Imputing Growth Rates
Recall that WIND reports revenues for Products separate from Services, but the former includes both Product License sales and Run-time fees. The key to WIND's mid-term growth expectations is the growth rate of Run-time fees.
To dig this out of past performance numbers, list all of WIND's quarterly revenues along with the cost of product sales. Product sales suffers a cost of goods sold, which actually accrues only to the Product License portion of these revenues, because the cost for royalties is zero. Further, it is reasonable to assume that the cost portion of Product License revenues is relatively constant over time, or equivalently that the Gross Margin for Product License revenue is invariant over time. This means that, for any period,
Product License Revenue = Product Cost / (1 – Gross PL Margin)
Denote the unknown Gross PL Margin for the unknown Product License revenue as GPLM. Since (1-GPLM) is a constant, the percentage growth of Product License Revenue is independent of whatever value GPLM turns out to be. This means you can plug anything in for GPLM, and analyze the calculated series, using the right hand side of the equation above, for growth, and always get the same answer.
The answer turns out to be about 37%. Further, this growth estimate seems not to be changing much up or down over the years, although it was about 40% for FY 1999.
Now, suppose we somehow knew the exact, proper value for GPLM. We could then determine Run-time fees as the residual of the reported Product Sales and the calculated Product License Revenue using the right hand side of the equation above. Go ahead and guess a number for GPLM and do the calculations for Run-time fees. Now check the Run-time fees as a percentage of total revenues.
You should recall, because we discussed it ad nauseam on the thread, when WIND went on the road show for the Secondary in the summer of 1996, they indicated that Run-time fees comprised 20% of total revenues at that time. Thus, add up your estimated run-time fees for FY 1996 and make sure they total 20%. You can alter the percentage by changing GPLM. I ended up with GPLM = 85.5%, which was a very comfortable number, since the gross profit of total Product sales has increased from about 87% to 92% over the years – presumably owing to the more rapidly growing Run-time fees (remember, the Run-time fee portion of Product revenues has a Gross Profit of 100% by definition.)
Also, in the last conference call, Ron indicated that Run-time fees comprised somewhere between 20% and 40% of total revenues (which I believe went unnoticed by the thread!). Check to make sure this condition is not violated for FY 1999. I got 27.8%, which feels right.
Using GPLM = 85.5%, or whatever you calculate, we now have succeeded in dis-aggregating Product sales into its two components: Product License revenues and Run-time Fees. And we know that the former is growing at a stable 37% per annum.
Now turn your attention to Run-time fees. Fit a growth curve to this series and examine it carefully. I found the growth of Run-time fees to vary over time a great deal more than Product Licenses, but not with any clear indication of direction. Basically I found the smoothed annual growth ranged from the high fifty percentage points to the low seventy percentage, depending on how far back I started the series. Since calculated, non-smoothed, annual Run-time fee growth dipped from the 60+ percentage in FY 1997 and FY 1998 to about 50% in FY 1999, I substituted a conservative 50% growth for the much higher smoothed values.
Armed with these figures, you can now put them altogether and make the following conclusions, assuming the patterns underlying WIND's past performance continue:
1. WIND should grow revenues by about 40% again this year, if Services grows by 30%, which should be the minimum achievable given the company's recent service emphasis.
2. Run-Time fees will comprise about 30% of total revenues by the end of this year, versus 20% in FY 1996.
3. If Run-time fees surge at the smoothed rate of 63%, revenues for this year will expand by about 42%.
4. From FY 1997 on, WIND's exceptional revenue growth over 40% owed much to rapidly increasing Run-time fees.
There are a bunch of other conclusions that follow from this analysis, but only if you also have a cost model that reflects WIND's expenditure habits. I'll spare you from these details by making just a couple observations. If revenues grow as conservatively indicated (specifically, 50% growth for Run-time fees), EPS should grow in the 43% to 45% range. Should Run-time fees rebound to the 60% growth level apparently experienced before last year, EPS would grow approximately 50% per annum going forward.
Let me remind you that a major weakness of this analysis is that it cannot “see” into the future like a model based on exogenous data – unless there are telltale signs embedded in current performance figures. There are no indications in the data of anything pending beyond the observation that WIND's product mix is benefiting from increasing Run-time fees, and these fees provide increasing visibility going forward. They increase visibility because Run-time fees stem from design wins bagged months, or even years, in the past. Last year, cumulative Tornado design wins probably exceeded 8,000, a number that strengthens my belief that Run-time trends suggested by this approach will continue apace, and probably increase.
What this model cannot sense is any kind of inflection point related to a positive market niche breakout, as always possible with the likes of I2O or Internet Appliances.
Allen |
