Belly Up to the Candy Bar

How Viewing Pedestrian Traffic in New York Helped a Dynamic Business Find the Best Location

By Karen Richardson, Esri

Dylan’s Candy Bar, a Willy Wonka-like retail outlet dedicated to the sweet things in life, gets a boost in choosing a new retail outlet — thanks to a new data layer that displays pedestrian traffic throughout New York City at various times of the day.

Forget the phrase “like a kid in a candy store” — Dylan’s Candy Bar has raised the chocolate bar to new heights. Adults feel right at home in the 13,000-square-foot, three-story flagship store, which also boasts a café and even a bar. Customers over 21 can imbibe in specialty drinks like Pink Cotton Candy, made with Pinnacle cotton candy-flavored vodka, and the with the fruity confection. Located on the Upper East Side in New York City, the boutique candy store provides more than 5,000 different types of goodies including novelties like Hello Kitty cookies and Wizard of Oz ruby slipper lollipops. There are themed sweets for holidays from Hanukkah to Halloween. And the company will help you hunt down those hard-to-find childhood treats or make your own personalized chocolate bars.

At Dylan’s Candy Bar, life is indeed sweet. When the company wanted to expand its flagship store to another location in New York City, it contacted start-up company Placemeter, as David Reid, Dylan’s Candy Bar’s CFO, says, to “get the biggest bang for our buck.” Placemeter staff uses Esri ArcGIS software ArcGIS to implement their network of optical sensors that measure foot traffic around New York City and in other markets. They analyze the information, including how crowded a certain area is at different times and how that changes throughout the day. This data is used to create the world’s first real-time dynamic data layer of side-walk and street traffic measurements from video feeds.

The data will be available as a data layer in ArcGIS Online in 2015, when the company will also be expanding to other major cities worldwide. “Just as you might check the demographic constitution of a place, along with what competitive retailers are already located in an area, now you can also look up how many people walk by a location at various times of the day,” said Jason Novack, a solutions architect with Placemeter.

Putting Analytics in Its Place

Like most successful companies, Dylan’s Candy Bar has internal analytics on its customers including how much each can be expected to spend after walking through the door. What the company needed to find out was whether it could attract enough of these customers to the proposed new location. That’s where Placemeter comes in.

“We were interested in finding the best analytics to ensure we find the right location,” said Reid. “Generally speaking, commercial leases usually last for 10 years. If you have the wrong base to start with, you are locked into that lease. We needed to figure out if we could attract the right amount of people to our store to negotiate the right contract.”

Dylan’s had its eye on a couple of locations in particular, but it couldn’t decide which would yield more customers to keep the chocolate fountains and length of the anticipated lease.

“When we started modeling the sites, we thought about the different ways we could do this,” said Reid.

“Unless we conducted a physical count with someone camped out at that location — and paid handsomely for that data — we were not going to get a good measure of the business passing by our proposed storefronts.”

The Placemeter data provided to Dylan’s Candy Bar helped it decide on the right location for the new store. “We have found out that there is a lower chance that people in the early morning will come into our store — they are usually on their way to work. Candy might be the furthest thing from their mind,” said Reid. “But someone strolling by at 11:00 a.m. in downtown New York — that might be a tourist or a person out shopping. There is a higher chance that person will go through our doors and make a purchase.”

ArcGIS Ensures the City Is Covered

Dylan’s Candy Bar used the aggregated data from the Placemeter sensors that recorded the comings and goings of pedestrians for the two locations. The sensors in this case were company-installed, live-video cameras, but Placemeter also runs a program in which people can apply to place a sensor at their place of residence or business.

These sensors can be as simple as an old smartphone that is propped up in a storefront or apartment window. Placemeter checks to see if the location has a street view and if the location is needed for coverage. If the application is reviewed and accepted, the company will pay a stipend each month for the data feed. Placemeter uses ArcGIS to position optic sensors around New York City. The company then uses computer vision algorithms to turn these feeds into real-time data about locations, streets, and neighborhoods. The information collected includes historical and real-time pedestrian and vehicle counts, speed, dwell time, and other measurements.

“ArcGIS provides us with the spatial analysis power we need to understand where we have gaps and which areas people are requesting the most,” said Novack.

“It’s a lot like trying to figure out what places we are missing in the middle of the ocean, but in this case, our ocean is Times Square. We couldn’t get that analysis in any other mapping software we looked at.”

When Less Is More

In Dylan’s Candy Bar’s case, staff chose a location that yielded consistent traffic throughout the weekday. “We weren’t too far into the project when we made our decision,” said Reid. “Our key driver is understanding how much foot traffic we need to make each dollar spent for rent the most profitable. Placemeter helped us achieve this quickly.”

Dylan’s use of the data doesn’t stop when the ink dries on the rental agreement. Reid is toying with the data to understand other operational functions at the store, such as analyzing why one store location might be doing better than another, or whether to give away coupons outside to drive traffic into the shop and when to do so.

“When you start to pull in lots of data from extended periods of time from thousands of locations in the city, you start to build an understanding of what to expect in daily patterns from the locations,” said Novack.

This is especially interesting for retail chains that are doing site selection. Many don’t have a long history of information in local markets, so they can get a lot of data-driven information based on the kind of pedestrian and vehicle traffic in an area and discern which locations are best suited for their needs.

Reid agrees: “If someone rents for a certain price, then it is assumed that the neighboring location will rent for a certain amount as well. But realistically, that location could be so different from the other. There could be significant idiosyncrasies between two sites. This type of data really helps retailers have more control over the costs of finding the appropriate site.”

One clap, two clap, three clap, forty?

By clapping more or less, you can signal to us which stories really stand out.