When Software Jumps the Shark

Posted By on Oct 6, 2011

Are you one of the million hipsters clogging my stream with Tweets about Arrested Development’s reunion for a half series and a following feature movie? If so, did you ever think that the right time for Arrested Development to end might have been when it did? This blog post was triggered by a discussion among my friends about how TV producers never have the sense to realize when a story has been told, and wrap it up at an ideal time. Most producers get greedy when their series are successful, and keep them going long beyond their optimal expiration date, a phenomenon known as Jumping the Shark, after Fonzie’s trip to LA in Happy Days.

When a TV show finds success; money, fame, and inertia make it so that as long as networks are willing to renew the contracts, producers are willing to go along. You can’t blame them for not saying no. They don’t want to disband the crew, they have sentimental attachments to the characters, and it’s really hard to say ‘No’ to a boatload of cash. So artistic legend be damned, they will keep making new shows well beyond the point when the story runs out of momentum.  While Happy Days is the iconic version of this, you can look at many of the currently running series to see this pattern – How I Met Your Mother, Desperate Housewives, The Office, and plenty more. And even HBO, where artistic license often trumps popular opinion, sometimes lets series go overboard with sequels. Sex and The City ended at the right place after 6 seasons. But now there are movies. The first movie wasn’t not that bad but please stop making them! When they got to the second movie, it was so bad that I fell asleep.

You must be wondering ‘what about the shows that ended too early instead of too late?’ The truth is there’s no such TV shows that ended too early if they are ‘successful’. They got canceled because the network didn’t consider them successful enough. Audiences’ preference or the fact that you or I think it’s a great show doesn’t translate to millions of eyeballs.

Does the same phenomenon happen with software?

The day after we talked about letting a TV show end, I was working on designing a few new features for Boomerang for Gmail. As the product designer and manager, I spend a lot of time prioritizing which feature requests we will build and which we won’t. Then I saw a tweet from someone evaluating the product. He drew the conclusion that we have added too many features so that we could justify the paid subscriptions. Of course, I got mad. How dare he call MY PRODUCT feature laden?!

Does he not realize this product started out as 2 buttons inside Gmail? It literally was 2 buttons integrated into Gmail, with nothing else – not even a way to see, cancel or manage the messages you scheduled. Every new feature you now see in the product, including the new features I was designing at that moment, have been requested by many, many users, over and over again. All of us are heavy users of our own product, but every step of the way we’ve made sure we’re building the product for our customers, not just for ourselves.

As I was fuming, I started to see how the same problem that bothered me about TV shows could manifest itself in software. The guy on Twitter wasn’t there when we coded up our management page the first weekend, when we realized it was a hole that we must fill or our customers would be mad. He didn’t see the hundreds of emails requesting response tracking. He’s the guy who’s starting to watch How I Met Your Mother or The Office for the first time this season. He doesn’t know how it got there, or  why each feature was added, but he doesn’t care. The product, like new seasons of a TV show, has to be able to stand up the way it is now. It doesn’t matter that each individual step made sense at the time. The product’s history is our problem, not his.


When Facebook came out in 2004, back when I was in college, all “The Facebook” included was a personal profile where you can upload a single photo and ability for users to add each other as Friends. That’s it. No Wall, no multiple photos, and not even private messaging. Aside from the ubiquitous shade of blue, it’s hard to even tell that the modern version of Facebook evolved from the original.

But it was successful. And as it continued to grow, the team decided to add more features, like the Wall and the News Feed. Likewise, people kept asking for new things: the ability to chat, ability to add videos to your Wall, and dozens of other requests.

These features have dramatically changed Facebook’s experience from what it was in 2004. But the numbers of their user growth (700 Million users as of 2011) show that the new changes have worked. Facebook, as it is today, adds value to new users who have no idea about its history and how it evolved. Maybe Facebook hasn’t reached the point where the product story has been fully told. But as they add more and more features, reacting to user requests, and more recently, responding to the competition from Google Plus and Twitter, that day may be quite near.

Microsoft Office spent the last two release cycles without adding very many new features at all. Instead, the development team’s focus was on exposing the features that were already there, since most of them were only being used by less than 5% of the total user base. They’re in a tough position – even a feature that’s used by only 1% of Office users is still used by millions and millions of people. The result is a product that does so many things that it took over five years of development for Microsoft to figure out a way to combine all of those features in a coherent way.

So how do you know when your product has told its whole story, and it’s time to work on something else?

If your product isn’t successful, it’s easy. Like a network show that nobody watches, the product is almost certain to get canceled early. You won’t have any users, and you won’t have anyone urgently demanding for features they must have. You won’t have any obligations to fulfill these demands either.

If your product is successful, it’s much harder. Feature requests will continue to roll in forever. If you implement them all, your product will jump the shark. I don’t believe we are there for Boomerang quite yet. I believe it will be important for us to know when the product will be complete in advance, and to have the strength to stick to the plan. After all, the most important part of a great story is a great ending.

Thanks to Alex Moore for his help with editing this post.

Submit a Comment

Your email address will not be published. Required fields are marked *