A recent podcast from “This American Life” told the story of a New York literature professor who set a goal to read one shelf of books in her local library. She wanted not just to experience the jewels of the library, but also the average, or even the mediocre books, too. In her quest, she discovered some unknown treasures. This kind of “planned serendipity” stuck in my mind.
Serendipity in CX
As customer experience (CX) experts, Tahzoo tends to focus on patterns: related products, related content, “Customers Who Bought This Item Also Bought” kinds of things. We measure, segment, do predictive analysis, and base our personalization on these steps. Usually, we’re right on target.
But aren’t marketing-driven personalization engines killing diversity? Recommendation engines aren’t really that random; usually they are a list of similar items sorted by some function of probability of you—the customer—liking the item.
In one of my previous projects on accessing news articles, the serendipitous finds were often the most interesting ones. We enriched news articles through entity recognition. Entities came from sources such as DBPedia, Freebase and Geonames. With any article or entity we showed related articles and related entities. General news we tend know, so these relations were easily ignored. The unexpected links were by far more interesting—and possibly more valuable.
The same holds for my personal newsfeeds: Who hasn’t blocked Buzzfeed and similar from their timelines? All in favor of the hidden gems, curated and endorsed by your friends and peers.
In business, serendipity has often played a big part in the success of startups. Muller & Becker wrote Get Lucky on setting the right skills to seize creativity and unplanned events. So, serendipity and business really do belong together.
Just how happy is the union of serendipity and CX? Well, like many marriages, it has been tried with to varying success:
- Stumbleupon tries just that, with tools to manage your serendipity (now that is a real oxymoron). SU started in 2002 and thrived for years but seems to be struggling now.
- Random is an app that tries to serve a combination of relevant serendipitous news. It seems to be going strong, being backed by Skype founder Janus Friis.
- Wikipedia includes a Random article on its homepage. How often do you hit it?
- Dating is all about serendipity, so for instance Happn uses it in its approach.
- The short-lived photo-sharing App Rando was not only labeled serendipitous but even antisocial for going against mainstream orchestrated and “like”-driven channels.
- The Highlight App seems to survive, whereas the Roamz App has already kicked the bucket.
- Spotify offers a serendipitous function: It plays a few seconds of songs that have been played simultaneously by two people in different locations on the planet (audio alert, and this is not a playlist).
- Google worked on a “Serendipity Engine” for years. Is Google Now the end result? Google Now is an App that pushes you information, based on your agenda, location and preferences without you requesting for it.
- For more examples read Sarah Perez’ TechCrunch article on Engineering serendipity.
Making Serendipity Work
From the above examples it looks likes serendipitous functions haven’t been a resounding success. So are there any tricks to make serendipity work for you?
One of the Roamz founders says that you need to have a strong primary product, where you can add serendipity. Serendipity by itself is not a killer feature, but it adds strength to an already strong proposition.
Kiip seems to succeed by capitalizing on “serendipitous moments” by choosing the right time to make an offer, thus conversion ratios can soar. This success is all about choosing the right moment, and by introducing a random factor, a real serendipitous experience is created.
Lastly, to achieve serendipity, you need access to (multiple) data sources that somehow gauge the users’ sentiment. You need to reach them at the perfect, happy moment or otherwise make serendipitous combinations the users cannot make themselves.
Your result list should really be A/B tested— A) with and B) without the serendipitous find. That is the only way to really know that ‘serendipity’ is a better experience.
Add a touch of serendipity to your CX, so your users feel pleasantly surprised by your offers, rather than feeling they are being funneled. Let your UX team be creative, but let the data team make it relevant and productive.