Many successful banks prefer to sell primarily through branches. This includes some large ones, such as Wells Fargo, plus a great many community banks.

You can't apply for any credit card at Wells' web site if you're not already a customer. You are invited to come to a branch. I suppose the average number of products per customer (Wells's declared goal is eight) will be higher if fewer people start by buying just one!

For these banks, spending heavily on branches and opening more and more of them implies trying to spread branch costs over as many customers and products as possible — rather than spending incrementally on other marketing channels, when expense levels are already under pressure.

Nonetheless, many other banks are happy to sell profitably any product, for instance a credit card or wealth management, to anyone at all. Which strategy is right?

The better strategy is to market any profitable product to get a new customer-and then cross-sell relevantly at the right times, without requiring a branch relationship. This is, for example, my understanding of the Capital One and Chase strategies.

Can't sell a checking account to a card customer? Better try an installment loan first. Anyway, do you really want more checking customers now? How long will they take to become profitable?

All this is true for small banks as well as large. A community bank is a small business. Maybe in 1911, virtually all small businesses served only local markets. That's not true in 2011. Improved technology broadened the opportunities. Almost every day we can read in this newspaper about small institutions that have become innovative and more effective marketers.

Consider also some examples in non-bank financial services. Prepaid card companies started small, but not local. They arranged the retail distribution they needed. PayPal likewise started as a small company, but negotiated for national distribution. USAA grew without branches from the beginning, when it sold insurance.

Western Union closed down all its branches! But an increasing number of banks, including U.S. Bank, are Western Union agents. Markets are no longer defined by locations.

It is ironic that 9/11 and anti-money-laundering regulation have ushered in a new era of Know Your Customer. By now it should be obvious that knowing your customer doesn't mean sitting across the desk from him. To know your customer's finances, persuade him to share his Internet banking credentials. But you can easily learn what you need to know on the Internet or telephone.

The other extreme, abject ignorance, means knowing only what the customer chooses to assert-true or false, such as stated income. One large bank told me: "In our branches, we ask for income when they open a checking account — but we don't require them to answer." Much less to answer honestly. Payday lenders are more demanding.

The superficiality of most bank branch relationships is glaring, irrespective of the size of the institution. What fraction of today's community bank customers ever talk about their finances with someone in the bank? Is the customer's being able to recognize the smiling teller going to get us the mortgage — even if our rate is high?

The branch experience is remarkably similar and equally sterile in most small and large banks. The mega banks are trying to diminish the small difference in customer perception even further by handing out titles, such as city, state, or region president. Soon maybe there'll even be branch presidents.

It follows inexorably that most branches are ineffective at cross selling. The fraction of customer visits to a branch that lead to a cross sale is tiny compared even to the fraction of telephone contacts with an insurance agent or stockbroker that leads to some kind of sale.

Another reason for this is that consumer banking, much more than any kind of insurance or even securities activity, has moved to a self-service model. At ATM's and at web sites, your customer must get by without asking questions. So, he won't come into the branch and ask for advice about his savings and payments.

Institutions with no reasonable expectation of making their checking accounts adequately profitable for years to come are nonetheless mired in an antique view of checking as the hub and touchstone of relationships — to which they nonetheless are unwilling or unable to contribute any human element of continuity.

Specify your target customers, determine how to reach them most efficiently, and learn and sell relentlessly to their needs.

Andrew Kahr is a principal in Credit Builders LLC, a financial product development company, and was the founding chief executive of First Deposit, later known as Providian. He can be reached at akahr@creditbuilders.us.com.