Investing in On-Demand
--
This post is part of a series from the On-Demand Conference, organized by Pascal Levy-Garboua, Semil Shah, and Tradecraft. For a full list of posts, click here.
My favorite part of the on-demand conference was the investor panel. This heated discussion, illuminated some of the fundamental principles on-demand startups need to be contemplating. Here are some of the most salient points.
TL ; DR (Highlights from the panel)
Don’t get into a market where the competition is seeing real growth and is starting to approach liquidity. Convenience is table stakes. Consumer delight and value (premium experience without a premium price) are the new differentiators. Don’t avoid investing or building just because you don’t have line of sight to the unit economics. If you want to get big, start small. Don’t replicate someone else’s playbook, create your own. Know where things have broken in the past. Don’t go to new cities before you’ve figured out the playbook. Benefits have to be decoupled from the organization and attached to the individual.
Don’t be the sixth company in a space
We’re not seeing any slowdown in the number of new companies in the on-demand space. Even in the areas that are really well financed (e.g. in transit). It’s ok to not be the first company in a space, but you should not be the sixth. The odds of five companies ahead of you falling apart is probably not a good thesis. Once someone in a space has seen real growth and is approaching liquidity, don’t get in. Even if that liquidity is in a narrow niche or geography, it’s probably a time to move on to another space.
Once someone in a space has seen real growth and is approaching liquidity, don’t get in. - Simon Rothman, [tweet this]
What investors look for in an On-Demand startup
Satya Patel (of Homebrew) spoke about how companies have to get away from convenience as the differentiator. Convenience is table stakes. There are three things they look for as must-haves from a product standpoint. a) Convenience - the core feature. b) Delight - how is this dramatically better than the existing experience or alternative experiences. c) Value - we like the feeling of a premium experience, but not for a premium price.
From the business model Homebrew likes to see companies who see their customers not only as the demand side but also as the supply side (e.g. Uber riders becoming drivers).
How to think about unit economics
Homebrew wants to see fundamentally sound unit economics. It can’t just be theory, it has to be somewhat proven. In part, this is due to Homebrew not having the luxury of additional pools of capital to support these companies if the unit economics don’t work out.
Worry about unit economics but don’t avoid investing or building just because you don’t have line of sight to unit economics. - Simon Rothman [tweet this]
Simon Rothman (of Greylock) pushed back on Satya Patel (of Homebrew) saying with unit economics it is very easy to get the wrong signal. If you look at history, everyone has been very focused on people who already make a lot of money, have large margins or spaces that feel really big. If Uber has taught us anything, it’s that spaces can grow. You can build out spaces.
Greylock took a swing with Sprig, got in early with the seed and led the A. At the time there was no way to understand what the unit economics were going to be. However, they had a belief in the space, a belief in the founder and it felt like it was part of the future. How are we going to live in a world five years from now where we don’t press a button and get quality food fast? “It just felt obvious”. They did some rough modeling and just believed they’d figure it out.
Sometimes you have to take a bet on a founder, a space, a road map and hope you’ll figure it out - Steve Schlafman [tweet this]
Steve Schlafman agreed, he said it was similar to how RRE Ventures approached Breather. The market for getting space on-demand didn’t exist and there’s not a proven model behind it. However, sometimes you have to take a bet on a founder, a space, a road map and hope to figure it out.
Need confidence that the team has their hands on the levers - Patricia Nakache [tweet this]
Only after these companies have invested heavily in backend operations and get to a certain amount of liquidity do the unit economics start working. This is why Trinity Ventures needs to have confidence that the team has their hands on the levers and that they understand what it will take to get there. It can be easy to put off the moment of truth where you prove that those unit economics don’t work and thats when companies get into the danger zone. Don’t put it off.
This is hard stuff. It bothers Simon when he sees great companies with great ideas that are operations (ops) intensive and don’t have ops as a core competency. It’s easy to believe we can just figure out operations but this isn’t an IQ thing. This is where the experience of measuring nickels and pennies matters. Patricia Nakache (of Trinity Ventures) agreed, adding they need to have a lot of grit. Simon thinks your first big hires should be someone who is great at operations and someone who is great at government relations. If you don’t have this early on, it will be a big problem down the road and that will be hard to unwind after the fact.
How narrow should you go?
Patricia Nakache says you really need to be solving a big problem that can justify a stand alone service. For example, groceries are definitely a big enough vertical to justify their own service. She doesn’t believe “Uber will eat the world”. Steve Schlafman said if you believe that ‘mobile devices are remote controls for the real world’ then it’s not just these obvious industries like groceries. Every industry where they have service based professionals is now up for grabs.
Patel reminded the crowd that there are a lot of these well defined niches, but if there is not a path to the billion dollar pay out, then it’s better off not taking VC money and bootstrapping. Simon’s view is that if you want to get big, start small. So find a tight focal point that can eventually wedge into something broader, not a thin niche that can’t expand. The founder should know the difference and often if you don’t know the difference, odds are it’s a very thin niche and not a wedge.
I believe every single professional service or product field can benefit from improved customer experience or efficiency. - Satya Patel [tweet this]
Patricia thinks that the more high stakes the service, the less likely you would be to use the on-demand service. For example, she joked that she doesn’t see on-demand surrogate mothers taking off. Simon disagreed with her premise that higher stakes deter use. Using medicine as an example where the outcomes matter, often meaning life and death. Most people would say that the doctor-patient relationship is intimate and needs to be personal. However, if we look at the range of quality doctors and people who live in markets that can’t get great doctors. If technology could remotely deliver that care, these people could get access to the world’s best and that trumps any relationship with a mediocre doctor. Satya agreed saying, there are big verticals where there is information obscurity or supply scarcity based on geography or other dimensions that can support really big on-demand business.
Making the playbook
First you have to have the playbook for how to roll out each new city, but even then, every location is almost like doing another startup. Each city is different, you have to get supply again, you have to customize it. You also can’t just replicate someone else’s playbook. Odds are it’s not going to be right for you, so create your own. You really need to know what gets a city humming. Take the time to really understand your own operations, know where things have broken in the past or where there are weaknesses in the system so you can properly structure your roll out.
It’s important to build a culture of knowledge management and learning. You really need to understand what the best practices are so you can replicate them across your network. The default should always be to keep things in-house close to headquarters vs decentralized. Local presence isn’t always mandatory. How much of this can be done through people vs how much of this can be done through technology? If you can create a national footprint it’ll be significantly easier to execute.
Liquidity is more important than geography
It’s very natural for founders to want to expand fast because they feel like it’s a land grab. However, you need to get it right in your first city before expanding. You have to understand all of the details before expanding. From how to unlock both supply and demand, make the user experience amazing, get repeat use, etc. If you’ve spread too quickly, while you may be bigger, you are actually much weaker.
The probability of success proportionately declines with every city added before actually fully figuring out the model. A bigger footprint is not impressive for investors and entrepreneurs who actually understand the space. You need to take your foot off the gas to be able to go faster. It may look like you are losing because people are going city to city, but you are actually winning because you will get super efficient at launching cities. In the long term you are going to be in a much better place.
Where To Start
It’s important that the use-case is universal and the way it’s being solved is clean and focused. It doesn’t matter too much where it starts, as long as the founders are living there. This is because they need a tight feedback loop.
Typically focusing in dense urban areas makes operational sense because it’s easier to build liquidity and get validation. However, suburban and rural environments are really interesting for business that don’t require much local presence. For example, doctors on demand, therapy, nutrition, etc; wherever there aren’t local experts but there is a demand for access to them.
What is ‘right’ for On-Demand labor?
The panel agreed the 1099 and W2 framework is broken. It was created in an outdated world. We need a third class of worker. We need to decouple things that have been traditionally attached to corporations (e.g. benefits) and make them independent so people have flexibility. These decentralized models enable the assets to also become decentralized. This opens it up to increased opportunity for new, more profitable models. For example, Uber doesn’t own their own cars. This makes it possible to make more money that can then be pushed back to the customer or the employee.
I call these the Uncollared Worker - Simon Rothman (in reference to blue vs. white collar workers)
However, what wasn’t clear was what we do in the short term. Satya was concerned that not enough of on-demand companies are doing what’s right. They need to treat their employees like customers. Not just giving benefits, uniforms and working wages, but also real career paths.
Simon said if you give control and cash to workers, allowing them to work when they want, how they want and pay them much better than their alternatives, there are enough economics to cover the other issues. He wasn’t sure making them full time employees is as good for them as that control.
Steve agreed. We need to put ourselves in the employees’ shoes. He sees this with ManagedByQ who often say people are our interface and keeping them happy will ensure they will be the best interface we can provide. He said we talk about supply being a problem, but really it’s churn and utilization of supply. If you empower workers, they are more likely to stick around longer and be really committed to the company, thereby solving some of the supply issues.
Patricia talked about when customers ask Uber drivers how’s it going, their stories impact the customer experience. She went on to highlight the danger of how employees could get locked into these jobs. For example, an Uber driver could be locked into an auto loan. Now they have to keep driving to pay it off. These type of lock-ins take away that control Simon was advocating for. His response got the crowd laughing, “for the record, I’m against indentured servitude”.
Satya said control and cash are nice, but not enough. The infrastructure doesn’t exist for employees to protect themselves from liability, health and financial concerns. When employees have to figure out that the infrastructure doesn’t exist to make it easy for people to protect themselves from liability, health and financial stand points. To which Connie Lolzos (Moderator) replied, “maybe that’s the next opportunity”.
In the end, the panel agreed benefits have to be decoupled from the organization and attached to the individual. However, they ran out of time before they could come to any consensus on what to do in the short term before these changes are made.
Summary
