Description

Introduction

Many years ago, I was riding along a dirt road in the Cascade mountains with my father-in-law as we traveled to a trailhead for a backpacking trip.  The road followed a small river, and we periodically passed signs along the bank that read "Registered Mining Claim."  Wondering if there were any wealthy Pacific Northwest families whose legacies had sprung from such claims, I asked my father-in-law if anyone had gotten really rich during the gold rushes of old.  He replied, "Yea, the guys who sold the picks and shovels."  Since then, I have discovered that some of capitalism's beneficiaries are companies like ANSS who sell the "picks and shovels" of innovation-the tools that let companies in competitive fields improve their products in the hopes of defending or growing their marketshare.  ANSS' tool is simulation software that allows engineers to perform tests on virtual prototypes.  This approach to product development has major advantages over traditional methods, and I believe that it is relatively early in its adoption cycle. 

Background

In the past, engineers developed new products by using Computer Assisted Design software (CAD) to create digital designs of their new products.  From there, they sent their designs to a service bureau who constructed a physical prototype and delivered it back to the engineering department.  The engineering department then ran the prototype through various tests to uncover flaws in the design and reported the results back to the design engineers.  The design engineers then began making changes to their original designs, and the cycle started all over again.  While effective, this was an very time consuming process since each prototype took a long time to make, and some companies had to go through several dozen prototypes before settling on a final design.  An indirect consequence of this approach was that it limited the number of design alternatives that engineers could consider since each iteration took so long to test.  In many ways, it had the same effect on engineers that typewriters had on writers.  Reworking a couple of sentences with a typewriter might necessitate retyping entire pages, and this discouraged writers from experimenting with different ways of rephrasing their thoughts.

ANSS makes simulation software that converts the digital design from the CAD software into a virtual prototype and then conducts simulations on this virtual prototype.  This dramatically accelerates the feedback loop between testing a design and reworking the design.  While the simulations are not accurate enough to dispense with physical prototyping altogether, they dramatically reduce the number of physical prototypes that are required.  I have heard anecdotally that virtual prototyping has allowed John Deere to use 75% fewer physical prototypes than they had used previously. 

I have heard of numerous other examples of how computer simulation accelerates the product development process.  In one case, an automotive engineer needed to model the flow of air conditioning through an minivan.  Using old methods, this required 30 days, but with computer simulation, this task was reduced to two days.  In another example, an automotive engineer needed to design a new gas gauge for gas tanks.  This was previously a year long project, but with ANSS software, it was reduced to eight weeks.  A third example involved a jet engine manufacturer who was trying to optimize the way they manufactured turbines.  Traditional methods had taken two weeks to complete this task, but with ANSS software, they were able to compress it down to two hours. 

There are four major benefits from realizing these type of productivity gains.  The first is that it improves "time to market," delivering new products faster.  The second is that companies discover many more design flaws during the design stage.  This provides considerable cost savings because designs become much more expensive to change once they enter production than when they are only on paper.  This is because design changes during the production stage often require new tooling or rearranged processes.  A third benefit is that simulation early in the design process can help optimize a product's unit cost.  Studies have found that 80% of a product's cost are locked in during the first 25% of the design process.  Consequently, being able to test basic design parameters early in the design process helps companies optimize unit costs and ensure that they can meet their cost targets (i.e. value engineering).  The final benefit is that simulation allows engineers to consider more design alternatives than in the past.  This increases the amount of innovation that they can introduce into their new products.

While this technology was originally focused on product development for manufacturers, it has many other applications.  I heard about a retailer who used ANSS' software to simulate the flow of air conditioning through their datacenter in order to optimize the datacenter's configuration and thus its energy use.  Several years ago, I saw a simulation of a heart valve from a medical device company.  Several months ago, ANSS put out a number of press releases describing how Speedo had used ANSS software to develop a new swimsuit used extensively at the Beijing Olympics.  I've also read about how medical researchers studying aneurisms are using ANSS' technology to translate medical images such as CT scans into simulations that predict how likely an individual patient's aneurism is to rupture.  The technology can also be used for marketing purposes.  I heard an anecdote about a wheel manufacturer who used simulation to demonstrate how their design would meet their customer's requirements as well as why their competitor's design was inferior.  The ability to grow into new applications such as these is one of the exciting aspects of this technology.

History of Simulation Technology

Simulation software has been around for decades, but for most of its history, it was only accessible to a small group of users.  Very expensive computers were required to run the simulations, and users needed PhD's to understand the underlying math and conduct the simulations correctly.  It was similar to the computer industry the 1970's.  Computers were available, but they were expensive, and using them required learning complex commands that were unintuitive to most people.  (Remember "C://startup.exe"?)  These financial and accessibility constraints limited computers to a relatively small group of users.  By the late 1980's, however, computers had become much cheaper and Windows had made them accessible to a much broader group of people.  After that, PC usage mushroomed. 

A similar trend is underway with simulation software.  Over the last several years, high performance computing has become much cheaper.  More importantly, ANSS has made the software much more accessible.  They have essentially child-proofed the software so you don't need a PhD to use it.  They have also automated parts of the workflow so simulations take less time to perform.  Lastly, they have made the simulations more realistic which makes them more useful.  In the beginning, a given simulation package addressed only one type of physics (i.e. mechanical simulation, fluid simulation, thermal simulation, etc.).  The real world, however, is "multi-physics" with many of these phenomena playing off one another.  ANSS began in mechanical simulation but steadily acquired companies with expertise in other areas such as fluid-structure interaction, and most recently, electronic and electromagnetic simulation.  Through these acquisitions, it has assembled the expertise and products to provide "multi-physics" simulations.  These trends of greater accessibility, and higher fidelity simulations have led to growing adoption of simulation software among the engineering community.  Based on management's comments and industry commentary, I've gotten the sense that we are still relatively early in this technology's adoption curve which means that ANSS probably has many years of double digit revenue growth ahead of it.

Financial Model

One unique element of ANSS' model is their high degree of recurring revenue.  Roughly 33% of their revenue comes from paid up licenses, 33% from leased licenses, 30% from maintenance, and 4% from services.  Consequently, about 63% of their revenue recurs every quarter.  In addition, as a software company, ANSS enjoys a very high degree of operating leverage.  Realizing this leverage is sometimes deferred in the year following a large acquisition as they tightly integrate the acquiree's sales force and technology, but there is a solid trend of margin expansion over time.

It's worth mentioning a couple other aspects of ANSS' financial results.  Having acquired a number of other software companies over the last several years, ANSS has built up considerable intangibles on its balance sheet.  The associated amortization, as well as stock based compensation result in FCF that is much higher than GAAP Net Income.  The acquisitions also create unique revenue dynamics.  Under EITF 01-3, ANSS is not allowed to report revenue that comes from an acquisition's deferred revenue balance even though ANSS gets the cash associated with that revenue.  After an acquisition, ANSS' GAAP revenue thus understates their sales in the marketplace.  As those leased licenses or maintenance renew in the following year, ANSS is allowed to begin recognizing this revenue which creates an additional revenue lift that begins 12 months after the acquisition.  It's thus necessary to make adjustments for this unreported revenue when estimating organic revenue growth in the second year following an acquisition.  A summary of ANSS' historical results is shown below:

Why The Stock Has Sold Off

ANSS reported their 4Q08 results in late February.  Excluding a recent acquisition, ANSS grew 5% on a tough 4Q07 comparison.  If you also exclude the f/x headwind, revenue grew 9%.  This was admittedly a sharp deceleration from prior quarters when ANSS was growing 18-22% organically. 

What I think unsettled the market, however, was that ANSS' sharply revised its 2009 guidance from the figures previously given in November.  This was particularly disappointing since ANSS systematically low-balls their guidance and has thus met or exceeded expectations for many years.  I have summarized the magnitude of their guidance changes below:

(Non-GAAP revenue includes the deferred revenue from acquisitions.  Non-GAAP EPS includes income associated with the deferred revenue from acquisitions, and it excludes amortization of acquisition related intangibles and stock based compensation.) 

As you can see, management reduced the low end of their revenue range much more than the high end, and as a software company, this is amplified on the bottom line.  The magnitude of their guidance revisions, however, was somewhat incongruent with the tenor of the call.  Management acknowledged economic weakness a number of times, but they also repeatedly talked about how demand for their products remains healthy because many companies want to maintain their investments in new product development.  They also said that the sales cycle has stretched and that some companies are waiting to see how things play out, but they were positive overall about the company's prospects. 

As I began to develop my 2009 estimate, it became difficult to see how 2009 EPS could come in close to the low end of EPS guidance.  Getting to the low end requires assuming that revenue drops 9% organically (it increased 3% and 4% organically in 2002 and 2003, respectively) and that there isn't a modicum of expense control.  This just seems implausible for a company with 63% recurring revenue and ongoing interest from customers.  I've assumed a 2.6% decline in organic revenue and expenses that are consistent with management's expense guidance.  (This 2.6% organic revenue decline includes 4-5% of currency headwinds.)  These assumptions imply an EPS figure much closer to the high end of guidance than the low end.

It's hard to make sense of managements' mixed signals, but I believe they lowered the high end of guidance to reflect actual economic weakness and slashed the low end of guidance as a precautionary measure to manage expectations.  This would be consistent with their dependable pattern of low-balling.  In order to low-ball-adjust my own estimates, I've kept records of how ANSS' actual results have come in over time, relative to their initial guidance.  This is shown below:

I admit that my conclusion may sound Pollyanna-ish, but it's the only way I can reconcile their guidance with their verbal commentary and 4Q08's results.  The market, by contrast, appears to be seizing on the lower guidance as evidence that ANSS' business trends are falling apart, completely ignoring the historical resiliency of their business model and the long-standing trends of broadening adoption.  This pessimism has opened the door to acquire an exceptional franchise with years of growth ahead of it for 10x FCF.  I've included ANSS' historical multiples below, but these include low multiples in the early years due to depressed small cap valuations at that time as well as the more limited size of ANSS market when the technology was less accessible.

Conclusion

I believe that ANSS' stock will recover over the course of the year as the low end of guidance proves to be far too conservative.  Even if I'm wrong, however, the stock is currently priced for very weak results which suggests that the downside is limited. 

Catalyst

Basic modeling and past experience suggest that the revised guidance given on 4Q08 CC will prove unduely conservative.  Great growth franchise at 10x FCF.