People as Attributes (or how Uber got it right)

In Part II of our series on people search today, online identity and privacy (read Part I ‘Searching for the Self’ Here) we examine how Uber disrupted the taxi industry by recognising that many industries that have traditionally been ‘searches for people’ are actually searches for a very limited set of specific attributes. 

Let’s start with a thought experiment. Suppose you were building an app to find tennis partners, how would you do it? The simplest way would probably be to have people input their location, availability and estimated ability on a score of 1-10. You would probably also want the app to verify that this was a real person, so you might insist on a social log-in but basically what you are interested in is the above three characteristics.

The app would then use a simple interface to make matching with other potential tennis partners extremely easy and would, in an ideal world, make it possible to book a court/set a time without ever having to leave the app. Hey presto, you have a tennis partner.

Essentially what has happened here is that person you are being matched with is being reduced to three key attributes: Ability at tennis, location, availability. Who they are beyond that is unimportant, for the most efficient service to enable the largest number of competitive tennis matches – this is all that matters.

Much of the way that people search is carried out in the 21st century is through this kind of ‘attribute search’. Uber are the most successful example of this. Despite their initial tag line of ‘everyone’s private driver’, Uber have actually done something rather different. Uber have successfully commodified the attributes of their drivers to the point where the personality of the driver has almost completely been erased, as exemplified by the relatively recent feature allowing the passenger to play their own music by connecting their Spotify. The driver is there to provide a journey; he or she is not there to be a person.

A CBS Insights Report

Unlike with our tennis player example, the ‘ability at tennis’ score is the product of a service-side verification by Uber. We do not really care that our driver is a 10 not a 7, what matters is that he/she is above whatever the threshold is that defines a ‘safe driver’. Our tennis player app is essentially ‘Uber for tennis players’, a tagline (Uber for X) that you increasingly see clogging up the pages of techcrunch and on the slides at demo days.

The Uber model of identifying attributes and commodifying them as efficiently as possible (by which I mean reducing the individual who possesses them to that limited set of attributes) is one that has been revolutionary for 21st century capitalism. Many industries that used to require relationships with an individual have no intrinsic reason for doing so, other than a relationship was required in order to get that service. For instance getting a plumber to fix a sink does not require any relationship with the plumber, provided you know their ‘ability at plumbing’ attribute is above the threshold for competency. A more efficient system would be an Uber for plumbers, meaning that you don’t have to wait for ‘your’ regular plumber to be available and that demand is more effectively spread around the ‘competent plumber’ work force. There are other useful market forces at work here that help make the market more reliable – i.e. more full of cherries than lemons. Clearly an ‘Uber for Plumbers’ leads to a desirable outcome for both consumers and competent plumbers, but it does mean that people no longer have a relationship with the person who comes to fix their sink (whether or not this is a desirable outcome is of course up to you).

However there are many industries that cannot be ‘Uberfied’. Hiring a full time employee is someone you see everyday, you can’t just bring in a rotating series of temporary workers summoned at a moment’s notice. There are skills that have to be learned, personal relationships that have to be established. Above all there is no ‘minimum competency threshold’ that applies for all cases. What makes someone a good employee in one team may make them a terrible one in another, people are different and work together differently.

Let us return to the tennis player app though experiment. Yes it’s efficient, but is it the most desirable system? Does it produce the maximum utility for its users? People don’t play tennis in silence, they arrive at the court a bit early and hang around waiting for their time slot, they take breaks between sets – standing in awkward silence during these moments would be far from desirable. Suddenly, wouldn’t the app be improved if you could see the interests of potential partners? Wouldn’t the tennis game be more enjoyable if you had a shared topic to talk about between sets?

This guy is probably too good to be on the app

Taking it further – what if the person was from your industry? Someone you might potentially bump into in meetings, or might be good for giving you user feedback down the road? Perhaps someone who knew similar people to you and so you could interact with both on the level of tennis but also turn it into something with potential business benefits, or whom you could interact with on a social level by gossiping about mutual acquaintances, or on the intellectual level by talking about your shared love of art or literature.

Someone who lives in the same location as you, is free at the same time and has a similar tennis ability, while also sharing at least one of the above attributes would be the ‘optimal’ tennis partner. At Wholi, this is the kind of relationship we want to help create in the long-term – by expanding the attributes to describe the person behind them rather than severely limiting the attributes to create ‘an Uber for X’. We want to create ‘people search’ not simply ‘attribute search’.

Life is short, why waste time playing tennis with people with whom you have nothing in common. Uber revolutionised how we think about hailing a taxi, and the lesson we can learn from this is that they were able to revolutionise people search taking the ‘personality’ out. With richer contextual information it will, long term, be possible to not only make people search incredibly efficient, but actually able to foster more meaningful connections than those that currently exist.

Recruiting for permanent roles is clearly one industry that cannot be ‘uberfied’, simply because (despite the best efforts of Meyers-Briggs devotees) what creates a good employee is not simply something where being above an ‘attribute threshold’ is what matters. Any time personality play a big role in a relationship, that relationship is ‘uberification-proof’. Hiring, finding an investor, a mentor, new friends, people to date, people to travel with, people to play sports etc. with are all examples of huge industries with a number of existing solutions – industries which cannot be fixed by tapping a button and immediately having someone who is ‘good enough’ sent to you. We believe that all of these industries can be improved through richer contextual people search, where the attributes of someone’s personality play a determining role in putting great people in the room together. Thinking about people as attributes has allowed Uber to shape the way we think about the service industry, but recognising its limitations is vital to building a world where the best people are more easily able to find and contact each other.

One thought on “People as Attributes (or how Uber got it right)

  1. Very interesting point you raise here about differentiating between the binary and minimum threshold services (the Uber model) and more nuanced/complex services (connecting people via search). I think personality is a key component, as you state, and would extend that concept to say that the match or interplay of personalities who connect is the goal of people search.

    I’ve have the good fortune to tackle some of the challenges within the people search field as well, having led a project to build a social network for Duke University alumni and students in which the profile and search were at the core of the product.

    In our research and discovery with constituents we turned the Reid Hoffman paradigm of Information Age vs. Networked Age ( on it’s head to create an alternative proposal, namely that current people search tools on the market represent (or guide) an Information Age approach in which a massive data set of attributes is perceived as the core product value. In our work we recognized that attributes were super important (especially those that only universities possess) but our horizon was not simply to serve an explicit need (“I want to search for Duke alumni to connect with”), but instead to deliver an experience which was attuned to impactful outcomes for people who used the product (as in your example above of searching for a tennis partner). Put another way, we based our work on the principle that really effective people search has to accommodate both expressed needs as well as implied needs. As opposed to big social networks, matchmaking apps have been going down this path for some time now, and thus have developed new innovations in this area, but it’s fascinating how no one has cracked the code on people search for non-companionship reasons.

    I applaud your approach and wish you the best of luck!

Leave a Reply

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