The Center for Data Innovation spoke with Josh Patchus, chief data scientists at CAVA, a fast-casual restaurant chain based in Washington, DC. Patchus discussed the similarities in how restaurants and e-commerce companies use data, and how CAVA uses sensors to improve customers’ experiences at their restaurants.
This interview has been lightly edited.
Joshua New: Most people have probably heard of ways retail chains use data science, but examples of how restaurant chains use data science are harder to come by. Is there much overlap between the two? What do restaurant chains have to consider when it comes to data that other industries don’t?
Josh Patchus: There is loads of overlap. I’d say that we take more of an approach of e-commerce, so we probably align more with a Warby Parker, Bonobos, or Amazon than a classic retail chain. Restaurant chains have to consider scale. If you only have a handful of restaurants it’s probably best to buy things out of the box. I think everyone working in restaurants should gain some experience with the numbers and metrics. Each restaurant has its own challenges and it’s up to everyone to figure out what those are.
New: Can you describe the kinds of sensors used at CAVA restaurants? What can you learn from this kind of data?
Patchus: We have a variety of sensors, mainly ones that monitor environmental variables, temperature, sound, light, and motion. We can learn what environments are best to consume our product as well as what makes the experience that much more enjoyable.
New: Once CAVA deployed these sensors, did anything surprise you in the data about how customers behaved?
Patchus: I think what surprised me most was the sheer amount of data we had. There is absolutely loads of noise in there, and that was what shocked me the most. Obviously customers do not want the volume level to be too loud, they prefer a conversation between the team members and the customers. I was surprised by how much more we could learn.
New: What’s the return on investment like for making an entire restaurant “smart?” Is this only viable for larger chains, or do you think we’ll see this kind of technology proliferate to small, mom-and-pop-style cafes?
Patchus: It is a combination of the two really. We can’t think about today, we really have to think about how our learnings are going to affect our customers four years from now. A lot of the investments we are making now may not impact the customer for a few years, but it allows us to better understand their experience map. I would say that these insights will make it down to the smaller cafes. Some of the best ideas we’ve gotten have been from smaller cafes and it has been a constant feedback loop. The reality is that it’s 2017, and customers do not care if you have one location or 100—they expect a great experience. That’s where the scaling issue comes into play: as you grab more locations, understanding your customers becomes harder.
New: CAVA has grown pretty fast over the past several years and has plans to add 18 more locations in 2017. How much of this fast expansion do you think is attributable to data science?
Patchus: I’d say CAVA is aided by data science, aided more by the tech. We aren’t held back by integrations because we can just build them ourselves. That’s really been the pedal to the metal.