More on Net.B@nk (Bore Port) June 11, 1999

<THE BORING PORTFOLIO>
More on Net.B@nk
What the company is doing right

By Dale Wettlaufer (TMF Ralegh)

ALEXANDRIA, VA (June 11, 1999) -- Today and Monday, my Boring Port duties have carried over into the Fool on the Hill (FOTH) column. Given the discussion Net.B@nk (Nasdaq: NTBK) has generated, it was really on my mind yesterday as I was thinking of what to write about for my regular FOTH. So please excuse my recycling of Evening News content in this forum. As long as we were on the subject of Net.B@nk, I thought it would be fine with everyone. What follows is exactly what you will find in the Evening News tonight. Here's the link to part one from yesterday. Part two follows below.

Continuing my look at Net.B@nk, there are a number of things the company is doing right. First, the growth of the company and the per-customer acquisition costs the company has achieved are both very good when you compare Net.B@nk to other online financial services businesses.

Second, when you consider that growing pains are the result of the company's extremely rapid growth, that's a whole lot better than a company that has seen some service glitches as the result of its own ineptitude.

Third, the company realized early on that the Internet is a great medium for the financial services business, insofar as it makes things easier for customers to handle their finances, it makes it easy for the company to service customers and offer enhanced products and services, and it can be very inexpensive on an average cost per transaction basis to serve customers.

Financial services analysts Jeff Runnfeldt and Joe Morford at First Security Van Kasper recently cited Booz Allen & Hamilton research that show the following costs per transaction for the banking industry:

• Internet: $0.01
• ATM: $0.27
• Automated call center: $0.44
• Call center personnel: $0.85
• Branch: $1.07

In their report, which rated both Net.B@nk and Telebanc Financial (Nasdaq: TBFC) "sell" when they were priced at $59 11/16 and $46 1/8, respectively (both adjusted for recent stock splits), Runnfeldt and Morford quite rightly questioned the practicality of those data. My impressions on the matter are this: Not every transaction between an online bank and its customers is going to be between a server and the customer. Even if they were AND you had a very low-maintenance customer base, the $0.01 per transaction datum is still hard to believe.

In this business, you're going to have inbound and outbound call center representatives 24 hours a day to service accounts. You're also going to have to deal with customers by e-mail, where a customer service rep's salary and support costs are at least $0.25 per minute. If you're judging the marginal cost of processing an Internet customer at a very large bank and then adding that to the total per-transaction cost at such a bank, then you might come up with $0.01 per Internet transaction.

But let's just take customer service expenses last quarter at Net.B@nk. That was $593,897 with an average of 21,021 accounts. Ascribing those overhead costs and only 40% of Q1 data processing to the cost of executing customer transactions (which is clearly generous if you look at the entire overhead structure), and assuming $0.01 per transaction, then each customer would have to generate 3,261 transactions per quarter for that $0.01 per transaction idea to be in the neighborhood.

Sure, there are flaws in that methodology, such as the fact that the average cost for a service business such as this drops when the customer base scales up and is more seasoned. But I think modeling a company at maturity based on the Booz Allen & Hamilton datum would probably be a little risky.

Why have I concentrated so much on the customer service aspect of the company? I talked about that yesterday, but there are a number of additional reasons that are central to the investment thesis here. Included in that is the hypothesis that the benefits of spending $50+ per customer to acquire the customer will encompass more than just the revenues and profits from that individual customer. When you spend $50 per customer, you're hoping that they will spread the good word to their friends and associates about their positive experiences with Net.B@nk. If the customer is pleased, then you have the chance at leveraging that $50 investment. If you churn the customer base, then your marginal customer acquisition cost never goes down and you don't reach the scale you need to get your average transaction costs down.

Before people start to lose it here, let me just say that I don't pick a single point in my treatment of a company and then argue that. I try to look at a distribution of possibilities and go from there. If you're thinking about Net.B@nk and assuming 100% that per-transaction costs do get to $0.01, I think you're thinking about it wrong. I also don't assume that the company cannot reach its low-cost business model goal. I'm trying to handicap what I think is most probable and still count in the distribution of probabilities on both sides of the mean assumption. In other words, I think I have an open mind on the issue here.

Even at a relative stage of immaturity, noninterest expenses as a percentage of average earning assets are surprisingly low (I've annualized Q1 results in the following data so the rapid growth of the company distorts the conclusions less. There's no way to completely separate discretionary operating expense elections from required, maintenance-like operating expenditure decisions). From what I can see, noninterest expenses as a percentage of average earning assets are about 1.9%, far below the 5-6% of the large commercial banks, the 4.4% at smaller mid-size banks, and below the neighborhood of 3%+ at the larger fiduciary-specialist banks.

Also, let's say 75% of marketing is discretionary and 25% is maintenance marketing. At that rate, noninterest expenses as a percentage of average earning assets would be 1.6%. Efficiency ratios (noninterest expenses as a percentage of revenues) in those two scenarios would be 66.2% and 57.2%, which is very good for, as I said, a relatively immature company. 1998 noninterest expenses as a percentage of average earning assets were 2.2%, according to the 10-K. I'm also surprised that the net interest margin is not as compressed as I thought it would be. This has been one of the more interesting findings so far.

Unfortunately, I'm very interested in banking, so I tend to really run on about the subject and use up all my space very quickly. Therefore, I'll have to extend this discussion one more day at least. See you on Monday and feel free to send me your comments on the message board where you can pretty much always find me.

Have a good weekend, and go Sabres.

Make a Living Foolin' Around.

Stock Change   Bid
APCC  +  1/16  19.38
BRKb  -9       2322.00
CSL   +  15/16 48.44
GTW   +2 7/16  61.25

                  Day     Month   Year  History
        BORING   +0.50%   0.55%   5.10%  41.12%
        S&P:     -0.70%  -0.62%   5.57% 115.48%
        NASDAQ:  -1.48%  -0.92%  11.64% 135.16%

    Rec'd   #  Security     In At       Now    Change
  8/13/96  200 Carlisle C    26.32     48.44    84.00%
  4/20/99  460 American P    14.48     19.38    33.84%
 12/31/98   12 Berkshire   2276.17   2322.00     2.01%
   2/9/99  100 Gateway 20    72.38     61.25   -15.37%


    Rec'd   #  Security     In At     Value    Change
  8/13/96  200 Carlisle C  5264.99   9687.50  $4422.51
  4/20/99  460 American P  6659.25   8912.50  $2253.25
 12/31/98   12 Berkshire  27314.00  27864.00   $550.00
   2/9/99  100 Gateway 20  7237.50   6125.00 -$1112.50


                             CASH  $17972.40
                            TOTAL  $70561.40