Getting smart about store locations

2017 has been the year of the “retail apocalypse” (the term is now legit, see its very own wiki entry). Since January this year, retailers have announced plans to shutter more than 6,700 stores in the U.S., according to Fung Global Retail & Technology, a retail think tank. The last peak of store closures was recorded in 2008 during the financial crisis, where 6,163 stores were closed, according to Credit Suisse.
 

As the retail industry evolves and tries to get smarter about its real estate investments, it is being joined by companies such as LocateAI, who hope to help retailers make smarter real estate decisions. Retail Atelier catches up with LocateAI’s founding team, Peter and Stephen, here in San Francisco.


Retail Atelier: How did you end up here? What’s the founding story? LocateAI doesn’t strike me as something that you knew you wanted to build as a child or as an undergrad at Stanford.

LocateAI: We’ve been working on this since 2015 when we part of TechStars in Seattle. We were conducting user explorations and one day we got connected to a restaurant owner in Seattle who was trying to find a new location. He’d just raised $1M for expansion but was struggling with choosing new restaurant locations because he didn’t have enough data to back his decisions and felt limited to his broker’s reporting and factsheets.

Retail Atelier: That was the moment, wasn’t it? I wish I’d been there to see you scrambling to build your prototype!

LocateAI: We realized it was a big problem, a data driven problem, the data he needed to make his decision was not transparent to all parties. Even worse, his brokers did not even have a view of the data he wanted and couldn’t answer his questions. After more digging, we also learned that the equation was even more asymmetrical; brokers are also incentivized to sell their own listings as they get a better and bigger cut. Our team of Stanford data scientists ended up working with this restaurant owner on our first prototype, which he loved and found useful!

Retail Atelier: What did you end up building?

LocateAI: There have been many versions after our first prototype with that restauranteur. Today, LocateAI uses predictive models powered by machine learning to help retailers, developers, and brokers make better real estate decisions. For our machine learning models, we use over 100,000 variables as input data - some of which include demographics, real estate data, traffic, businesses, and consumer mobile data. Within each of those categories you can derive hundreds or thousands of features for machine learning.  We work with each customer to fine tunes models specific to their businesses; no two models are the same as each customer will have custom requirements specific to their operating models.

Retail Atelier: I know that retailers typically don’t have teams of data analysts/scientists working with them to choose new store locations. What are some of their initial questions to you?

LocateAI: One of the first things we usually work with them on is filtering out locations that they’ve already shortlisted, based on scoring composed of demographic signals. We’ll help them eliminate and then prioritize locations based on the criteria they care about, for example: age ranges, competitor proximity, income levels, and ethnic distribution.

LocateAI flow.PNG

Retail Atelier: Working individually with each customer must be really gratifying, what are some of the most exciting questions you’ve helped them answer?

LocateAI:  Our customers usually have a working hypothesis before they come to us, we then help them gain clarity and speed up their go-to-market process. Examples of what we’ve partnered with them on include:

  • Scoring metro statistical areas for a prepaid mobile company by diversity ratio and income level

  • Revenue correlation for a fast food chain’s proximity to grocery co-tenants such as Trader Joe’s, Whole Foods, and Safeway

  • Weighing airport proximity against competitor distance for a modern fried chicken joint

  • And sometimes they want to look internally at saturation: are their own stores cannibalizing each other?

Retail Atelier: These sound like top-of-mind strategic initiatives for some of our favorite brands. How many customers are you working with today?

LocateAI heatmaps.PNG

LocateAI: We’re working with more than a dozen customers across the retail sector, including Sprint and Boost. I think our customers experience a moment of delight when they see our user interface: the heatmaps are a departure from the Powerpoint brochures and spreadsheets that they are used to agonizing over. Those that have paid for this sort of intelligence in the past are similarly delighted. They are used to older and more static models that leverage linear regression. They say that the speed and accuracy of our LocateAI software is unmatched. We also have the technical advantage; machine learning not only allows us train our models on tens of thousands of data variables, it also provides the ability to iterate through millions of scenarios to pick up on insights which are often completely missed by traditional models.

Retail Atelier: If I were a retailer trying to plan to achieve my monetization goals by choosing the right retail sites, I’d be worried about this stat recently released by the Bureau of Labor Statistics: apparently the number of restaurants is growing at about twice the rate of the population. Whether they are in the food service vertical or otherwise, are you seeing a slowdown in your customers’ appetites for growth?

LocateAI: Actually, we’re still getting queries about new site selection. Our customers are just going about growth in smarter ways than before. Making a calculated risk based on site revenue projections is the way to go when the market is uncertain. The retailers we are working with still have their mind on growth and we are still providing a lot of support for new site selection as a service, they are asking, “how do I get the most from my chosen market? How do I balance expansion with cannibalization?”

Retail Atelier: Well, that’s good news for me as a shopper and as a retail blogger. How can we see the product?

LocateAI: You can check out our website here, or book a demo! As a reminder, we’re always hiring too!

LocateAI base.PNG