Defining the Virtual Branch: How hard can it be?

In my last blog post, I raised the issue of branch accounting and the associated profitability. If you haven’t already read that blog post, you should stop now and read it before continuing. You can read How important is your Virtual Branch here. Unless there is specific profitability assigned to the virtual branch, how can you know how much investment in new technology is warranted? This question raises the issue of how accounts are assigned to a particular branch and whether all of the associated income and expense of that customer are allocated to it. Along these lines, I had illuminating conversations recently with two CEOs regarding how to account for the virtual branch.

The two FIs are very different (one is a bank, the other a credit union, one is in the Southeast, the other is in the Midwest), yet both have the same dilemma. They are both trying to determine the appropriate way to account for the impact of decreased branch activity with a corresponding rise in online activity. So, if a customer rarely comes into a branch, should their overall relationship be assigned to a particular branch just because that is where they first opened an account?

This is where it gets interesting. Both of the bankers I spoke with had tools available to them to calculate virtual branch profitability. One FI used a tool that allowed specific transaction types to be associated with customer/member activity, including the origin of the activity. That allows them to track where all transactions are originated. Using that data point, they chose 65% as the percentage of virtual transactions would constitute a virtual branch customer. Meaning that if 65% of a customer’s transactions were originated outside of any physical branch, then the entire customer relationship, all income and expense associated with that customer, would be assigned to a virtual branch G/L. The tool performs the calculations every month and if a customer has more than 65% of the targeted transactions originate from a specific channel (ie: physical branch, call center, online) in 3 successive months, then that relationship is moved to that branch on the G/L. The effect is that if a customer/member opened accounts at the Pine Street branch but was performing more than 65% of their transactions online for 3 successive months, they would be moved, income and expense, to the virtual branch.

A review of the branch data for that institution revealed that (surprise!) the virtual branch is significantly more profitable than any of their physical branches. This makes sense as there are huge costs associated with operating physical branches. This also aligns with nationally published data from industry analysts from Monitor, Jupiter and javelin that says that online customers represent significant cost savings. However, it’s one thing to intellectualize the information from a national survey and quite another to see an institution’s own G/L data that confirm the profitability of the virtual branch. For that CEO, making a decision to invest in the virtual branch is easier since they can see what the total profitability of the online branch is relative to the expense of maintaining the other physical locations.

And yet, it’s not a perfect system. If in the example above, the customer used the Pine Street branch 34% of the time, and the rest was online activity, there is still cost associated with that activity at the Pine Street branch that is not offset by any relationship revenue. This is where the idea from the other CEO I spoke with comes into play. Their institution is implementing the ability to examine all relevant transactions, determine where the percentage of transactions are allocated across channels and then apply those percentages of revenue to each location. For example, if the activity breakdown for a particular relationship is 58% online, 30% Pine Street branch and 12% Oak Street branch, then the total revenue associated with that relationship could be allocated to the three locations according to the actual usage. This would yield a highly David Peterson Chief Innovation Officer accurate picture of total profitability by location, as each location has specific costs associated with its operation. The problem lies in the effort to do this accounting and have it impact the actual general ledger. Turns out: it’s a lot of manual work.

And even if we could get the accounting allocated automatically, it still not perfect. Expense is not 100% allocated to a channel. The call center serves the online customers but also provides services to branches as well. How should those costs be allocated? If the actual revenue percentages cannot be easily allocated, then what is the amount of transactions that governs what channel gets the relationship? Is it 65% as the example illustrated above? Or just 50% plus 1 transaction? Is it just the accounts associated with the primary individual or all of the related accounts? These are important questions that will not universally fit all FIs in all situations. But I do think this is a discussion worth having.

Regardless of how you calculate it, making a serious attempt to define a virtual branch, however imperfect it may be is still heads and tails better than doing nothing. Because if you don’t figure out profitability of your virtual branch, you are not just limiting its success, you are overstating the profitability of all of your physical branches. Which may cause you to over invest in a physical branch while significantly shorting the virtual branch. Every institution is unique, if you are an FNBB customer FI and have an interest in exploring this topic with me and other interested institutions, shoot me an email to Or call me at 225-231-5032. I would love to listen to your thoughts on how to properly account for the virtual branch. Ultimately, it is possible to create a framework for evaluating investments in the virtual branch and definitively answer the ROI question.