Serendipity Isn’t A Use Case

When I started formulating the idea for my current startup, Roamz, the vision for the company was serendipitous discovery of the things happening around you. I imagined that we'd say to people "Want serendipity? Great download our app" and they'd respond "Yeah sure I was just looking for some serendipity".

I quickly realized (the hard way) that you don't leave your house in the morning hoping that you will chance upon serendipity. You have a specific intent in mind: looking for a park for your kids, a cool cafe to grab a coffee or even simply to find a traffic-free route to work. If serendipity happens to occur in other ways then that's a bonus and if not…then it is just another ordinary day.

Having discovered that serendipity on it's own isn't a strong enough motivation to open an app, we've had to figure out how to change our product to account for the way people are actually using it; as a tool to find a place they're looking for. At the same time, we didn't want people to miss out on discovering other cool things nearby that they weren't necessarily searching for.

A lot has been written about this subject since SXSW where many ambient location apps made their debut (see for example http://techcrunch.com/2012/03/30/522411/) . The main discussion focused on apps that help us meet other people nearby with common interests. This class of apps faces a similar problem in that finding interesting people nearby isn't a strong enough use-case in itself to get you to regularly use an app.

The potential value they add is when they send an alert to us that someone nearby is doing something interesting which you otherwise wouldn't have known about. Again, though, it has to be at a time when we are free and willing to act on it. But if the only value these apps add is an occasional notification, are people going to continue to download and use them?

It will be interesting to see how the future pans out and whether people are willing to have a bunch of single use-case serendipity apps sitting on their phone and occasionally coming to life. I believe only a handful of apps will be successful in baking in serendipity as secondary features to support a compelling primary use-case. This is where a technology like Siri could be really valuable; interfacing between multiple apps and platforms to give you interesting information when you want it:

"Jonathan, Foursquare says 3 friends are nearby at Mint Plaza, Waze says the traffic is bad today so you are better off walking downtown and I've seen your calendar is free this afternoon so should I look for a movie that you might like?"

This sounds very Minority Report but the only way we'll be able to truly nail serendipity is by coupling big data with user interaction.

Engineering Serendipity is Tough

Given serendipity is by its very nature random, in its purest form it is very difficult to engineer. Many of the current suite of apps playing in this space use notifications to come alive when they think they add value to the user (and potentially increase the chances of a user having a serendipitous experience) but all too often notifications are sent at the wrong time and add to the noise, instead of cutting through it. This is all part of the learning process for developers, and the more mistakes we make the closer we get to engineering serendipity.

The difficulty of trying to engineer serendipity by analyzing a person's digital trail is that the data exists on many separate services and platforms. Netflix has a good idea about what entertains me, my bank has a good idea of my income, where and when I shop and even my personal circumstances – the constant diaper purchases online give away that I have a baby.

Then there's Facebook which has amazing data about not only me but also the things my friends like, while Foodspotting knows the cuisine that I dig. There are very few services that have the complete picture of who I am. I'd argue that not even Facebook, Apple or Google on their own have enough data.

Even if one service had enough information to paint an accurate picture of who I am and what I'm into, there is no API to read our emotions or state of mind. Hopefully one day we'll have the data, sensors and algorithms that will perfectly be able to synthesize all of this data to engineer serendipity but that future isn't here quite yet.

The answer isn't to give up on trying to add serendipity to peoples' lives but rather help surface things that increase the chances of having a serendipitous experience based on the data we do have. What I've learnt in the process of trying to engineer serendipity is that it is tough to put forward serendipity as a primary use-case.

For an app to be compelling and sticky you need to have a strong primary use-case and bake in serendipity in other ways. For Roamz, this will be through using the iPhone's background location feature coupled with the data we have about what's going on nearby, surfacing interesting local content for the right person at the right time and at the right place.

Brooke A via Roamz