In Defense Of Bank Branches

Subscribe Now!

Stay on top of all the latest news and trends in banking industry.


I got this from a CNBC article on bank branches:

“Mobile transactions are easier for customers and cheaper for banks to service, according to Diebold, a company which specializes in ATM and branch transaction services. In the company’s 2010 investor presentation, it estimated a $4.25 per transaction expense at a bank branch versus only 8 cents through mobile banking. A 2013 Deloitte study found 40 percent of consumers were willing to pay more for the ease of mobile banking, too.”

My take: There’s so much wrong with that paragraph — both stated and implied — it’s hard to know where to begin.


So let’s begin with the first sentence. Why would mobile transactions be “easier” for customers? Because they don’t have to go to into a branch to conduct the transaction? What if the transaction (or interaction) requires some discussion or involves some level of complexity?

If we’re talking about checking the balance on a account, or transferring funds between accounts, then sure, a mobile transaction may be easier for a customer to do than doing it other channels or through other methods. But the blanket statement “mobile transactions are easier for customers” doesn’t hold water. Unless, of course, you assume that the only transactions that exist are those that more easily done through a mobile device.


The implication of the second sentence — which states that branch transactions cost an average of $4.25 per transaction vs. $0.08 per transaction for mobile transactions — is that shifting transaction volume out of branches and into the mobile channel will result in huge cost savings for banks and credit unions.

Won’t happen. Not unless you shut down a large number of branches, which is a whole lot easier said than done. In addition, these cost estimates are terribly misleading. They are not variable costs. The branch does not start the day with $0.00 in costs and add $4.25 (on average) every time someone comes in to conduct a transaction.

The CNBC article quotes Brett King as saying “Customers, on average, visit a branch 85% less than they did in 1995.” Assuming that the branch transaction volume declined by 85%, then a driver of the supposedly high transaction costs in a branch is the fact that the volume of transactions in insufficient relative to the cost of operating the channel.


And if you do shut down branches, there might be negative side effects. Again, from the CNBC article:

“Even in the face of real estate and transaction costs, bank branches are a critical tool to attract new customers—if only serving as expensive billboards for the company in a choice-heavy world. “It’s going to be very difficult to convince people…that you’re a major presence in a market and you’re here to serve them if you don’t have any physical presence,” said Jonathan Larsen, Citigroup’s global head of retail banking.”

This really gets to the problem of the channel costs that people throw around. The $0.08 mobile channel transaction likely produces no revenue, while the $4.25 transaction might.

It’s akin to why I want to slap people who think direct mail is dead upside their heads. It’s about ROI. If a <1% response rate produces $1 million in revenue for a $10k investment in direct mail, and a 10% response rate in another channel (e.g., social media) produces $10k in revenue for a $1k investment, the larger response rate doesn’t matter. Sure, the social media campaign cost less, but the direct mail campaign produced more revenue.

It’s the same with channel costs. Looking at average transaction costs ignores the composition of those transactions. Smart managers don’t ignore the transaction composition.


My guess is that Apple could save a lot of money by cramming its products into much smaller stores, and locate those stores in the seedy sections of the cities where they do business. In fact, they could just shut down those stores, and take all product orders online. I’m sure the company could develop a mobile app to merchandise products and take orders.

Yet it doesn’t, and everybody with a Twitter account falls over each other to tell the world how great their Apple store experience was.

Nobody brags about their bank branch experiences, though (except for my dad).

The real problem with bank branches isn’t a higher cost per transaction. It’s a two-fold problem: 1) transaction composition is (still) skewed too much towards service (vs. sales) transactions, and 2) those sales transactions suck.

OK, that last point (#2) was unfair and unsubstantiated.

But the fact of the matter is that many of the sales-related transactions that occur in branches are conducted by employees unqualified to help consumers make smart decisions about their financial lives. And the information that the employers — banks and credit unions — provide to those employees to help conduct those sales transactions is woefully lacking.


There are industry participants and observers who think that branches will become places where consumers will go to discuss their financial needs and lives, and become more sales-oriented than service-oriented. Others think branches are dead (or rapidly dying) and have no shortage of data to prove their point. 

I think the debate is stupid. There’s no reason why any particular bank or credit union couldn’t go branchless. And there’s no reason why any particular bank or credit union couldn’t make their branches the equivalent of an Apple store. 

It’s not a matter of whether or not branches are a good idea (or not), or whether they’re alive or dead — it’s a matter of execution. It’s about having a commitment to making the branch work (or getting rid of them), and understanding the inconsistent and conflicting decisions that so many banks and credit unions make that undermine channel strategies.

For related posts, see:

What To Do About Bank Branches

Distorted Visions Of The Branch Of The Future

This article was originally published on . All content © 2022 by The Financial Brand and may not be reproduced by any means without permission.