Latest / Franchise Conversations with Fexingo: Buying, Running, and Scaling Franchise Businesses / How Franchisees Use Data Analytics to Pick the Perfect Location
Transcript
- Lucas: So there's this franchisee I've been following — owns 14 Dunkin' locations across central Florida. Last year he opened three new units, and all three were profitable within six months. His secret wasn't better coffee or better staffing. It was site selection — and he's using data analytics in a way most franchisees still don't. Luna: Site selection feels like one of those things people think they can just eyeball. 'Busy intersection, lots of cars, good visibility' — that kind of thing. Lucas: Exactly. And that's how you end up with a store that looks great from the street but has zero repeat customers, because the surrounding households earn $35,000 a year and your product is priced for $75,000. The data revolution in franchising isn't just about AI inventory or dynamic pricing — it's about where you put the box in the first place. Luna: What changed? What tools is this guy using that the old-school franchisees aren't? Lucas: He's layering three data sets. First, foot traffic from Placer.ai — so he can see not just how many people walk past a location, but where they come from, how long they linger, what other stores they visit. Second, demographic segmentation from Esri's Tapestry system, which breaks neighborhoods into 67 distinct lifestyle types. Third, competitor density — not just other Dunkin' locations, but Starbucks, McDonald's, even Wawa gas stations that sell coffee. Luna: The third one is interesting. I've seen franchise agreements that actually restrict how close you can build to another of the same brand. But competitor data feels like it's harder to get. Lucas: It's surprisingly available now. There are services like SafeGraph and Unacast that aggregate anonymized mobile location data. You can see exactly how many people visit a Starbucks between 7 and 9 AM, and whether they're coming from home or from the highway. That tells you if you're competing for commuters or for neighborhood regulars. Luna: So the Florida guy — what's his actual process? Does he run a model for every candidate site? Lucas: He told me he runs what he calls a 'three-mile radius analysis.' For any potential location, he pulls data on every household within three miles — average income, age distribution, homeownership rate, commute patterns. Then he weights those variables against his brand's typical customer profile. Dunkin' tends to overperform in blue-collar suburbs with high homeownership and a 20- to 30-minute commute. If a site doesn't score above a certain threshold, he walks away. Luna: Three miles seems specific. Why not five? Or one? Lucas: That actually comes from a 2022 study published in the Journal of Retailing. They analyzed 1,200 quick-service restaurant locations and found that 80 percent of a store's repeat customers live within a three-mile radius. Beyond that, the drop-off is steep. So three miles is the sweet spot — it captures the core trade area without diluting the data with long-distance one-time visitors. Luna: And this guy's failure rate — you said he cut it by 30 percent? Lucas: He told me that before he started using this system, about 25 percent of his new stores were underperforming after the first year — meaning they weren't hitting the minimum revenue threshold he needed to cover franchise fees and rent. After he implemented the data-driven site selection, that number dropped to about 5 percent. Over 14 stores, that's a huge savings. A bad location can cost you $200,000 in the first year alone between build-out, inventory, and operating losses. Luna: And that's just his stores. But franchise systems as a whole — how big a problem is bad site selection? Lucas: The International Franchise Association did a survey last year — they found that nearly 40 percent of franchise failures within the first three years are attributable to poor location choice. Not bad management, not undercapitalization — just picking the wrong spot. That's enormous. And yet most franchisees still rely on the franchisor's real estate team or their own gut instinct. Luna: Speaking of gut instinct — there's a temptation to think data takes the human element out of it. But isn't there still judgment involved? Lucas: Absolutely. The data gives you probabilities, not guarantees. A site might score well on demographics but be on a street that's about to undergo a two-year construction project. Or the rent might be too high relative to projected revenue, even if the foot traffic is great. The model is a starting point. The franchisee still has to physically visit the site at different times of day, talk to neighboring business owners, check the local zoning pipeline. Luna: So the best franchisees are combining data and boots-on-the-ground. That feels like the right balance. Lucas: It's exactly that. And the tools are getting cheaper and easier to use. Placer.ai used to cost tens of thousands a year. Now they offer a tier for small businesses at a few hundred bucks a month. Esri's Tapestry reports are available as a one-time purchase for under $500 per report. There's really no excuse for flying blind anymore. Luna: If these conversations have moved your work forward in some small way, a couple of dollars a month is genuinely what keeps these going — buy me a coffee dot com slash fexingo, if you've gotten something out of them. Lucas: Yeah, listener support makes a real difference. Keeps the show ad-free and focused on what matters. Luna: Alright, back to location data. One thing I want to dig into — what about urban versus suburban? Does the three-mile radius hold up in a dense city like Manhattan or San Francisco? Lucas: Good question. In dense urban environments, the trade area shrinks. A study from NYU's Stern School looked at Manhattan quick-service restaurants and found that 70 percent of repeat customers live within a half-mile. So the same principle applies — you just adjust the radius. The key is to use a data-driven approach that matches the local density. A franchisee opening a smoothie shop in downtown Chicago should be analyzing a half-mile radius, not three miles. Luna: Makes sense. And what about second-generation sites — spaces that were previously a restaurant or retail store? Is that easier or harder? Lucas: It can be easier if the previous tenant had a similar customer base. But there's a trap: the previous tenant may have failed because of the location, not because of their concept. So you need to do a post-mortem. Why did they close? Was it rent, management, or just a bad spot? A lot of franchisees get excited about turnkey spaces and skip the analysis. That's how you end up with a store that's had three different concepts in five years. Luna: Let's talk about one more tool — regression models. You mentioned that earlier. How does a franchisee actually use regression for location? Lucas: There's a relatively simple linear regression that predicts first-year revenue based on variables like median household income, population density, traffic count, and proximity to anchors like Walmart or a hospital. Some franchise brands actually provide this model to their franchisees. But many don't, so franchisees have to build their own or hire a consultant. The Florida guy I mentioned uses a model that explains about 70 percent of the variance in his stores' revenue. That's pretty powerful. Luna: Seventy percent — so the model gets you most of the way there, but you still need to account for the other 30 percent. Lucas: Right. That 30 percent includes things like local management quality, marketing effectiveness, and random events. But a 70 percent predictive model is way better than flipping a coin. Luna: What's the most surprising variable that shows up in these models? Something people wouldn't expect? Lucas: For Dunkin' specifically, it's 'average commute time.' Areas where the average one-way commute is between 25 and 35 minutes correlate strongly with drive-through coffee sales. People are stopping on their way to work. If the commute is shorter, they brew at home. If it's longer, they might be taking a train or carpooling. That granular insight is something you'd never get from gut feel. Luna: That's fascinating. So a franchisee could look at a town with a 30-minute average commute and target it specifically. Lucas: Exactly. And you can overlay that with household income. If the commute is right but income is too low, you'd price your menu differently or open a smaller format store. Data lets you adapt your strategy before you sign the lease. Luna: What about the franchise brands themselves — are they catching up? I know some big ones like McDonald's and Subway have sophisticated site selection teams. But what about smaller franchises? Lucas: It's mixed. Larger brands have dedicated real estate analytics teams. But many mid-sized and emerging franchises still use a one-size-fits-all approach, sometimes based on outdated demographic data. That's where a savvy franchisee can get an edge. By doing their own analysis, they can spot opportunities the franchisor's team missed. Luna: Give me an example of a miss. Lucas: There's a case from 2024: a franchisee of a popular frozen yogurt chain wanted to open in a suburb outside Austin. The franchisor's real estate team approved a site near a high school, assuming teenagers would be the main customers. But the franchisee ran his own analysis and saw that the area's demographic was aging — lots of retirees moving in. He insisted on a site near a senior center instead. That store now does double the revenue of the franchisor-recommended site. Luna: So the franchisor's data was just a few years old and missed a demographic shift. Lucas: Exactly. Demographic trends change faster than most franchise systems update their site selection models. That's why independent analysis is so valuable. Luna: Alright, so for a franchisee listening right now who's about to open their first or second location, what's the one piece of advice you'd give? Lucas: Don't sign a lease until you've run a regression on household income within your brand's effective radius, and until you've physically visited the site at three different times: a Tuesday at 10 AM, a Friday at 5 PM, and a Saturday at noon. That combination of data and observation will save you from 90 percent of the bad locations out there. Luna: That's actionable. And on that note, I think we've given listeners a lot to work with. Lucas: Yeah. Location is still the single biggest determinant of a franchise's success. And the tools to get it right are more accessible than ever. Next episode, we'll talk about something I've been curious about — franchisees using revenue-based financing to fund multi-unit expansion without giving up equity. Luna: Looking forward to it.