From Computer Business Review, a sister publication…

When IBM’s chief executive Lou Gerstner sat down in May last year with Jerry York, his finance chief, to discuss a takeover price for Lotus Development Corp, the software company’s revenues and profits were at the bottom of the agenda. Higher up on the list were brand name, its research and development and distribution reach, the caliber of Lotus technical staff, and perhaps most importantly, the groundswell of sentiment that was moving in favor of Lotus’ popular Notes groupware package. Also on the agenda was the question: What’s the going rate for hot software technology? The value of Lotus, IBM and its advisors concluded, could not be calculated on the basis of incremental cash flows, multiples of earnings or a valuation of intangible assets – the pricing parameters used by every Wall Street analyst. It was a calculated gamble, based on unpredictable trends and the popularity of market-leading products during the next five years of increasingly network- centric desktop computing. If anything, Lotus’ recent history suggested that a low offer might suffice. A month earlier, on 19 April, the spread-sheet and groupware software supplier had announced first quarter net losses of $17.5 million on revenues down by a disturbing 20% to $200 million, equivalent to annualized sales of perhaps $800 million. In an accompanying statement, Lotus attempted to soothe anxious shareholders by pointing out that its Notes package had already shipped a respectable but insufficient 220,000 copies in the first quarter alone. But the statement appeared to be a plea for lenience. Lotus looked desperate. Notes, recognized as its best hope for a flagship product to replace the famous 1-2-3 spreadsheet and related products, was a success, but it contributed less than half of Lotus, revenues and nothing to profits. As such, it was not yet ready to fill the gap left by falling desktop software sales.

Potential for exploitation

Shortly after that first quarter results announcement, Lotus’ market value fell to $1.5 billion on the Nasdaq exchange. But a month later, with takeover speculation boosting the share price, IBM and its advisors reached a different calculation. What was Lotus worth now? IBM was initially willing to pay $3.3 billion, or $60 per share. A month later, when the deal was completed, IBM paid $64 a share, valuing the company at just over $3.5 billion – almost four and a half times Lotus’ annualized sales. Justification for the high figure centered – not surprisingly – on Notes and on IBM’s potential to exploit the product. If left to Lotus, Notes was worth far less. The uncertainty surrounding the company might affect sales, and the prospect of a groupware price war with Microsoft might affect research and development funds. In IBM’s hands, the value of Notes, and so Lotus, increased dramatically. But was that value set too high? In the context of the subsequent emergence of Internet software supplier Netscape Communications Corp and the Intranet – company-internal Web page servers acting like groupware servers – many have argued so. But in the context of the software industry as a whole, the price was not astronomical. Whether measured by the value of companies at IPO (initial public offering), by takeover price-tags, or by straight forward open market valuations, software companies have never been so valuable.

Startling trends

Figures collected by the IT mergers and acquisitions (M&A) consultancy Broadview Associates highlight the acceleration of IT acquisition valuations. Between 1992 and 1993, the total value of IT acquisitions worldwide increased substantially from $21 billion to $28 billion, but in 1994 the value of transactions really took off, increasing to $68 billion. In 1995, the figure had risen to $134 billion, a 48% rise. The most startling trend was – and still is – in the software sector. Software companies single-handedly pushed the value of 1995 deals up by $35 billion, a year-on-year jump of more than 100%. More to the point, the average transaction size among software companies during 1995 soared, to at least double 1994 levels (which had already doubled in 1993). And the high price-tag trend continues: In December, database supplier Informix bought Illustra, a supplier of object database technology with sales of $10 million a year, for $350 million, with a further $100 million payable to redeem share options. In January, IBM offered $743 million for Tivoli – a supplier of systems management software – with revenues of less than $50 million. These two valuations represent Price-to-Sales Ratios (PSRs) of 35 and 15 respectively; figures that can be compared with the 4.0 to 4.5 range that Microsoft usually occupies (i.e. its Nasdaq market value compared to annualized sales). Calculating such valuations is a mixture of solid arithmetic and gut feel. But has there been a dramatic shift in emphasis from one factor to the other, leaving the purely numeric all but redundant? Valuing IT companies is not a science, says Patrick Seally of Broadview Associates. If it is anything, he says, then it is an art. And that art is about knowing what is going on among the core technologies and how much companies have been paying recently to buy into them. When it comes down to it, that’s really all you can do. The financial aspects of the calculation come at the end. Often then only to rationalize a decision that’s really already been made about price. Broadview’s methodology is not quite the finger in the air that Seally might suggest – but it does seem to trash the established economic models relied on by corporate financiers since the 1960s. And as the largest M&A deal broker, Broadview does enough business to make it representative of the industry. It has advised on, among others, Intel’s takeover of Phoenix Technologies, EMC’s purchase of McData, Symantec’s of Delrina and Sterling Software’s of Syncsort. The view of the M & A advisors to the IT industry is that it is meaningless to begin to talk of long-term earnings forecasts and discounted cash flows (DCF). These calculations work for car manufacturers and hoteliers, but fall to pieces for companies which are experiencing rapid growth and rapid developments in technology. In the context of rapid change, looking more than a couple of years ahead becomes fruitless. But if this is the case, what yardsticks can be used? With the option of valuing businesses on hard assets and stable profitability removed, Broadview and similar consultancies have been forced to consider alternative approaches to IT valuation. We have to look at qualitative issues. We’ll take a classic approach to competitive analysis which involves looking at market positioning, products, quality of people, competitive threats and the trends in technology, says Seally. The thinking is that if you get these issues right, they will translate into the right numbers in the balance sheet and profit and loss account. This is the theory. In practice, the IT industry is moving so fast that people now speak of ‘Internet years’ and ‘infanticide’ to describe the speed of change and the early obsolescence of new products. In this context, the value of recent high-price acquisitions may never be clearly apparent as the perspective is clouded by the technological and business changes.

On the trail

Nobody knows the truth of this more than the executives at Symantec, the Cupertino- based utility software provider. The company has made over a dozen acquisitions in the last five years: Peter Norton Computing in 1990 for $70 million, Fifth Generation Systems for $48 million and utilities specialist Central Point Software for $60 million (both in 1994). Yet no-one there boasts that they have got it 100% right. Michael Bailey, one-time Silicon Valley venture- capitalist and now senior vice president of business development at Symantec, believes that the rate at which technologies are rendered obsolete is the most destructive influence in the IT industry right now. Not only do investors and suppliers run the risks of backing the wrong technology, but it makes even medium-range planning immensely difficult. You can only justify your decisions beyond, say, one year in purely strategic terms – not financial, says Bailey. Last year, Symantec acquired Delrina, supplier of the well-known Winfax software and other document management products, for $415 million. At the time, Delrina was turning over around $90 million a year. Why did Bailey recommend paying $415 million and not, say $350 million, or $550 million? For the numerical purist, his answer is initially unsatisfactory. He believes that a look to the market is all one can do to establish ‘ball-park’ valuations. Both the seller and the buyer will do the same exercise. The deal was hard to negotiate and the price we arrived at was based on clear notions of value on the part of both vendor and acquiror. We always want to pay less and the selling company always want more. Sometimes the two ranges overlap and sometimes they don’t, says Bailey. But the bartering does have some quantitative aspects. As we develop an idea of what our range is, we factor in a qualitative judgment about the long-term growth prospects for the business, coupled with quite a rigorous short-term analysis of the incremental revenue and profit impact on our own business, he adds. The longer term analysis, says Bailey, is based on trends in technology and Symantec’s ability to address these trends. But the short term analysis is based on a detailed spreadsheet projecting incremental revenues and expenses over six to eight quarters. If we are issuing shares, the incremental profits-per-incremental-share should exceed our own recent earnings per share. In other words, it can’t be dilutive for our existing shareholders. If he can achieve all that, says Bailey, he’ll buy. A key short-term parameter for valuation seems to be the natural and rigorously defined limitation imposed by shareholders, who tend to resist dilution in earnings per share (EPS). Bailey is clearly mindful of this measure, as is IBM’s head of finance, Richard Thoman (who replaced Jerry York in mid 1995). Thoman sent the message to Wall Street analysts that any acquisition IBM undertakes will have to be earnings enhancing within two years.

Business imperative

But the ‘earnings enhancement’ rule is not always applied. Last December, the $1 billion-a-year database provider Informix announced that it was to pay $350 million for Illustra, a business with revenues of less than $10m in its year ended 30 June 1995. Geoff Hudson, Informix head of business development, acknowledged a probable minimum 10% dilution in EPS. To justify the price-tag we had to make two key assumptions: The first was about what it would do incrementally for our current businesses, the halo effect of simply possessing Illustra. Then we had to look at the revenues from Illustra’s product line and how we could accelerate those by integrating them with ours. On our calculations, we believed that the deal would be maybe 10% to 12% dilutive in the first year and marginally dilutive in year two. But after that, things would take off. Illustra represents Informix’s ‘Holy Grail’, according to Hudson. The company, set up in 1992, designs object database technology extensions which could make Informix’s existing database products far more useful in rich data environments – such as supporting Web pages and storing images, video, sound and 3-D spatial information. It is, Hudson says, the ideal multimedia complement to the current Informix portfolio. Although a year ago Informix was reticent about how it would be able to support rich data types, it now feels confident that it has the edge. High performance support for complex data is the next great battle ground for the DBMS vendors, declared an Informix white paper on the deal. At its 35-to-1 PSR, Informix clearly believes it is a battle it will win. According to Hudson, there are two ways to view these software acquisitions. You can either do what someone like Computer Associates does and buy up companies simply to strip them down and sweat cash out of them, or you can take the approach which entails buying technologies, or people, in the belief that in the future they will add to the developing markets in which you want to operate. Hudson presents it as if he had no choice. We had to decide whether to buy in, or develop certain technologies in-house. We had to buy and realized that Illustra had what we wanted. Once we identified them, we just had to decide on price. Hudson insists that he could not have paid a dollar less for Illustra than the $350 million plus that it paid. Even that price was tough to achieve because there were a lot of precedents available for the vendors to point to – recent IPOs, current stock market valuations and other acquisition price-setters. In fact, just looking at current stock market activity, it wasn’t difficult for Illustra to point to a likely flotation value of $750 million, even a billion, twelve months down the line, and they were some way down the road in preparing to do just that, says Hudson, adding We probably got them cheap.

A question of value

Would shareholders tolerate the dilution? In the end, this was easy to: In the months before the acquisition Informix’s share price was in the upper $20s. When the takeover was announced, the price actually rose, despite the heavy short- term earnings dilution, to around $32 to $33. Now it’s at $31. So we feel that the institutions and the analysts are thinking the way we are about valuation says Hudson. (Late April, Informix stock had fallen to $25.) One company which will almost certainly argue that Informix paid too much is its big rival Oracle, the leading database supplier. Last summer, Oracle, explored the possibility of purchasing Object Design Inc. (ODI), a supplier of object databases, which was then running with annualized sales of around $30 million. ODI reportedly demanded $200 million, a price Oracle refused to pay. The deal was never consummated. ODI’s chief executive, Robert Goldman, is reluctant to rake over the failed Oracle deal, but his comments suggest that he believes Oracle was foolish to pass him up on the basis of a marginal mismatch on their respective expectations on price. Price, believes Goldman, is normally a clear-cut issue. It’s not a question of what the value of any particular acquisition is at the moment. You have to look at an acquisition in two stages: First, what the company is going to cost. The answer is usually easy to work out, just by looking at prevailing valuations for similar businesses in the public markets or the private sector. That’s the easy part. The really tough question comes next. Is this company worth what I’m going to have to pay? Like the M&A specialists at advisors such Broadview, Goldman takes the line that price calculations are more determined by the market than by any scientific measurements. This may be heresy to say, but if you hire any number of highly qualified and sophisticated corporate bankers, they will all use their own elaborate financial model and they will all come up with a different valuation. The answer lies in whatever view you take of the way the technologies are going. In this context, Goldman thinks the recent valuations of software companies are reasonable. Many deals at the moment – with valuations based on high multiples of revenues – seem incredibly expensive, but at a time when the world is changing so dramatically, really driven by the Internet, many of these deals will prove to be very cheap in the long run. IBM’s justification for the price it paid for Tivoli – almost $750 million for a business with revenues of $50 million – enters another dimension. Bill Bock, Tivoli’s vice president of operations, explains: We are not just being acquired and tucked away into a corner of IBM. We are being given management responsibility for IBM’s own product lines in this product space. In other words, Tivoli is charged with revitalizing IBM’s whole systems management business. IBM has transferred a significant business unit – its client/server division and mainframe system management division – and combined it with Tivoli. We will be managing all of it for IBM. These activities represent revenues in excess of $1 billion, and we will also be managing our own R&D as well as IBM’s in this area, says Bock. IBM’s shareholders have no problem with the agreement. Those that follow IBM closely enough will know that IBM’s recent attempts to unify its systems management tools under one banner – SystemView – ran into trouble, forcing several rethinks and causing some loss of market credibility. Tivoli, however, has a worldwide reputation for excellence in the field; part of its value is its credibility and the ability of its managers to take charge of IBM’s projects.

Dissident voices

The role of precedent – recent deals and flotation values – in the decision mix emerges constantly. But some voices are expressing doubts about the wisdom of continuing with this approach. One senior director with a leading IT M&A advisor asked not to be named when he said: These me-too valuations have a nasty habit of becoming self-fulfilling. They ignore the dramatic cyclical variances in valuations. With everyone pointing to everyone else’s deals as a justification of their own bid price, it doesn’t take long before things get out of hand. Valuations just trend upwards and upwards. We’re getting dangerously overheated now. Such a view is widely held – but rarely said publicly. But does that mean consultants will now quietly begin to guide their clients’ valuations downward? Even if they do, their message is unlikely to fall on receptive ears. The same executive describes the scenario: It’s almost impossible to point to tangible evidence which proves that prices are overheated at the moment. The purchasers – our clients – believe that we are in a once in a life-time, perhaps once in a century phase in technological development. Internet, multimedia and digital communications are exploding in their faces and they believe they’ve got one chance to hook into it. If we advise them to shoot low on price and then they are outbid for the next standard-defining technology, we would have very little in the way of a defense. It’s a dangerous, uncontrolled and uncontrollable lurch that we are all participating in and none of us knows how to get off. ODI’s Goldman also senses that a built-in destruct mechanism is ticking away under the floor- boards of Wall Street’s IT brokerages. Prices are as high now as they’ve ever been. Private companies, especially, are looking at public markets and the Internet, and as long as [investors and acquirors] pay multiples of revenue, they’re being drawn into it. It’s not a question of whether they are worth the money, it’s simply a question of supply and demand. If I’m a private company and I can be valued at 10 or 20 times my revenues, then I’ll demand that, plus a premium to give up control of my business. Beyond Goldman and one or two insiders, it is almost impossible to find a dissident voice in the IT industry. In the short term. everyone gains. No-one has time to stop and think too hard, and there is little tangible material on which to base valuations. So it will continue in the near term, until the computer market or the pace of innovation slows down. There is even an argument that ultimately, it does not matter if the investment proves a good one in the long term. Chris Galvin of IT specialist brokers Hambrecht & Quist, based in San Francisco, makes the point that for a lot of these companies, it is tough to pull out their contributions a year or two later. The gamble is that attractive technologies, if successful, can be better exploited by the bigger acquiror. So really, the prices paid for these acquisitions have become an abstraction. Investors and management are more interested in whether the stock price goes up.

Traditional valuation models

How do investors value companies ? At the heart of many economic valuation models lies the concept of Present Values (PVs). To calculate PV, the future cash flow is calculated to a ‘present value’, with the cashflow discounted to allow for future risk and inflation (a certain dollar today is worth more than an uncertain dollar tomorrow). The further away a cash flow is (e.g. three years, five years), the less its PV becomes. By estimating all future cash flows, a single PV is calculated. If this figure is greater than today’s acquisition price for a company, then the deal is attractive. PVs represent the most rational system available for numeric valuations, but even for more stable industries than IT, they are flawed: PVs can vary wildly according to (i) the estimates of future cash flows and (ii) the discount factor chosen. Small changes in the chosen discount factor can have a dramatic effect on the PV, making the same transaction appear either attractive or unattractive. So in a sense, such methods are just as unscientific as those which prevail in the IT business today. In any case, few involved in IT company M&As place too much emphasis on this figure. The unpredictability of the IT business can make ten-year cash flow forecasts seem a little redundant.

Wall Street’s search for high-tech ratios – Price:earnings ratio

The most used and quoted measure of valuing shares and companies is the price:earnings ratio (PER). This is a ratio of the share price to the last reported (historic) – or sometimes forecast (prospective) – earnings per share. A PER provides otherwise incomparable share prices with a common yardstick. Higher PERs indicate one of two things: Either investors recognize higher growth expectations for a company or alternatively, the shares are simply over-valued. But PERs have severe limitations – especially when applied to young technology stocks. Companies with no profits, or even losses, will have an infinite, and so meaningless PER that will fail to indicate whether the company is in recovery or terminal decline. In addition, the ratios denominator – earnings per share – is pulled from published accounts and so varies dramatically according to differing accounting conventions. It is often distorted by such intangibles as goodwill, depreciation, amortization of R&D, and provisioning. Nasdaq’s average PER stands at 23, compared with perhaps 17 for less technology-oriented markets. The average PER for US-quoted software stocks has varied widely over the last decade, but has usually stood at a 20% to 50% premium to the Nasdaq market average. Microsoft has spent much of its life in the 30 to 40 range.

Price:sales ratio

The price:sales ratio (PSR) is an invention of Forbes magazine journalist and author Kenneth L. Fisher, and is now an established measure of value. PSR is the ratio of market capitalisation to sales. By drawing the emphasis away from earnings to focus on revenues, short term profits volatility is disregarded. This is useful when valuing immature businesses participating in rapid growth IT sectors where short-term profitability is sacrificed in favor of market share gains and heavy R&D spend. In the absence of profits, the PER is useless, but the PSR comes into its own. PSRs have rapidly gained recognition as the pre-eminent valuation metric in the more speculative sectors such as IT and bio-technology. On Nasdaq, the overall average PSR is higher at 1.7 but software companies average 3 to 4. The average PSR for the London FT all share index currently stands at 1.2.

Net asset values

The net asset values (NAV) ratio is used more to value property companies than for IT valuations. NAV is calculated by deducting all of a company’s liabilities from its asset value. Most software companies have few assets which might appear on a balance sheet and so have a negligible NAV.