketing tech loading with each Verge page, leading Orchard to drop this beauty:
I feel like someone just set up the entire vendor hall from an awful tech conference in my living room.
Then he raised a similar question to my last post:
If you want to talk about browser performance - mobile or desktop - don't you think that maybe, just maybe, some of that 7 megabytes of JavaScript code might have an impact?
Down in the comments section, a brave reader with a vested interest (a NextWeb reporter) played devil's advocate, arguing that careless web programming leads to browsing bottlenecks.
That's a factor, but not all media sites are created equal. Orchard did a run down of the number of "requests" and loads for some big media sites, and saw plenty of variation. Here's a few (see his blog for the full list):
(I took some delight in seeing Forbes.com so high on the requests list, given their pop-up obsessions). Scrolling through Orchard's reader comments, I found only one or two "I will NEVER pay for content under any circumstances" sentiments. Most readers said they would pay for content that was simple to consume and not bogged down in obtrusive ads.
Orchard brought up several news sites he thinks strike a better balance, including Flipboard, Facebook's Instant Articles, and Apple's News app.
The NextWeb commenter assumes that advertising is a given, and we must learn to put up with it, but readers rejected that. I'll make a different argument for our enterprisey readers: poor reader engagement is a failure of imagination.
There are plenty of approaches that can serve readers at scale and further commercial agendas. A well-argued piece from Techdirt Editor Mike Masnick takes up that point (Don't Blame Your Community: Ad Blocking Is Not Killing Any Sites). Masnick shares their decision to try an obtrusive form of advertising (that was lucrative), then pulling it down when readers complained. He makes the case for "advertising that works, not annoys":
As you may recall, UPS sponsored me doing a series of "whiteboard videos" about topics that we regularly talk about here -- one on the economics of abundance, one on the innovator's dilemma and one on the difference between innovation and invention. Before releasing these, I was actually a bit nervous about how people would respond. But these videos, which were clearly labeled as being sponsored by UPS, actually were a huge hit, and we received lots of compliments about them. Even more interesting? Numerous comments on the videos thanked UPS for sponsoring them and making them happen.
Funded video is not the only way forward. Pando is taking a different approach, putting most of their content behind a subscription paywall in the self-described push to become more independent of Silicon Valley investors. The intent to avoid over-dependence on advertising dollars is clearly stated. I'm not convinced that a strict paywall approach will work for Panda, but the point is - experiments that buck the existing model are welcome, and readers may surprise you when they perceive value.
I won't rehash the enterprisey lessons I spelled out last time; I'll simply add that great things happen when you put "delighting users" at the top of the priority pyramid, test the heck out of what you publish, and let the logistical priorities stem from there (e.g. fast page loads trump programmatic ad servers).
It beats the heck out of placing bets on Safari.
Image credit: Boredom, Yawning, Nerd. © BillionPhotos.com