Surplus Value     Time of Turbulence

Information Letter 12

This Information Letter was written for the first Founders' Meeting. The Founders' Meeting was held at Dan Drake's house in Oakland not long after the public offering had been accomplished. It was an opportunity to review the path we had traveled and the challenges we might face next. I wrote this the night before the meeting, racing down to the Marinship office at midnight in search of references in my bookcase there.
Autodesk, Inc.
Information Letter # 12

by John Walker
Revision 2 — Moon Day, 1985

I resolved to make such an attempt at ‘clarification’, fully realising that it would increase the size of the pamphlet and delay its publication; I saw no other way of meeting my pledge I had made in the article “Where To Begin”. Thus, to the apologies for the delay, I must add others for the serious literary shortcomings of the pamphlet. I had to work in great haste, with frequent interruptions by a variety of other tasks.

— V. I. Lenin, What Is To Be Done, 1902

We did it.

In January, 1982 we got together and decided to build a software company which would become an industry leader. We agreed that our goal was to build a large, conventional, tightly-coupled company which provided all the services needed to become an industry leader. We all committed a major component of our time, and put at risk a substantial portion of our financial assets.

Today, Autodesk is one of the leading software companies in the world. Our goal was to build a company which would be one of the top five. By Mike Ford's analysis of the SoftLetter 100, discounting game companies and people who have collapsed since the list was published, we are about number seven today. Autodesk has joined the elite world of public companies, placing it with Lotus, Ashton-Tate, Software Publishing, and Micro-Pro in the top 5 visible players in the microcomputer software industry as seen by the financial world. (And don't discount the value of this: a public company gets press coverage as a matter of course that a private company can't buy at any price. Also, public companies are perceived as more solid citizens with more staying power in a competitive situation.)

We agreed to do it right and do it fast. We did both. Name the companies which have moved from start up to public companies in three years. That's a pretty select list to start with. Now look at the ones who have done it with no venture capital, with the original founders still in control, and with not even an outside director at the time of the public offering. You're down to a pretty damn short list. Now filter for the companies with their principles intact: who still believe and practice consistently rewarding the people who do the work, of getting the best people and cutting them in on the pie in a real sense.

So what we've pulled off here is, if not unique, awfully rare in the contemporary business world. This is a good time to reflect on what we've done and to look at how we can best apply the techniques that got us this far to the difficult task in getting to the next plateau.


Because it's not over, folks! The process of building a company and reaping the rewards of our collective efforts is something I look at more and more as an ongoing brutal winnowing process. Three quarters of all start-up companies fail within the first two years. Only about one in ten thousand companies reaches the stage of making a public offering. And, yes, most public companies languish in the ranks of the NASDAQ Bid & Ask tables where we've taken up residence, rather than becoming the shooting stars who are perceived as the movers and shakers of the industry.

There's a phrase politicians use that I detest. I translate it as “we don't have any idea what the hell to do about this, and things aren't going to get any better”. The phrase is “redouble our efforts”.[Footnote] I think that this phrase is only used by people who have never in their lives ever doubled their efforts in the first place. I assume that everybody in the founding group of this company is currently working flat-out. What we need to do is continue this, at a sustainable pace, through the next stage in the company's development.

The public offering purchased effective immortality for the company. We now have over ten million dollars in the bank. This means that if our sales went to zero, we could survive for over a year, without any cut-backs or layoffs at our present expenditure level. Given the retrenchments we would make in should such a dire and unlikely (though certainly not unprecedented) scenario eventuate, we could cut back to the core group and spend five or ten years figuring out what to do next (and defending ourselves against shareholder suits). Hell, we could pay a reasonable group reasonable salaries just from the interest on the cash we raised in the offering.

To put the company's liquidity into perspective, I'm sure you remember when we all ponied up our $1 per share to buy Autodesk stock, accompanied by my incessant bleating about how we could run it to the moon. Well, adjusting for the two stock splits we've done, that dollar per share works out to six and two thirds cents per current share, $0.06666. Now today, the company has on the order of $2 in cash for each of those shares. So if we all went home and divvied up the pie tomorrow, we would have a gain of over 3000% on our initial investment, or 1000% per year. Not too shabby.

Please bear with me while I do a reverse presentation of these numbers. While the way Wall Street adjusts for stock splits is absolutely correct, people who buy an asset become attached to the price they paid for it (and much investor psychology derives from this). So rather than adjusting the historical numbers for splits, let's look at our performance assuming we never split the stock.

All right, on April 29, 1982 we sold some stock for $1 per share. We issued some options to people who contributed at $1 per share in May '83. By the time we got to the next round of options, it was all we could do to beat the price down to $2.70 per share in November 1983. Then things really started to cook. The next time we had to name a price was August 1984, when we had to move it to $7.50 (and remember that during this entire period we were doing everything we could to justify as low a price as possible, so that we could issue options as worthwhile as possible to the recipients). The next time we played “pin the number on the stock” was April 1985, and by then it had jumped to $10.50. There's nothing like a public offering to move the stock price to the “industry multiple”, and ours sure did. The public offering sold on June 28, 1985 for a price of $165 per share! And last week the stock traded as high as $210 per share. Doesn't that seem different from the quotes you see in the paper?

Does anybody wish he had bought less?

You know, it's really fun writing some self-congratulatory prose after so many “crisis letters” and exhortations to exertion. Just so it doesn't become a habit….

Burning Questions Of Our Movement

So, what happens next? What should we, as the founders of this company and owners of the largest piece of it, be doing to maximise the value of what we've built? What should our company be doing to advance within the industry? How can we best apply the principles upon which we built this company to the very different circumstances and environment in which we now operate? Obviously we've done a lot of right things, but what have we done wrong? What significant opportunities are we, at this very moment, overlooking? And why are our sales only $2 million per month and not, say, ten or twenty million? Can we get there? How?

One fatal luxury of success is a failure to question one's assumptions. We must constantly be looking at what we're doing and the general environment and watch for indications we should be changing our strategy. There is, to my mind, a growing spirit of “we're number 1”, “we're unbeatable”, and “all the competition is garbage”. This can destroy us! We have to maintain good morale and believe in what we do, but we have to remember that we got where we are by running scared. There is no shortage of competitors out there with a lean and hungry look. We should be continually reviewing their products and strategies and taking the best ideas for incorporation into our own.

The only function of economic forecasting is to make astrology look respectable.

— Ezra Solomon

We've talked a lot informally about just what is involved in being a public company. I'd like to put it on paper, just so everybody has the same information all at the same time. Once you become a public company you operate in a fishbowl. Not only is the value of your company and therefore your performance rated daily in the open market, many business decisions you were free to make in private now become open for the world to see. This can lead to making decisions which may be bad for the long term future of the company in order to prevent a cataclysm in the market for the company's stock.

Those who hold and trade the stock obtain information about it primarily from the quarterly and annual reports the company files, from press releases the company issues when important events occur, through reports by financial analysts who follow the stock and are in regular contact with management, and to a lesser extent from presentations the management makes at various financial conferences. Our stock is held largely in institutional hands. This means that it is mostly in the accounts of pension funds, pooled investment accounts run by banks, and in mutual funds specialising in high-tech. The money managers who run these funds are accountable to the people who put the money in them, and their results are evaluated on a quarterly basis. If a fund is significantly underperforming the market, the money can evaporate as fast as the morning dew on the surface of Mercury. In fact, if a pension fund is underperforming the market, the custodians of it can be personally sued for malfeasance of their fiduciary responsibilities under ERISA. So to put it lightly, these money managers are under a lot of pressure.

They, in turn, look at the quarterly results issued by the company as the major indicator of the company's progress. It's a gross oversimplification, but worthwhile nonetheless to consider the stock price as made up of two components, the earnings (usually expressed as earnings per share or EPS), and the price/earnings ratio or PE. Thus:

Price = EPS × PE

The reason for breaking things down this way, is that similar stocks, such as banks, auto companies, aerospace companies, copper mines, and CAD companies will, in the absence of outstanding information peculiar to an individual company, trade at about the same P/E ratios. Thus one talks about the “market multiple” of a given industry. The P/E band moves up and down constantly; in an ebullient market such as 1983, P/E's overall may be twenty times those of a gloom and doom period such as 1974. Autodesk is in a somewhat strange position in that if it is considered a microcomputer software company it will probably settle at a P/E about half that it would command if seen as a CAD company. And of course next year, if software is in and CAD is out, the numbers may reverse. But in the minds of those looking at the stock on a daily basis, the P/E is relatively constant.

Thus, the primary determinant of the price is the earnings per share. This is very simple to calculate: you take our profits after taxes and divide by the number of shares outstanding. Zooming in a bit more, and assuming the number of shares as a constant, our earnings are broken down as follows:

PreTaxEarnings = Sales − Expenses
AfterTaxEarnings = PreTaxEarnings × (1 − TaxRate)
GrossMargin = PreTaxEarnings / Sales
AfterTaxMargin = AfterTaxEarnings / Sales

Now let's look at these numbers and what they mean in the minds of investors. The two key numbers everybody's trying to guess are Sales and EPS. Thus, if you overhear me talking to an investor trying to probe us for information, you might hear me say “we're sticking with 90 cents on 25”, which translates to “Look, I hope we really blow the top off the industry and end up with the whole pie, but I sure don't want to be dumped on if ‘all’ we do is increase our sales by 250% this year. We're 95% confident that, assuming no changes in the current competitive environment and the economy as a whole, that our sales will be at least $25 million and we'll earn at least 90 cents per share of outstanding stock.”

As each set of quarterly results are issued, they will be eagerly digested for indications as to whether the company is ahead, on, or behind expectations. Investors want to see each quarter increase both Sales and EPS from the last, and compare each quarter's results with those of the comparable quarter in the previous year to see if growth over the year matches the expected growth rate. To date, our business has not been seasonal, so straight quarter-to-quarter growth will be expected.

Before we move on to the edgy relationship between the company and the financial analysts who cover it, I'd like to define “visibility”, a key term in that relationship. Visibility measures to what extent outsiders can predict the business trends of a company overall. Consider a defence contractor. In that business, you receive contracts to do work, and the size of the contract and the payment terms are specified in advance. Any changes in the contract are public documents and are disclosed immediately in any case. Thus income is calculable by anybody who reads the paper. Expenses tend to also be pretty well predictable from historical measures, so all you need to come up with pretty reliable sales and earnings forecasts is a subscription to Aviation Leak and a pocket calculator. This is a business with high visibility.

Now let's consider a hypothetical company whose sales are almost entirely booked over the telephone. Most orders are shipped within 48 hours of receipt, so there is no backlog and no sales contracts to forecast. If the phone stops ringing, the money stops flowing. This company's sales flow through many different kinds of outlets and into numerous markets, which may behave differently as economic conditions change. The product costs almost nothing to manufacture, and is sold for a high price which is justified by difficult to measure productivity measures. The high price is largely the result of a lack of competition in the market; a determined competitor could sell such a product for $100 and make money doing it. The expenses of this business are mostly sales and marketing expenses, which are determined by the need to respond to competition and open new markets. Such a business would have really lousy visibility. I leave to you the exercise of naming such a company.

So who do those whose jobs are on the line turn to in order to decide if they should buy or sell our stock? The security analysts. These analysts usually work for the various investment bankers, and “follow” a group of stocks, usually in one industry. The analysts initially following our stock are Peter Schleider of L.F. Rothschild and John Rohal of Alex. Brown. We hope additional analysts follow our stock in the future. The analysts write regular research reports on the stock, and talk to management in order to prepare their own estimates of the company's future. These reports are then used by the institutional sales forces of the bankers to sell stock in the aftermarket. An analyst will probe to get as much information as possible, and then issue his own forecast. In some cases this forecast may be much more optimistic than that issued by the company. If the company fails to meet the forecast, the analyst will then write a report which says that the company “had disappointing earnings”, even if they represented a new high and exceeded the company's own expectations. Now this may be a little hyperbolic, but it has happened, and it does happen. Maintaining close contact with the analysts and seeing that they reach the conclusions you want is an ongoing task for a public company.

In the offering process we “signed up” to a set of performance criteria. Our investors will be watching these and, having been sold very many high tech stocks that went south soon after the offering, will be using them as triggers to dump the stock. We cannot let this happen. Therefore, here are the numbers by which we live and die. All of these numbers are consolidated, i.e., the sum of domestic operations and all foreign subsidiaries. We must do $25 million in sales this year (FY ending January 31, 1986). We must generate 90 cents per share after tax profit. Our gross margins must be in the band from 35% to 40% and therefore our after tax margins should be about 20%. We must build the company and our distribution channels and product line to support $45 million in sales the following year and $1.45 per share after tax profit.

These are the company's must-meet goals. We hope to do a lot better, but we must not do worse. If we fail, the management will be battered by the shareholders, and our stock will be gored. But the management cannot make these goals happen. The company as a whole must do this. I tried to involve as many as people as possible in formulating these goals. Now we have our job to do. Let's get on with it.

What will be the environment in which Autodesk will be operating in the future? First of all, we cannot spend the proceeds of the offering on virtually any of the needs we perceive the company to face. Since in the software business we don't use any expensive capital equipment, virtually everything we would spend the money on shows up as an expense on the income statement. If we hire people, that's salary. If we do an advertising blitz, that's promotion. From the standpoint of accounting, spending the money we raised in the offering is precisely the same as spending money we get from selling an AutoCAD. Now please refer back to the equations given above which calculate the critical numbers. If we spend the money from the offering to hire people, or to advertise, or to do any of the obvious things, those dollars are added to Expenses. That gets subtracted from Sales and reduces earnings and margins. Assuming a marginal tax rate of 50%, the reduction of the pretax numbers is twice that of that of the after tax numbers. And remember that our performance is being watched quarterly. Even if we can spend the money knowing that it will generate a major return in six months, that's not good enough. The added expenditures will affect the one or two intervening quarters, and Autodesk will be perceived as having “disappointing earnings” or, even worse, “eroding margins” and look out below. (If you think for a minute you'll see why eroding margins are a superb leading indicator of competitive pressure.)

Now this may seem to be a lot to digest, but it really is crucial to the way in which the company will continue to operate. When I say, “We have all this money but we can't spend it”, I am not setting up a smokescreen to deny people in the company what they want. I'm just describing the reality which I hope the above has somewhat clarified.

What Business Is the Company In?

This is a computer software company, y'hear. Maybe this is sufficiently obvious that it doesn't need restating, but as the company grows there is a tendency for every department to look upon what it does as central to the mission of the company. Departments then tend to see if they can make a direct contribution to the till by adding products to the company's line. For example, training could offer courses to users around the country for a fee. Technical could offer consulting services to driver developers for an hourly charge. QA could perform screening of third party software products for vendors. Marketing could prepare promotional materials for OEMs and third party vendors. And Production could manufacture third party products. Within the next year, we may be doing any or all of these functions, and these activities may be contributing dollars to the company's revenue totals. But they are incidental to the business the company is in, which is designing, developing, manufacturing, selling, and supporting computer software! Our company's value largely derives from the fact that what we do is so extraordinarily profitable. It is so profitable because we are selling intellectual property; virtually pure value added; pure reason without the critique. As I said three and a half years ago in the original Working Paper, there are few legal businesses as attractive. If we wish to go into another business, we must review that proposal as we would review the purchase of an operating company: looking at capital requirements, sales projections, pro forma income statements and balance sheets, and risk factors. This is one of the most profound decisions a company can make, and is not to be taken lightly or backed into inadvertently.

If you aren't used to thinking in the terms expounded above, the impact of stumbling into a new business can be less than obvious. Suppose we were to start doing direct contract support to major user accounts. Suppose that this was so extraordinarily successful that by the end of the year we had generated 5 million dollars in support fees, and had managed to do this with expenses of 4 million for personnel and travel. We would then add 1 million dollars to the pretax profit number and $500,000 to after tax profit. Sounds great, right? Wrong. That component of the business would be operating at a 20% pretax margin and a 10% after tax margin. When these numbers were consolidated with software operations, they would reduce our operating margins, and Autodesk would be perceived as having eroding margins. The analysts would then look at the numbers to find out what was happening and discover that we had gone into the education business. Education is not a stunningly profitable business (as thousands of colleges know, and Westinghouse and CDC learned to their dismay), and Autodesk would lose some of the attractiveness of being a “pure high-tech CAD play”. This would reduce our P/E, and the stock could be clobbered.

I don't want to dwell on this too much, but it is a problem that growing companies typically have. We'll have to keep focused on the ultimate goal of selling a lot of software if we're to avoid it.

Product Style

I'd like to talk for a moment about our products and their general style. I've spent some time recently using other people's software packages and fooling around with some new product ideas, and it's clear that many of the things we talked about happening three years ago have happened. The micro software business has become very professional very rapidly. The standards for user interfaces and ease of use have risen extremely rapidly. I'd like to talk a bit about some of the implications of this.

Why do we make clunky user interfaces? I think that some of our much-vaunted “mainframe approach” to software may be leading us into some poor decisions in the current environment. We always build software to be easily ported, machine independent, and easy to maintain and enhance. These are things much to be desired and unqualifiedly good, as long as there is no cost to the user. If ease of development or support imposes a performance, convenience, or learning cost upon the customer, this must be looked at as a tradeoff, not decided preemptively in favour of the developer.

I would invite you to spend 5 hours using a program Dan Drake turned me on to, “Managing Your Money” by MECA. This is a $120 program of extraordinary complexity. I would rate its connectivity and integration as approaching AutoCAD. It fills three discs. You have to read about three pages of manual to get started on it, and it contains hundreds of pages of intelligent, useful, and witty on-line assistance. All response is absolutely instantaneous. No error is fatal. You can always back up.

I think our tendency is to adhere to the reggae rule of “all killer, no filler” in designing our programs. We tend to eschew user interface “fireworks” such as instantaneous screen updates, fill-in-the-forms data entry, pop-up menus, and function keys in favour of solid, well engineered but prosaic programs.

This is one of the assumptions we should question.

How long are users going to accept a product which requires mastery of a 300 page manual? In a market which is dominated by IBM and compatibles, what is the opportunity cost of not deriving greater advantages by tailoring to these machines? Should we do a Macintosh or Atari product? I don't have the answers to these questions. But I think we have to consider them.

What do we do next?

On the operational side, I think we should pledge a significant effort to doing this job of being a public company right. We showed the Silicon Valley cynics and the venture capitalists how a bunch of dedicated, talented people could do a start-up company right. Now let's show Wall Street how a small public company should be done. What I propose is that (all within the limits of the law, of course) we treat our stock as an Autodesk product. Let's put together a shareholder communication program that rivals the ones we've created for our dealers and users. Let's put on shareholder forums at our annual meeting just as we do for the dealers and conference speakers. And let's have an informal shareholder forum at the 6 month point between meetings. Shareholders can be a pain in the ass, but they do own the company. And let's see what we can do to make this perceived as a very special company to own stock in. We must be open and candid. We must present realistic numbers and always meet them. If we're not going to meet our numbers, we must give warning as soon as we know, and have explanations ready to deliver. These acts will build loyalty that will stand us in good stead when times get tough.

The Next AutoCAD

There's absolutely no way we could put a man on the moon by 1994. No way.

— NASA Official, quoted by Jerry Pournelle at the 1984 L-5 Convention

It's also time to look for some new product ideas. Let's not settle down into this “going concern” mentality where we're maintaining a program that came from “somewhere” and is going “somewhere”. We built this from zero. We can do it again. Can a company our size continue to develop multiple products on the cheap, test market them, and get behind the winners? Can we rekindle the old “ten wolverines in a barrel” technological ferment we had when the idea of this company was fresh and new? Can we have more yelling and less nodding? Will somebody please come to me and ask to spend a week in another department to learn what the hell they do in there anyway? What is the minimum time in which we can bring a product to market now? Could we bring a new product to market this COMDEX? Look at the development logs to see where AutoCAD was at this point in 1982.

The American Dream

Conrad Hilton didn't make his fortune by building only one hotel and then opening a car wash. He may have had a car wash before he had a hotel, but once he found hotels and they succeeded, it was hotels, hotels, and more hotels. Did Ray Kroc who owned McDonald's start one hamburger stand and then open a dry cleaners? No. He had a winner, he stayed with it, and wealth was accumulated.

— Dr. Bruce Gould, “How Fortunes are Made”, in Bruce Gould on Commodities (newsletter)

There are two ways to look at what we've done here. We're either an awfully lucky bunch of weirdos, or we're really on to something. With every passing month and every milestone we reach, the luck seems less likely and the strategy seems more astute.

Let's franchise it.

Franchise it? Well, not exactly, but here's the idea. We started out as a group of people with limited financial resources but a great pool of diverse talent and willingness to work hard. We built this company as opposed to going to work somewhere else because we felt that this was the best way to achieve the success we wanted to reward our exertions.

I don't think we were the only people in the world with these goals. Let's build ten, twenty, or fifty more Autodesks. How? By offering the same kind of partnership to entrepreneurial people that Marinchip offered to the founders of Autodesk. We publicise the following proposition:

You want to start a company? We know how to do it. Look at our results. Put together your group, count your money. Here's a set of information letters that tell you how to do it. Come to us with your proposal. We don't want a long business plan; you don't know what's going to happen any more than we did, and if you say you do you're a bullshit artist and aren't worth listening to.

We'll look at the people and the product ideas. Is there that sparkle you can see in the first 5 minutes? If we're believers, we'll match your investment dollar for dollar. In return we get 35% of the company, held as soft preferred which basically protects us against being ripped off, but we're in there with all the other founders. But it's the founders' company. Our investment in these companies will not be an expense on our balance sheet. The investment will purchase stock which will be carried as an asset. I'm not sure how often such illiquid assets get marked to market, but we can probably let it just sit there until we either write it off or begin to get proud of it.

Autodesk will provide limited support to the venture. We will see that the legal details are taken care of correctly and that the accounting is of public company quality. We'll provide a pool of talent to the management that has “done it before”. We'll offer technical resources and the facilities of our lab on a sporadic basis. Our distribution channels, marketing and promotion resources, and our ability to promote products at shows and in publications at small marginal costs will be important resources to companies affiliated with us. Our manufacturing and shipping operations can provide those services at low costs per unit.

The founders of the new companies can choose to use our services, which will be billed at attractive rates, or to build their own or go outside—it's their company, career, and destiny. We will be providing what all the venture capitalists claimed they did, “bringing more to the table than just money”. The only difference is that we really will.

Our goal is to give the founders of these new companies the same shot at success we had when we started. Autodesk will provide some cash that we sorely could have used, but not enough to mess things up, and the ongoing establishment the creation of which cost us so many critical hours we could have better spent elsewhere. We won't run the show, but we'll try to be there when we're needed. Many of these companies will probably fail, but if 20% work, they will contribute mightily to Autodesk's success.

Just imagine if we pull this off. I hope we always retain some of the rabble-rouser elements of our creation. I can think of nothing I'd like to do more than drain the talent out of these Silicon Valley companies that are screwing their key people and giving the equity to the venture capitalists. Instead, here will be Autodesk, with one face talking to security analysts and breaking new ground as the model for small public companies, and at the same time erecting a rickety, low-rent conglomerate built on talent and hard work, of hardscrabble start-up maniacs who, just once, want to do something right and own it.

I think that people would be well served to take the chance we'd be offering. They'll have to have real commitment, real performance, real responsibility, and real professionalism to make it. If they're interested in making that kind of commitment, we can't guarantee that they'll succeed, but we can guarantee that together we'll have a once in a lifetime experience as we try. (Working paper, 1/82, Page [Ref]).

Now putting this together will take some work. But how much, really? Let's think about it, and see if we can pull another sleeping shocker on the industry. Can you imagine, just imagine, ten companies, all loosely affiliated, working like Autodesk all at the same time. Why they'll say it's a movement.

And that's exactly what it will be.

Surplus Value     Time of Turbulence