As on-demand rides racked up some controversy this week, the on-demand meal space had its share of news too. First, dark horse, Peninsula-based company Gobble announced its big relaunch comeback, pitting it against the likes of Munchery, Sprig, and SpoonRocket. Then, Sprig finally came to Android and expanded lunch service to four more neighborhoods.
It recruited Angela Wise, who formerly helped run supply and demand optimization at Uber. That’s right, this is the woman who was partially in charge of surge pricing — although before you storm her house with pitchforks, it’s worth noting she was not the inventor. Wise’s resume is formidable. She has two computer science degrees under her belt — a B.A. and a master’s from Rice University — and a Harvard Business School degree to boot.
She’ll be helping Sprig optimize, using data science to predict food demand ranging from favorite meals in favorite neighborhoods to when to expect surge ordering time. With that focus, Sprig is officially taking a page out of the ridesharing playbook, attempting to use technology to make an industry more efficient and cheaper.
Pando spoke with Angela for a Q&A about her new gig, the challenges ahead, and whether surge pricing is coming to the on-demand food space. Best bits below, edited for clarity:
How does supply and demand optimization apply to the food space?
Do people want chicken in the Marina at 6 pm? Maybe everyone who orders chicken has kids and they order at 5:30. It’s that level of analysis. Then you proactively position your supplies to meet that. The goal is to keep delivery time low. One of the things people love about Sprig is this wow factor when food shows up and it’s hot and it has only been twenty minutes.
What are the challenges specific to optimizing the food industry as opposed to transportation?
I think Sprig and Uber have a lot of similarities. There’s a huge amount of hard engineering in making that work. What’s harder about Sprig than Uber is the number of people who are going to want a ride in an area is a relatively simply period of time, compared to who’s going to want the pork entree or chicken in what area at what time.
There’s so many variables when it comes to food that it makes it a harder data problem to solve and a more interesting one as well. We’re concentrating on creating those capabilities on our engineering team so we can do the hard engineering to be as efficient as possible about where our supply goes.
How many data scientists does Sprig have working for it?
I am hiring a data science team now. We’re actively looking. I told Gagan I need free range to hire as many people as I need if I join Sprig, and we’re aiming to hire about two.
So far we’ve been doing light data science with the engineers we have now, but as we scale it’s going to be a bigger problem and the stakes are higher. It ties directly to the customer experience. The only way we can deliver that ‘wow high quality food in a short period of time’ experience is with a team to do heuristic predictive analytics.
It doesn’t seem like supply and demand data science worked for Uber. We still get surge pricing. What makes you think you can optimize an even more complicated endeavor like Sprig?
Occasionally you have a marketplace and things don’t work according to plan. Even great data science doesn’t work 100 percent of the time. But we’re so early in Sprig’s attempts to attack these problems that I don’t think we necessarily know what the solutions are going to be. The problems are the solutions in the end.
Could you see surge pricing coming to the on demand meal space?
It’s too early to say one way or another. But we’re definitely going to be evaluating each situation realistically so it can scale not just to San Francisco but theoretically everywhere.
We’ll do whatever supply and demand things we need to to make sure customers have a great experience. But surge pricing isn’t necessarily the right solution for us at this point whereas it definitely was for Uber.
What are the first problems you’re going to tackle?
We’re trying to build an operational foundation that’s scalable. That’s priority number one and that’s what we’re going to be putting these data scientists on as well. Occasionally we’ll have a really busy night and delivery times won’t be what I want them to be. Those are the things we’ll tighten up in the future so people never have a bad experience with Sprig.
What are the biggest challenges you’ll be facing?
It’s really hard to know, but the unpredictability of food is something no one understands. With Uber deciding whether someone gets a car or doesn’t, it’s binary. But with food it could be a cloudy day, people have a different feeling and want to eat different things. Solving these problems is going to be really hard, understanding it at a granular level.
That’s why it could be such a huge opportunity. If we can solve some of these questions and understand at a deep database level what people want from a food perspective, this could be a bigger opportunity than transportation.
[illustration by Brad Jonas for Pando]