On-Demand: It’s About the People, People.

Photo: Patrik Nygren

I love my on-demand services. They are a lazy man’s dream. No need to plan ahead; no need to even move. Yet, I’ve heard more and more anecdotes about how the quality of on-demand services has plummeted recently. Despite all the technology investment and improvement, service qualities have started to revert back to those they replaced. The magic of consistency and convenience hasn’t been easy to scale.

After a particularly brutal recent food delivery where food showed up cold, late and made my wife promise to never order from that restaurant again, I got to thinking about where these services were getting it wrong. In an effort to automate more and more with technology, humans were getting pushed further and further out of the equation.

Many people see the end state of the on-demand world removing people entirely. Software is designed to remove friction and handle complicated, multi-variable problems. It is great for providing connectivity and optimizing systems, but still a long way away from understanding people and handling unexpected situations. I love ordering through an app, but I want a person helping me when something goes wrong. Why wouldn’t I? I fight like hell to reach a person when I call customer support, and still wait in a longer check out line with an actual cashier at the grocery store.

The assumption that customers seek total automation flies out the window the first time something goes unexpected. This is especially acute for real-time services. The reality is that now, and for years to come, people will remain central to mobile service businesses, just as they are to offline service businesses. Don’t replace the workers, help them become more productive and consistent. When something goes wrong, provide resolution tools for them to solve problems on their own.

The companies that own the bridge between automation and great human service have the best chance of avoiding price wars and commoditization. If Tesla develops its own on-demand service with self-driving cars, what will Uber’s largest asset be besides its pool of drivers and the cars they independently own? It is time for mobile service businesses to embrace their workers and create structures for the best ones to shine.

A Look Backwards

Iconic brands like the Four Seasons, Nordstrom and Virgin Airlines were built on the principle that handling these ‘exceptions’ is the key to building insanely loyal customers. The staff at the Ritz-Carlton, for example, uses the motto: “We are Ladies and Gentlemen Serving Ladies and Gentlemen.” The motivating principle is that service isn’t something that tears people down, but builds them up. Staff members view service as a source of pride and camaraderie.

And it’s not limited to the luxury sector. Shake Shack, Southwest and USAA incorporate a high-standard of service into a cost-effective price-point. Zappos built an entire brand around exception handling by making returns free for 365 days. Before Zappos, customers were blamed for most issues and return complexity made apparel purchases online scary. Instead of competing on a lowest-price guarantee, Zappos won with a broad service promise and a unique culture that embraced the consumer’s inherent missteps.

Differentiating with Service

The ridesharing market in the US is primarily dominated by Uber and Lyft, trying to duke it out on price. Uber announced price cuts in more than 100 US and Canadian cities earlier this month for the third year in a row. A week later Lyft followed its lead. According to Bloomberg, Uber’s rates never rose again in a third of cities affected last year — in the long-term, a war on price is unsustainable, especially if you’re Lyft.

Historically, in every other consumer services category, there’s some group of people willing to pay more for a better or more consistent experience. But this is uncharted territory for on-demand. No one’s really built a brand where a level of consistent, quality service is simply expected — these companies have been focused on scaling operations. Perhaps this is Lyft’s opening? Instead of competing with Uber on price, should Lyft try to win with exceptional service? Maybe a return to the human-roots of their brand would be welcomed?

Design for the Exception

While some users might be willing to pay more, the reality is that great service can be delivered without a price increase or margin hit. The key is designing apps to handle the exception without pushing customers to email or social media. As an engineer, exceptions are often thought of as ‘bugs.’ If something went wrong, someone must be to blame — the expectation is that customers are incompetent. However, to the customer, the exception is what’s remembered and where services need to shine. By treating workers and customers (especially the more experienced ones) as intelligent agents, the system can improve with time. People want to feel heard AND they want to contribute to making the system better. An invested audience is a loyal one.

A quick look through the Uber driver forum, uberpeople.net, shows hundreds of posts from drivers unable to handle exceptions. The first flaw here is that the response from support is so infrequent that people feel compelled to go to a third-party site for help. One example on this forum was about charging the right amount when a driver misses a turn — the customer shouldn’t pay for the longer route, but it also doesn’t make sense for the driver to eat the cost of the remaining ride. Honest mistakes happens, but there is currently no “in-app” solution. In fact, shouldn’t letting the driver and rider figure out the optimal solution be the first step of any remediation? With a set list of solutions for them to choose, satisfaction would go up and actual support requests would go down.

From the consumer perspective, why not let users move the pin to make the location more favorable given the matched driver’s actual location? Maybe even let experienced users cancel because they know the pick up time is unrealistic? Let me send a “traffic-related” cancel request to the driver where the driver can decide to opt out of the ride — it is in nobody’s interest to have a driver be stuck in traffic before the rider even gets picked up. These actions create signals that can feed into the matching algorithm and make the system better. Zappos makes it easier to return items, but they do ask for data around the reasoning, which is then used internally and shown to future buyers to help them make a better decision.

People plus Technology

Humans possess the flexibility and initiative to make magic happen, even in the smallest of ways, given autonomy and freedom. As a teenager, I used to work by the pool at the Phoenician Resort — ranked #3 in North America at the time (who’s counting though). If it was a child’s birthday, I could bring her an ice cream sundae. Even as an extremely low-level employee, there was the latitude to ‘err’ on the side of the customer.

HotelTonight (an investment of mine), a company praised for the simplicity of finding a hotel in three clicks, recently added people back into the booking experience. Their “Aces” concierge product connects customers with real people whether you want to request an early check-in or make a reservation at a hot restaurant. The lift in repeat booking more than covers the cost given new technology that helps agents multi-task and solve problems faster. Twenty years after Expedia was launched to make travel agents obsolete, next-generation apps are finding ways to take the best of both worlds and integrate them into one seamless experience.

On-demand services have opened the door for millions of people to earn extra cash. In the race to achieve scale with the complexity of providing training, the default approach has been to remove as many decisions as possible. Standardization is key, but the optimal solution empowers people and lets the best ones shine by earning responsibility and building a career. Pride and opportunity can do magical things when it comes to quality, retention and culture. I’m excited about the next wave of services where the lessons of automation and scale ring true, but people return to the forefront of the equation.

P.S.: Marcela’s recent post on this topic was flagged to me after I posted this. Just wanted to share here since she’s obviously been thinking about this for a long time and putting it into action with her work at Hello Alfred. If anyone else has good posts on this topic, please share in the comments and I’ll update to add to the conversation.