At first glance, Flagstar Bancorp’s acquisition of 52 branches from Wells Fargo looked like a perfect opportunity. The deal, announced in June, offered the promise of cheap deposits for a company with a pressing need for funding. In total, the Troy, Mich., company said it would pick up roughly $2.3 billion in deposits, as well as $130 million in loans, through the acquisition of Wells storefronts in Michigan, Ohio, Indiana and Wisconsin. One analyst described the planned acquisition as potentially “transformative” for the fast-growing Flagstar. But after the deal closed last week, a different reality set in. More customers than expected opted to keep their accounts with Wells — likely, analysts speculated, because some had moved away, or simply felt comfortable banking digitally without a branch nearby. Flagstar, by the final count, acquired $1.8 billion in deposits, or 20% less than initially projected. Then, on Dec. 3, the first day after the conversion, Flagstar’s website crashed, leaving all customers— including its nearly 200,000 new ones — unable to access their accounts. The company is currently investigating the root cause of the crash and says that service has since been restored for all but a small number of customers. “If I could apologize to every customer that couldn’t access their account on that day, I sure would,” CEO Sandro DiNello said in an interview. “It’s probably one of the more disappointing things that’s happened to me in my career.” The issues at Flagstar, while specific to the company, also underscore some of the risks involved with branch deals in the digital age. Among them: As big banks continue to improve their digital banking capabilities, customers, particularly from larger banks, have become less willing to make the switch after their local branches are sold. There’s also little room for error in the conversion process, any sort of online glitch can damage the company’s brand. Over the past few years, banks have begun to shy away from branch deals — once an attractive option for banks to scale up in new markets. So far in 2018, there have been 45 branch deals — or 48% fewer than three years earlier, according to data provided by D.A. Davidson. As recently as 2012, by comparison, there were 104 deals. While banks are trimming back their branch networks, many have simply been unable to find other banks to buy their brick-and-mortar locations, said Kevin Reevey, an analyst with D.A. Davidson. In some cities, branches have sold to and redeveloped as restaurants or other retail stores, he said. “A lot of banks have been closing branches since they haven’t been able to sell them,” Reevey said. “And a lot of companies are relying more on mobile technology as a distribution model.” Still, for some banks, particularly those with branches in rural areas, branch deals make sense to increase scale. Associated Banc-Corp in Green Bay, Wis., for instance, announced this week that it would acquire 32 branches, including $850 million in deposits, from Huntington Bancshares in Columbus, Ohio. The deal is expected to bolster Associated’s presence in its home state, particularly outside of the state’s metropolitan areas, such as Madison and Milwaukee. After the deal closes, Associated plans to take out 45% of the costs associated with the new branches, the company said in a presentation. Those cuts will likely come in the form of branch closings. The biggest challenge facing Associated after it completes the deal, which it expects to do in early 2019, will be cross-selling customers on a new set of products, and helping them adjust to rates Associated offers on deposits and loans. “The question becomes, can they do it?” Reevey said. Flagstar, meanwhile, is simply focused on stemming the potential loss of new depositors — and holding onto the cheap source of funding it acquired. The CEO, DiNello, said that technical issues have mostly been fixed, and employees across the company, from human resources to mortgage banking, have worked together to address Flagstar’s “huge backlog” of customer calls. He acknowledged that Flagstar may lose some of the customers — and much-needed retail deposits — it acquired because of the bad first impression. The $18.7 billion-asset company had said in June that it expected deposit attrition of 17%, though privately executives had hoped to keep it to 10%. In the days since the glitch, Flagstar has kept close tabs on account closures at the new branches — and “there are more than I would like,” DiNello said, though he declined to provide details. Still, the new branches have consistently also added between 100-200 accounts a day, he added. According to local media reports, competitors in the area, including the $17.6 billion-asset Old National Bancorp, have begun offering promotions to poach dissatisfied customers. Flagstar had expanded recently at a rapid clip, boosted in part by strong growth in commercial lending. Over the past year alone, its total assets have increased by 11%. “The company has a very specific need for funding,” said Scott Siefers, an analyst with Sandler O’Neill. Just months ago, the acquisition was viewed as a milestone for Flagstar. According to Siefers, some of the deposit attrition was likely caused by customers moving out of the Midwest, to another location where Wells has branches. So even though a customer’s account may have been assigned to one of the branches in Michigan that Flagstar acquired, it would have been impractical for that customer to switch banks. Siefers said he will be keeping tabs on deposit attrition, particularly given the reports of customer frustration. “It’s just not helpful that [they’re] starting in a hole,” he said.