Randy Befumo (MF Templar) tries to piece through the value equation and figure out what has been holding down the shares of America Online lately and what the heck the fourth quarter earnings have to say about the future, looking at the positive and the negatives. All this and more, kids.

Transmitted: 8/11/96

BURNED BY CHURN

Churn is a term that has become virtually synonymous with phrases like "America Online," "Steve Case" and "The stock is down 12% again!?" But what exactly is churn and why is everyone so worried about it? Churn occurs when someone who has been counted as a registered member of America Online leaves for any reason. Whether they leave because they are an old member who has just tired of the service; because they found nothing compelling in their initial free 15, 20 or 50 hours; because they said one too many fresh things in a chat room or on a message board and got TOS'd off the service; or finally, because they were using a fraudulent or stolen credit card and had to be canceled.

In order to figure out what the churn was, you need to be able to figure out how many people signed on the service and how many people canceled within the quarter. Unfortunately, America Online does not release those numbers, which leaves us number-crunching analysts to our own devices when it comes to figuring out how bad churn really is. In order to figure this out, we need to get three numbers: the net number of subscribers added, the average cost per subscriber acquisition and the total amount spent on subscriber acquisition.

THE DATA POINTS

In the last quarter, America Online reported that they had 6.2 million worldwide subscribers, 5.8 million in the U.S. and 400,000 internationally. This is up from 5.8 million worldwide subscribers in the third quarter, 5.5 million in the U.S and 300,000 abroad. Net new subscriber additions were 400,000 for the service and 300,000 in the United States. There is some confusion over how many subscriber additions that America Online had in the current quarter because during this quarter they switched from releasing U.S.-only numbers to telling people how many subscribers they had on a worldwide basis. The above numbers are the most accurate numbers I can find and jibe quite well with what was said in the third quarter and fourth quarter press releases and conference calls.

America Online tells you in the quarterly conference calls how much it cost them to acquire their subscribers on a total basis. In the just reported quarter, America Online's Chief Financial Officer (CFO) Len Leader said that America Online spent an average of $50 to $55 per subscriber. They also forecast that this would probably increase to $55 to $60 over the next year as they stepped up their marketing plans.

America Online's total amount spent acquiring subscribers was roughly $84.7 million in the last quarter, according to information allegedly released in the fourth quarter conference call. I got this number from an analyst who follows the company and cannot personally recall hearing it during the call, but frankly, you miss some things in a call when the information is being pumped out at you.

CALCULATING THE CHURN

If America Online added a net 400,000 subscribers at somewhere between $50 and $55 per subscriber, this cost them roughly $21 million:

Cost for Net Subs Added = 400,000 subs * $52.50 per sub = $21 million

Now, if we know how much America Online spent on subscriber acquisition costs in the quarter, we can get an idea how many total subscribers were acquired in the quarter. The number we have here is $84.7 million. Thus the total number of subscribers added is:

Total Number of Subs = (Total Costs/Costs for Net Subs) * Net Subs Added =

= ($84.7 mm/ $21 mm) * 400,000 = 1.61 million Total Subs Added.

This tells us that over the quarter, America Online added four subscribers for every one that it kept. This is the principle reason why the stock reacted so poorly on Friday to Thursday evening's earnings release.

COST PER SUBSCRIBER

Another number that people hear a lot about these days is cost per net subscriber added. Some call this number cost per subscriber added, but either way, this is not an accurate number. Essentially what these analysts do is take the total number spent on subscriber acquisitions and divide by the net number of subscribers added. The argument here is that this is de facto what it costs America Online to add a subscriber, an argument I think does have a lot of merit to it. However, we should label it correctly and then make this case, rather than mislabeling and causing confusing contradictions with the information America Online is putting out.

In the past quarter, America Online spent allegedly $84.7 million and added only 400,000 subscribers, meaning that they spent $211.75 for every net subscriber added. This number is not a tremendous improvement over the third quarter, as it is essentially flat, and is another factor that is weighing on the shares in post-earnings trading. Churn and cost per net acquisition are related problems, although they are not entirely the same problem. Solutions that address one almost have to address the other in some way, which is why we can safely view churn and the cost per net subscriber added as the bearish binary.

THE VALUE EQUATION

How does America Online deal with churn and the high cost of subscriber acquisition? Hanker down and pray it will go away? The battering the shares have sustained over the past few months is a measure of Wall Street's confidence that Case and Co. can handle this newfound dilemma. Any observer who has seen the company's transformation from the number three online service provider in 1994 to number one with a bullet in 1996 is probably going to be a little more open to possible solutions, however. The short-term mentality of the Street sees no easy solution in a quarter and therefore puts aside the shares for a rainy day. What is an investor to think after all of this?

Steve Case sums up the dilemma in a very tidy phrase, "the Value Equation." Essentially, the service has to be compelling to the consumer or else they are not going to stay on after they sign on. America Online's strategy to deal with this has been to improve the basic service and reduce the cost to the consumer, the kind of basic blocking and tackling than any consumer-oriented company needs to do to stay in business. Note that America Online did not announce a high-profile campaign to invest faster access or report some vaporware that would deliver full-motion video to your desktop. When confronted with the Value Equation, they improved the service and made it cost less.

AOL 3.0 AND THE $20/20 PLAN

America Online for Windows 3.0, America Online for Windows95 3.0 and America Online for Macintosh 3.0 are all out now, either in mass use or beta testing. WAOL 3.0 has had more than two million downloads and has kept the average user on for 10% longer than they otherwise would have stayed. (Buddy List alone probably accounts for 90% of that increase.) The fact that America Online delivered on its promise to the Macintosh crowd after months of hedging demonstrates a level of execution we have not seen out of Reston, Virginia in quite some time. From Buddy List to progressive art downloads, almost all of the major negatives of the service have disappeared (with the trashy pop-ups screens hawking infomercial-esque products when you sign on and the Hub being the lingering major thorns in this Fool's hide.)

Another step in the right direction is the gradual reduction of costs to heavy users, which hopefully will end one day soon in a total flat-fee for access system. The $20/20 plan has meet with pretty decent success despite the pain the switch caused many vocal subscribers. About 800,000 people are signed on and two-thirds of these were using less than $20 a month of service before they signed up, implying that the people moving up the spectrum in price are helping to pay for the heavy users who are getting a much deserved break. Is $20/20 the end of the road? Well, the average person watches television for 7.0 hours a week. If you extrapolate from this fact that the average person has 28 hours a month to view cool stuff, $20/20 ain't a flat fee but effectively fills most if not all of their need. (I mean, come on, you still have to watch X-Files, right?)

I wouldn't write off future price cuts completely, however. I have said more than once that if America Online cannot afford to offer flat fee service at some price point within two years, they have done something wrong and have not diversified their revenue stream enough. This quarter's guidance of a 50% increase in subscriber revenues for the year while the number of subscribers would increase from 6.2 million to 10.2 million subscribers does not fit together well unless Case & Co. have another price cut of some kind planned in the next 52 weeks.

GROSS MARGINS

What was in the results that should make investors excited? Well, that 5.4% increase in gross margins was nothing to sneeze at. It was well above what any analyst imagined and represented more real cash flowing to the company's bottom line. This ain't some number that gets monkeyed around with because of deferred subscriber acquisitions costs... that occurs in the sales and marketing line in the operating expenses. America Online attributed this rise in gross margins to the fact that more than 50% of its users are on its TCP-IP network, AOLNet, which dramatically improves their cost structure.

Is this latest surge in gross margins sustainable? From the conference call, the guidance suggested that it was. Len Leader said, "Expect margins in December through June to be higher than the just completed quarter," and I think that this is something we should take him at his word on. With the company shooting to have 85% of its traffic on AOLNet by the end of this fiscal year there is quite a bit of spending they can do on the operating line for marketing that suddenly does not impact the total financial structure quite as much. If America Online can move gross margins to the 55% range, a jump possible as the magnitude of people moving onto AOLNet at 85% is equal to the shift they have already seen, this has extremely positive long-term implications.

OTHER REVENUES

Anyone who has read a word I have written on America Online knows I focus on the company's paradigm shift from a fee-for-service organization to a broadcast and media powerhouse with multiple diversified revenue streams flowing into the bottom line. This quarter's results consequently gave me quite a bit to be happy about. Advertising, merchandise and transaction revenues increased sequentially over last quarter by 29%. Advertising alone was up 57% and the company sold more ad space in July than it did in the entire fourth quarter. Transactions were up 37% with the Banking and Mutual Fund Centers yet to open and America Online unwilling quite yet to see the long-term benefit of having more than one discount brokerage on the service.

These other revenues figures are key and it was gratifying to see the company perform as well as it did. These are numbers that I will continue to pay careful attention to as anyone can tell you something growing at 57% sequentially becomes a significant part of your bottom line pretty quickly. America Online guided analysts to expect the other revenues portion of the total sales to become 15% next year, up from 10% this year. This means the company plans on having these grow twice as fast as the usage revenues, no mean feat.

CASH-FLOW POSITIVE

The old saw about America Online has been the fact that it amortizes subscriber acquisition costs over a twenty-four month period, creating a future deficit that has the near-term effect of pumping up earnings. Frankly, I think this is a little overdone. All sorts of expenses are pushed out over their lifespan in GAAP and FASB approved accounting -- there are no tricks here. The sting of this potential liability is reduced greatly by Len Leader's prediction that the company will become cash-flow positive and therefore self-funding by the end of 1997.

Aren't all businesses self-funding? Actually, you would be surprised by the number of growth companies that are not cash-flow positives, deriving much need operating funds from issuing equity to partners. America Online netted $28 million this quarter when it made a deal with Mitsui and the Japanese conglomerate decided to buy some shares on the cheap. In the words of a buy-side analyst friend of mine, "If America Online becomes self-funding, it is dangerous." If the company is generating the free-cash flow to pay for its own marketing expense and no longer has to cut big deals or come to the Street for capital, a lot of the concerns about its prolifigate spending on subscriber acquisition will disappear entirely.

(ACTION-PACKED) CONCLUSION

In the end, you have the bearish binary and the bullish triad duking it out for the hearts and minds of America Online's shareholders, as the company struggles to solve the value equation and come up with a dominant consumer service that will attract millions of users every day. From almost a standing start in 1994, the company now has a billion bucks in trailing revenues, no small feat. This is a company that is quite a bit bigger than a whole raft of popular blue and red chip names and one that will continue rapid growth into the next fiscal year.

Quite honestly, the bears may dominate these shares for the next two to four quarters. It is going to take some time for us to see if the value equation will knock the bearish binary out of place and let us see the positives a bit more clearly. For someone with a time frame of longer than a year, however, these shares seem to offer uncommon value. With gross margins 5.0% higher than most analysts have been estimating and $1.5 to $2.0 billion in revenues coming next year, $1.00 EPS seems like a pretty low estimate to this Fool. Somewhere between $1.10 EPS to $1.25 EPS is where I think the Street will eventually come to rest, implying the company trades at 23 to 27 times forward earnings. At $556 per subscriber right now, we are at a trough of the company's valuation on that metric as well. Time will tell whether this Fool is fair or shill, though.