Hacker News Reader

Back in May, TheNextWeb made some changes. Their mobile experience at the time had two bars, one at the top and one at the bottom. Even on a large phone display, viewing the content between the two was very distracting. Also, they both had icons for twitter and Facebook. I gave them some feedback regarding the design and said that I felt it was a UX mistake. The VP of design got back to me and said, very politely, that he disagreed with my assessment. He went on to say basically that an increase in shares means an increase in page views which means an increase in ad impressions, and thus income. With TNW being an ad-supported company, this directly relates to the quantity and quality of our content. This lead me to feel that the single driving motivation was not that their users had a good experience, but that they maximized revenue. From a user point of view, I feel that's wrong. From a business point of view, it does make sense.

Regarding who is responsible for the final content, I would assume that if you have a VP of Design, that's the person. Maybe I'm wrong.

I notice now that they only have the one bar at the top. Interesting.

They're possibly optimizing for the wrong metric, or at least a metric you don't care about. I'd think they'd want repeat viewers. A bad UX will mean fewer repeat viewers, and possibly an unsustainable business. Classic short-term vs long-term view. Problem with this is it's generally harder to measure long term, so short term becomes easier to justify.

Maybe yes, maybe no. A classic example is tourist trap restaurants in heavy tourism areas. They literally don't care about repeat customers because all of their business is transient one-time customers. Sure they might get bad reviews ultimately, and that could hurt them, but that hasn't really seemed to slow them down, and that is easy to game.

The ad exchanges don't care for repeat viewers. Its purely a numbers game. More clicks more revenue.

The content sites themselves largely don't care about repeat visitors. Social networks and other aggregators are the primary drives of traffic to news/content/lifestyle sites. In other words, most people do go look at the home page of Buzz Feed or Wired or Huff Po and read multiple stories. They go to Facebook and click into different articles on different sites

ITYMTW "most people do not go look at..."

Optimizing for specific metrics is dangerous, and makes a big difference

This is typically the role of a project's Product Manager: to balance the needs (or perceived needs) of all internal stakeholders (development, design, marketing, sales, legal, and all the various other isolated silos that don't talk to each other) with the interests of the end user. You need someone empowered to say no to crap, because every company is filled with people trying to load crap into the products, even developers.

Weak/ineffective Product Manager roles are set up to simply report to one department, for instance Marketing, and just do everything they want. Effective ones have the power to say things like, "We're not doing this awful UI element just because legal says they want us to." and "We're going to stop taking requests from sales until we improve product quality." You need one person who is directly responsible for the entirety of the product, who can't escape responsibility with the standard "not my department".

The problem for these Product Managers is that the incentives between the different stakeholders are drastically misaligned at a fundamental level.

The people in charge of community, design, content and publishing are actively fighting the demands of the parts of the organization that drive profits.

Companies that use advertisements to subsidize free content have sales and marketing teams that need to interrupt the goals of all of these other functions in order to keep the organization alive.

However, if you look at the incentive structure for say a traditional book publisher, the marketing and sales don't actively fight the content because they don't put adds inside the covers. They just sell the finished product.

Sure, market appeal might affect the kind of content that gets invested in by a publisher, but it doesn't affect how the content is presented.

That makes me wonder - why do companies treat design and content as something that does not drive profit? Profit is holistic, otherwise why bother with the so-called "cost centers"? Kind of like a restaurant saying "the cash register is where our revenue clearly ends up, let's just cut out all the other cruft like the kitchen".

It's harder to point to specific items in those areas and say, "this feature X saved us $Y", whereas the sales departments have easy figures to point to.

Sadly, this whole "cost center" versus "profit center" exists at all types of companies. I used to work in food manufacturing, and the incentives in play there were beyond strange. The sanitation budget was constantly being slashed, because Quality Assurance as a department doesn't bring in a profit (that they can see), whereas hordes of "process engineers" would come in and rearrange boxes and claim that they have just eliminated 10 steps out of the workflow and calculate this to be saving $92/day. They quickly get promoted before anyone can actually see if this made orders any faster. Another common tactic was to cut positions on the manufacturing line, claim the reduced headcount as a savings, and move on before the problems of fewer headcount (slower throughput) became apparent.

It's a matter of accounting.

With a professional kitchen, yes, the cash register is a place where value is stored, literally, but what are the inputs? Settled bills coming from customers. Who were paying for what? Food coming from the kitchen. The kitchen is creating valuable things that can be tracked and accounted for by both the customer and the cash register and the bill and the receipt from the debit card and all tracked in a double-entry style in to books, organized by many different kinds of departments.

This way the kitchen itself can be viewed as an isolated financial entity, separate from the concerns of the cash register or the landlord. We can then see exactly how profitable the kitchen is because it has well defined inputs and outputs. With this information in hand, we would probably not want to get rid of the kitchen.

Imagine if the food were free and covered with ads. That's what we have now.

Content, when treated as property on a market, like food sold to customers, retains these properties. Now the "content department" is most definitely creating valuable things for the salesmen to hock.

Design is tough because it is qualitative and it sort of falls apart from an accounting perspective. It's really hard to see it as anything other than an expense, which is probably why it is so undervalued in most corporate settings. It certainly is important but only when it is novel as it can be easily copied. You can't copyright an entire design paradigm, so the minute someone invents "Art Deco", well, the cat's out of the bag.

The iPhone is interesting because it had such a big head start but don't for a minute think that there aren't fleets of accountants with a lot of say running around Apple and helping them figure out how to bridge the economic realities of engineering and supply chain with the delightful new features being dreamed up for the products.

Most long-lived corporations that make money back for most investors end up in a situation where they design and ship lots and lots of products precisely because they can't ever quantitatively figure out how to ship a "hit record". So if you turn the product of all your subjective creations (design, art, content, creative) in to intellectual property that participates in a marketplace, you're creating a way to run a business that you might be able to make some sense out of.

But a company that relies on ad revenue that sees no change in revenue when the experience is slow or "bad" wouldn't hire a Project Manager to put UX first. Why would they? It wouldn't fall in line with their company objectives (read: make more money)

It seems like there are two separate kinds of cruft: stuff that is distracting and gets in the way (social buttons, privacy policy modal, big nav-bar with temperature, ads, etc.), and stuff that happens in the background and doesn't get in the way, but makes things slow (multiple ad network embeds, optimizely, analytics scripts, etc.). While I agree that the first category has a lot to do with organization structure and incentives, the second category seems like it may be a good business opportunity for a new generation of "low-cruft" solutions to the same problems. For instance, I was kind of astounded that optimizely only has a crufty javascript solution, rather than a (probably far more complex) less crufty pure server-side solution, and I spent awhile looking around for such a thing, but found nothing much. Or something that puts all the tracking and analytics that different departments want to do behind a single optimized script (maybe segment.io already does this?).

tl;dr; The problem that lots of different people that contribute to a website are vying for the user's attention in different ways doesn't seem to have a technical solution, but the problem that we're making too many third-party requests that respond with too much data and/or run too much javascript seems like one that could be solved if tools focused on it.

I would say the biggest problem with pages like this is that it "jumps" around as content is loaded. The fact that ads are lazy loaded is not bad (probably good actually!).

This is easily fixed by simply having the container set to exact width/height so that it doesn't mess with the rest of the page when it loads.

CNN also has a sidebar that pops in with other stories. Don't do that. Render it on the server.

Just a couple of fixes like this and the page would be fine, even if some extra stuff takes a while to load. They key from the user perspective is that the article should immediately be readable when the page loads.

The fact that ads are lazy-loaded can be a big performance problem if it's implemented with a method that causes browsers to turn off async pan/zoom. Especially sad is when images are loaded this way, because most browsers will load images on scroll automatically, so this technique of "lazy loading" ads for "performance" actually cripples scrolling performance with no benefit.

> if it's implemented with a method that causes browsers to turn off async pan/zoom.

What technique would cause that?

Segment(io) doesn't have a solution for this, in case you decide to go searching. And their server side APIs have plenty of issues of their own (eg no way to get utm parameters to the destination - hugely frustrating and we are still working with them on that).

Thanks for the info. When I first heard about their product, I got excited that they were trying to solve this issue, but after reading their marketing materials and documentation decided that it didn't work how I expected, but I wasn't sure.

... they are largely built by committee

Any organization that designs a system (defined broadly) will produce a design whose structure is a copy of the organization's communication structure.

-Mel Conway ( http://www.melconway.com/Home/Conways_Law.html )

This might also explain why mobile apps by the same company serving the same content and, presumably, contributing to the same business model somehow manage to have less cruft[1]: there's a single gatekeeper team for each app.

It's funny how the important difference between apps and web pages is entirely arbitrary, social and historical and not based on the underlying technology—the concrete difference between the two.

[1] Or, at least, that's been my experience. As a concrete example, take a look at the BuzzFeed app[2] which only has the prominent social buttons as visual cruft.

Of course, the rest of the cruft on the Buzzfeed mobile page were two (!) links to their app...

The CNN app[3] is probably a better example, with the app being almost completely cruft-free.

[2]: https://play.google.com/store/apps/details?id=com.buzzfeed.a...

[3]: https://play.google.com/store/apps/details?id=com.cnn.mobile...

Pessimistic me says that's mostly because the market has not matured. Ads and cruft will arrive eventually.

that is exactly my point. And even when you have a single professional webdev/design team, you have no power over legal restrictions (EU cookie widget) or contractual obligations boss signed (ads must appear next morning). Real world just goes forward. Users have more power with adblockers over their UX than UX designer.

One way you can avoid the need for an EU cookie widget is by not setting unneccessary cookies.

In many cases to determine its necessity you set a cookie. Seriously, can you name a mainstream news site in Britain that does not set a cookie?

No, but I can name lots that shouldn't.

If I'm not logged in, there's no good reason for a news site to set a cookie.

Most uses of cookies that are not tracking are exempt from the cookie law and don't need consent: http://ec.europa.eu/ipg/basics/legal/cookies/index_en.htm#se...

Conway's Law writ as Web Cruft.

I always find it amusing that our capitalist system basically consists of competition between command economies as succinctly described above.

> And that hero image that's actually 1600px wide and then resized in CSS to 400px? It's there because that was the source image from the external agency and no one thought to modify it.

And because of retina iPads and Macbooks.

True, but they could be doing better by serving the appropriate image size depending on the client.

Much easier said than done, sadly.

Not that hard. Want to make it a background image?

    @media only screen and (-webkit-min-device-pixel-ratio: 2), only screen and (min-resolution: 192dpi), only screen and (min-resolution: 2dppx) {
        background-image: url(image@2x.png);
    }

Do you consider the image to be part of your content? (Something you'd want to appear on printouts and RSS feeds)

    <img src="image-src.png" srcset="image-1x.png 1x, image-2x.png 2x" />

Retina images are a solved problem for modern browsers. Very few users with Retina screens (and thus new computers) are on non-modern browsers, and those users who are using old computers and browsers will end up getting the 1x image. (Good luck finding a Retina Macbook Pro that runs IE9)

Sorry, I just saw this now and I doubt you will see it 6 days later, but just in case:

The main problem with that approach is (or at least used to be; I would love to learn that it's changed) that the browser will still load the 1x image first. Since the large chunk of retina users are on mobile devices, that sucks.