First of all, what you are about to read, I did not write.
It is an analysis of todays net leaders and why many investors "missed the train" BECAUSE of doing such thorough DD.
In your reading please see if you can see the parallels between these net sector leaders, hardly known a scant few months ago, and our diamond in the rough, PNLK.
This is a long piece and I believe you would be best served to print it out to read at your leisure.
Rich
Margin Of Error
The cost of a mistake is almost always equal to the magnitude of its opportunity. And while any choice sets limits by closing off alternative paths, it may also be possible to miss the only road that leads to a true and happy life. How does this happen? And what cruel fate allows one to act with such incomplete information? We have no idea, since we blithely assume we haven't forever lost that glorious road. But, just in case our current comfort suggests the proximity of a brick wall, we harbor a faith in the possibility of redemption, the ecstasy of second chances.
For the reluctant investor in Internet companies, second chances arrive every day. But habits of analysis are often the hardest to break, since they reflect, by definition, much of one's world view. So the same incomplete logic that underlies the rejection of Yahoo! solely because of extreme revenue multiples also compels a dismissal of America
Online, Amazon.com, and DoubleClick because of relatively low gross margins. Don't arch that eyebrow (especially you investors in twenty percent gross margin hardware companies), we're hearing this low gross margin
schtick a lot recently. And what it misses, besides the other fifty-seven value drivers of these stories, are the gross margin expansion strategies that three of our favorite Internet content and commerce companies are currently executing.
When AOL reports earnings this Wednesday, May 6, the gross margin lines should be the center of attention. We expect every other element of the March quarter to be at or above our model, but the ever-expanding usage of AOL by consumers could drive network costs (which account for 70 percent of the cost of goods sold line) over the $447 million we're expecting, bringing the gross margin below our projected 35.5 percent. With total AOL network usage now exceeding 240 million hours per month, and with 50 cents of every dollar spent by AOL in a given period
allocated to telecommunications costs, any efficiencies gained in providing access to its service has a tremendous earnings impact.
That's where the WorldCom deal comes in. You'll recall that WorldCom, as part of the deal that brought CompuServe into the AOL family of brands (it currently plays the role of Uncle Fester), agreed to provide AOL with incremental network capacity at a declining cost per hour over the next five years. Since that deal takes effect in the June quarter, we look for the March quarter gross margin to be a low watermark. And the rising river of network savings could ultimately be a Mississippi. In fiscal 1999 alone, we look for over half of our 600 basis point gross margin expansion to be directly attributable to WorldCom (most of the rest is from a revenue mix shifting toward more high margin non-subscription revenue). It's not hard to see why when you make an educated guess at the underlying terms of the deal: by the end of next year, we look for the blended cost of a marginal hour of network usage to drop from 47 cents to 33 cents per hour, an improvement of 30 percent. Ultimately, we think 50 percent
efficiency gains are possible, with an initial minimum of 40 percent of the network's traffic benefiting from this lower rate. So long, gross margin overhang.
A low gross margin is also one of the first reasons investors choose not to jump on the Amazon.com train, despite the ongoing positive trend of this line revealed by even a quick look at the company's financials. Amazon generated a 19.5 percent gross margin for all of 1997 and then proceeded to expand that by 250 basis points in the recently reported March quarter. The dominant online bookseller (and likely to be masterful online retailer) is now selectively stocking inventory ($11.7 million worth at the end of March, versus $2.7 million two quarters ago). Sourcing product directly from publishers costs 10 to 20 percent less than from distributors, and that kind of improvement in the primary element of cost of goods is huge, even if only a fraction of sales are from in-house inventory.
While critics suggest that this inventory strategy is fraught with uncertainties, Amazon has a decided advantage over its land-based competitors. Amazon knows precisely what types of books sell, it knows exactly what its customers buy, and it knows how the customers get to the point of purchase within the Amazon site. As a result, it knows
what books to purchase, it can more accurately estimate the quantity, it can provide real-time flexible pricing, and it knows how to promote which books to which customers. Thus, it can order directly from publishers with a greater degree of confidence than Barnes & Noble or Borders, and it turns inventory four times as quickly. Additionally, Barnes & Noble buys its books on consignment, returning unsold inventory. This practice, though favorable for B&N, affects the pricing policies of the publishers, as does the additional marketing charges B&N extracts (amenities such as book-jacket-facing-out are paid for by the publishers). The upside of search-based and subject-linked online retailing is that it is far less necessary for a publisher to pay for these amenities. Again, this affects the price that Amazon receives, which is mostly what drives the gross margin line.
Having walked through these book selling details, we can add up the numbers. Factoring in Amazon's estimated average price discount of 20%, we project that the 60% of the suggested retail price Amazon pays to a distributor results in a real cost of goods sold to Amazon of 75%, on top of which we add telecommunications costs, a portion of customer service, and the marginally profitable shipping and handling components. The net result on distributor-supplied merchandise is a 20% gross margin. Inventoried merchandise, however, should bring Amazon's cost to,
at most, 50% of the suggested retail price. Factor in a 25% discount from list price (got to move that inventory) as well as the 5% incremental telecommunications and customer service costs, and the realized gross margin comes in at just under 30%. Nice 800-900 basis point improvement. As Amazon continues to get more aggressive with its
inventory strategy, we can see the blended gross margin for the book business climbing above the 25% range. As is the case with Yahoo! Et al, that incremental gross profit emerges against a marginally (in the economic sense) discretionary set of operating expenses, dominated by the brand marketing line. Can you say earnings power? And we haven't even talked about digital distribution, especially in likely future vertical categories for Amazon such as music and videos. We won't try to walk you through the math on this variable, especially since its years away. But we're not insane to suggest a gross margin expansion of world historical proportions for Amazon when online delivery becomes more feasible. Stay tuned.
Finally, our favorite small-cap name, DoubleClick, has a financial model that provides clear visibility on gross margin expansion. The DoubleClick Network, which is the advertising sales component of the DoubleClick model, currently accounts for 95% of DoubleClick's revenue. The DoubleClick ad sales force represents a variety of sites in a variety of categories and sells that traffic to advertisers. These sites range from USA Today to PBS, with search site Alta Vista accounting for over 50% of DoubleClick's traffic and revenue. DoubleClick books 100% of Network revenue and pays between 60% and 70% of those sales back to its sites, with Alta Vista at the high end of this range. Despite what you may have heard, ad sales is still a seasonal business; it's just that in the Internet realm during Q1, search engine advertising proved to be the least seasonal. But in the second half of the year, within the DoubleClick Network, we expect non-search engine advertising to account for more than half of Network revenues, which drives down the amount paid back to Alta Vista as a percent of total revenue. This should expand the blended gross margin by 100 basis points (from 32% in Q1) over the course of 1998, and 300 basis points over the next three years.
More important for DoubleClick's gross margin is the revenue mix between the Network and the "other" DoubleClick revenues, comprised primarily of DART, the DoubleClick ad-serving service. DoubleClick charges an
average of $1.00 per 1000 impressions served, which costs DoubleClick approximately $0.20 per 1000 ads served (an 80% gross margin). While we like the growth characteristics of the Network, we see multiple growth drivers for the DART business, including both organic impression growth per existing DART site and incremental impression growth from added DART sites. The total number of DART customers doubled in the March quarter, and over the next three years we see DART revenue going from under 5% of total DoubleClick revenue to over 20%. The combination of expanded Network margins and increased higher-margin DART sales leads to a long-term gross margin approaching 40% (we see a pattern of 800 basis-point gross margin gains forming here). And, in the wonderful world of scalability, there are nominal incremental operating expenses associated with the DART revenue. Once again, we hear the sound of an Internet advertising model making the happy noise of a cash register
(which is "cha-ching", for those of you who send nice e-mails when we spell things out).
At this point, we normally end with a short soliloquy that hammers home the main thesis we've been promulgating for the last few pages (which, in case you missed it, is that gross margin expansion is good; it's unspoken but critical corollary is that it commands a premium valuation). |
