On-Demand Economy Toolbox
or How To Make Sure Your Uber for X Startup Doesn’t Suck.
“Make something people want” is something of a North Star for entrepreneurs in startupland. This pithy saying has CEOs navigating their startups in search of user demand at the expense of everything else.
“The business model is secondary, we’ll monetize later!” you’ll hear them say.
This guiding principle might sound simplistic but it’s true north for pure software startups because they can build meaningful consumer products and experiences without any real supply-side costs. Bandwidth and hardware tend to be the only costs of goods sold and Moore’s Law drives these factors consistently lower. That’s why single minded focus on user demand has been a viable startup strategy.
Unfortunately, this strategy is suicide for on-demand startups. Real world logistics and costs-of-goods-sold need to be managed from Day-1. You can’t fake it till you make it when you have real per-unit costs. So “make something people want” isn’t enough for navigating to on-demand success. We need new tools that guide us not just toward demand (the service people want), but also towards an understanding of supply (the cost of providing the service). Here are two frameworks that cover both sides of the equation for on-demand businesses that I’ve found very helpful.
Subjective vs Objective
When using an on-demand service, you can think about supply by asking yourself this question:
“Will I be judging my experience subjectively or objectively?”
If your answer is “subjective” then I bet there are a lot of problems going on under the hood.
1) Subjectively judged work often leads to unhappy customers. Unhappy customers are expensive.
“Did my driver do a good job getting me to my destination” is an objectively answered question; the answer is usually yes. Driving ability is so well understood and objectively evaluated that ride-share companies don’t even need to meet drivers in person to vet their skills and they’ve on-boarded hundreds of thousands of drivers globally. Yet less than 3% of drivers have a rating below 4.7 out of 5 on Uber. Screening and training costs are minimal and empirically we know that customers are happy.
“Is your home clean?” has a more nuanced answer. Some people have borderline OCD standards, others are more lax — outcomes in the cleaning space are subjective and workers have a hard time succeeding. Screening is difficult because soft skills like time and expectation management are hard to evaluate yet they’re requirements for doing a good job in an unfamiliar home. The subjective standards of cleaning mean there are many ways for a customer to end up unhappy. These factors create bigger cost centers in the forms of larger screening, training, and customer service orgs within a startup.
For subjective services, finding a reliable workforce will be costly and customers will suffer as a consequence. This will make churn high on both the provider and customer sides of the platform.
2) Gigs with objective outcomes have fungible providers. This means the relationship value accrues to the platform, not the providers.
When hailing an Uber or Lyft, you don’t care who’s driving you. The nearest 3 or 4 drivers are nearly interchangeable, the only difference being who’s marginally closest. This fungibility of drivers allows the platform to function as a liquid market — your needs are satisfied by any nearby provider and your relationship is with the platform. This effect flows both ways because customers are fungible too. Drivers get work from the platform any time they want it and don’t need to form relationships with customers.
Subjectively judged services don’t work this way. If you’re using Carelinx and you find a caregiver who meets your subjective criteria around cost, reputation, experience, rapport, etc., you will form a relationship with that caregiver once they’ve proven themselves in your home. At that point they’re known and differentiated from other caregivers on the platform and this creates a natural disintermediation scenario. That’s why Carelinx is structured as a lead-gen marketplace. They extract financial value right when a customer connects with a caregiver and they don’t wait to be cut out of the loop later.
Platforms with subjectively judged providers will have a hard time fighting this disintermediation and should look for other ways to provide value. Lead-gen marketplaces can be great businesses and not all markets for services should have an on-demand offering.
Now or Later
Everywhere you look there’s an Uber for something. Uber for Puppies. Uber for Haircuts. Uber for Uber. Looking at Uber and surmising that people must want all the things on-demand is not a recipe for success. To evaluate a new Uber for X, ask yourself:
“Do I need this service NOW, or will I wait for a better provider later?”
Many services look like candidates for Uberification but don’t actually benefit from Uber tactics (haircuts, landscaping, home cleaning, caregiving, chefs, massages, handymen, etc.). These services have many criteria for provider selection like cost, trust, competency, experience, reputation, rapport, etc., and customers will wait to get better providers. Immediacy may not be an important selection criteria.
By contrast, Uber got initial traction because their market valued realtime access. When you want a ride somewhere, you want it NOW. Realtime access always mattered and the winners in the taxicab era had more cars providing broader coverage and therefore better realtime access. Uber created a wedge into this market with technology. The proliferation of GPS enabled cell phones allowed Uber to increase realtime access to rides via a driver-rider-network app and a relatively small fleet. The market valued realtime access so Uber was able to gain a foothold.
Markets that don’t value realtime access won’t value an on-demand offering. The market for lawn-mowing services is driven by cost. A lawn-mower NOW is not better than a lawn-mower that’s cheap. That’s why many young people get their first taste of the working life by mowing their neighbors’ lawns for a pittance. People aren’t picky with lawn-care — they’ll let unskilled children get the job done. Instead of just applying the “Uber for X” recipe, it’s more instructive to think about ways to provide lawn-care at low cost. Let’s build a super cheap service with scheduled, autonomous, lawn-mowing-bots instead of a platform for connecting human landscapers with lawns that need mowing urgently.
With this lens, “on-demand” is sort of a misnomer for many of the startups that get the “Uber for X” label. They’re more accurately labeled “vertically integrated service” companies. Verticalized caregiving startups probably care about using technology to manage the trust economy. Lawn-care startups might have a foothold into the autonomous-robotic-labor economy. On-demand access may not be the primary value proposition to lure customers away from the existing options in the marketplace. And if the technology landscape hasn’t made it possible for a company to optimize for the criteria that actually matters, it might not be the right time to start a verticalized operation.
Tools for thinking let us execute quickly without evaluating from first-principles every time a new situation arises. The two frameworks shared here aim to quickly deliver insight into the supply and demand dynamics of a services marketplace.
Subjective vs Objective helps us think about the supply, and the workforce, and unit costs in a services market. It also gives insight into customer retention.
Now or Later helps us think about the nuances in consumer demand and helps orient businesses to compete well in service categories that have a lot of incumbents.
When dreaming up an Uber for X idea, try to aim for the Objective and Now.
If you found this helpful, follow me on Twitter and let me know what you think.
Thanks to Siqi Chen, Roy Sehgal, Arya Asemanfar, Amitt Mahajan and Tikhon Bernstam for help with drafting this essay.
