Learned Behavior: how the best products eventually change their users

When I evaluate product, either at the early or late stage of a startup, I’m always looking to see if the product has the potential to permanently modify user behavior outside of the product itself

A few examples of this:

  • Instagram : double-tap
    Apple designer Bill Atkinson devised the double-click, and soon it became the de-facto way to open files across all desktop GUI’s (Windows, Linux). The clever folks at Instagram appropriated it for their Mobile application to “like” a post. I’ve found myself mindlessly double-tapping photos in other applications. (NB: In light of this, I’m not sure I agree with secret’s choice to make the “like post” gesture as a left-to-right swipe.)
  • Netflix : binge consumption
    The way today’s content travels is much more binary – you’ll either never see the light of day or go viral (think: YouTube videos, 2048, any Upworthy article). To put it differently, today’s content distribution models enable and encourage binge consumption. Netflix is one of the first large tech companies to realize and capitalize on it by releasing traditional content in a way that enables binge consumption. Will that eventually put pressure on other content distributors?
  • Uber : pay through app
    It has become so easy to take an Uber that I’ve found myself walking out of regular cabs without paying. Embarrassment aside, it comes to show that Uber is creating a new behavior of seamless payment, something that Square and other startups have been attempting to do. 

The best products eventually change user behavior because they simplify those existing user behaviors; making it so natural that users have incorporated the habits into their regular lives. 

Uber and the Economics of Price Surging

There’s been tons of emotional outrage and blog posts on the repercussions and implications of Uber’s price surges. Uber’s been addressing this criticism with the protest that price surging is just a result of free market forces. As an Economist by training, I can’t help but add my two cents, because as an Economist by training, I also know that free market mechanisms rarely work as perfectly as they do in textbooks. 

  • Inelastic Supply: Uber explains that by raising prices during demand peaks, they are “putting more Ubers on the road” by increase driver incentives. Uber supply, however, is not perfectly elastic – there is a limited supply of Uber drivers in a given location and each driver is often faced with other constraints that doesn’t let her respond immediately to increased demand. (In Economic terms, while Uber attempts to move the supply curve to the right, it can mostly only move along the same supply curve.) If Uber’s market mechanics were elastic, when we would see price surging for just the amount of time it takes supply to increase to meet demand. Instead, I’m pretty sure price surging typically occurs for the entirety of a peak demand period.
  • Driver Incentives: I’m not even certain that price surges align driver incentives properly. Sure, getting 3x the fare seems great, but I would imagine that demand probably drops materially during price surges. Perhaps not enough to completely off-set the fare multiple (ie. demand doesn’t drop 66% in a 3x price surge situation), but probably very close to it. 
  • Income Targeting: While this is a minor point, income targeting may help explain some of the supply inelasticity. Economic research has shown that taxi drivers tend to be target earners – meaning that regardless of demand, taxi drivers tend to work until their daily income target are met. By the same token, drivers who have already hit their daily income targets will likely stop working, instead of increasing their hours to meet higher demand.

Uber also insists that price surging assures rides to those who need the rides the most, but it’s difficult to take that statement at face value. For one, Uber’s price surging is done as a multiple of the base fare, but a more accurate way of pricing according to demand would be as a % of income. The latter is an obviously more difficult methodology, but wouldn’t automatically out-price lower-income customers who may just “need” that ride more than someone with more purchasing power. At its core, Uber’s price surging is essential to its mission of guaranteeing rides when you need it, only because it out-prices the majority of the market.

I’m disappointed by Travis Kalanick’s comparison of Uber’s price surging to the high cost of air travel during the holidays. Sure, the supply/demand mechanics work in a similar way, but airlines are faced with constrained supply in a way Uber insists it’s not. Moreover, airlines probably have one of the lowest customer satisfaction ratings and probably isn’t and shouldn’t be what Uber aspires to be.

I’m an avid Uber user in San Francisco and will continue to be. I have, however, been surprised by some friends (mostly in NYC) who have decided to no longer use Uber due to the politics of price surging.

Uber, but for X

A couple of days ago, I attended a #devden talk at Atlassian featuring Chris Chambers, head of Engineering at Uber. He walked us through many of the challenges that Uber faced in its early days and the subsequent improvements they’ve made to their stack and UI.

While most people tweeted about the Google + Uber partnership as the beginnings of a fleet of self-driving cars, Chris focused more on the maps and infrastructure aspect of the business. The ETA’s calculated by Uber’s platform were more precise than those calculated by Google Maps. In fact, Uber’s back-end totally blew me away. It was incredibly interesting how they were trying to forecast demand, in real-time, as well as supply matching & positioning.

I couldn’t help but think how this commercial technology can also serve the public / non-profit sectors. People often complain about policy / ambulance / fire dept. response times – imagine if Uber can help reduce that by better positioning patrol cars. Maybe even effectively reduce crimes and accidents by having patrol cars positioned pre-emptively. Effectively dispatching help would also be crucial during states of emergency, such as national disasters.

How do you see Uber’s technology deployed in other areas?