The Decline of the Optical Networking Industry Forecast and How it Can Be Stopped

In the wake of the carnage that has befallen the "optical stock market", the industry has not only been asking itself why things have gotten so gloomy but also why expectations for the industry got so absurdly high in the first place. CIR believes that blame can be assigned to various causes, but increasingly there is an active debate in the trade press and behind closed doors that exaggerated industry forecasts were a major contributory factor.

When the stock market is convinced that a certain industry sector will turn in revenues of several billion every year based on overly optimistic (or perhaps even deliberately exaggerated) forecasts, valuations and expectations will then obviously scale beyond reason. Companies that, at best, would command a price in the low hundred millions instead get to add an extra zero to the valuation without any real business fundamentals to support the price of the stock. Then along comes a market slowdown with its inevitable reality check and the results become quite devastating.

Unfortunately, major stock market corrections are not the complete picture of what happens when forecasting becomes a game of, "Whose numbers look higher?". Such forecasts can lead to misallocation of resources, not just in terms of the stockholders funds, but also in terms of internal planning. Companies struggle to meet market expectations and end up chasing after unachievable sales targets and living quarter-to-quarter. The need to have a story about the "All Optical Washing Machine" forces firms like Nortel, Lucent and Cisco to purchase companies with unproven and even non-functioning technology and meanwhile, staffing becomes a veritable bidding war for talent. Resources are allocated to expand manufacturing capacity and even more money is spent on trying to distinguish the company in the chorus of market hype necessary to obtain investment or maintain the growth in stock prices.

Now, it would be wrong to claim that the problems that the optical stock market is facing right now rest solely on the unrealistic forecasts that analysts produce or companies often quote. However, there is a growing sentiment among many of the people that we have spoken with that market forecasting has been biased or lacking in credibility and that has not only added to the correction but damaged the industry itself. Furthermore, this sentiment reflects a strong belief that if we are going to see a return to realistic and sustainable growth, companies need to do what they say they are going to do and that the projections based on hype, wishful thinking or client appeasement are going to have to stop.

This paper explores two questions: (1) why exaggerated forecasts occur in the first place and (2) whether anything can be done to stop the decline in quality of the optical industry forecast.

In answer to (1) we will see that this decline is rooted both in forecasting methodology and the self-interest of the forecasters. CIR does, however, believe that the answer to (2) is that the decline can be stopped.

As someone once quipped, forecasting is a hard thing to do, especially when it is about the future. . . and the optical future is especially hard to forecast. It is a far simpler task to forecast new car sales -- there are years of data on which to base such forecasts and advanced statistical techniques that can be used to manipulate the data into convincing market projections. When you want to forecast metro DWDM, tunable lasers, optical switches, etc., there is no historical data per se and with immature technologies, uncertainty as to available applications and unproven demand, it would appear hard to build a market forecast that would convince anyone.

Let’s suppose we were trying to build a market forecast of the optical switching market. We are immediately faced with a bunch of almost unanswerable questions such as, "Will the Corvis approach of many wavelengths on a single port beat out the approach used by most other companies in which there is one wavelength on every port?" or "How will optical integration impact switching chip pricing?" or "How far will optical switching penetrate the network?". Meanwhile, talk to router vendors and they will have a very different view.

Forecasts embody huge uncertainties, including assumptions about the future that can make differences in market estimates measured in orders of magnitude. This can be very frustrating, especially – as is often the case – when assumptions (or agendas) are not made explicit. And, even when such explanations are available, there can be profound differences among reasonable people about assumptions, which, in turn, can invalidate an entire set of forecasts.

One response to this methodological mess is to give up entirely, take the lazy way out, and make the claim that objective forecasting in our industry is an impossibility. One view that has recently become popular is that when dealing with disruptive technologies, forecasting is not needed.

So why bother with forecasts at all? Unfortunately, this question is soon answered by someone in the company who notes that forecasts are needed to help win favor with various constituencies. Managers need forecasts to sell their projects to senior management and senior management needs forecasts to keep investors happy by providing "visibility" and thus keep their company’s stock price high. And, of course, start-ups need them to impress the venture capitalists, validate business plans and impress the investment bankers so that they can become the next hot IPO.

And if forecasting the optical industry is not to be taken seriously anyway, why not take an "anything goes" approach, come up with some forecasts that keep everyone happy for a while and give your company a bit of a boost?

The answer to this question was actually given at the beginning of this paper and this is that in the long-run, exaggerated forecasts can contribute towards a serious loss of confidence in individual companies or in the industry itself. But, the "long run" can seem like eons to managers in the optical industry where product life cycles are down to about a year and where the average tenure of an engineer or a manager at a company is about as long.

So in the short-run, forecasting is gradually taken over as part of the marketing and PR function and is seen by many as little more than a way of helping companies win favor or impress an increasingly skeptical market. Exaggerated forecasts, of course, may be produced internally or by outside analyst firms whose forecasts have tended to appear more impressive because of supposed objectivity. A recent article in America’s Network magazine however, reveals what lies beneath this veneer, namely the cozy relationship that often exists between many market analyst firms and the PR departments of equipment and component vendors. It has been claimed that because of these relationships, analysts’ forecasts ultimately reflect the hopes, dreams and fantasies of the vendors and that any attempt by analysts to be more objective will be punished by vendors buying analysis and consulting services from another, more docile, analyst firm or in the case of a financial analyst, saying good-bye to underwriting fees or other lucrative investment banking opportunities.

The purpose of this paper is not to examine the sociological aspects of market forecasting but suffice it to say, CIR believes that the kind of criticism made above is well deserved. The end product of the relationship between market analyst firms and vendors is typically what we call the hype-based forecast. A hype-based forecast sounds something like this:

"The market for [Insert favorite optical component or equipment category here] will grow from under a billion dollars in 2000 to nearly (fill in the blank) billion dollars by 2004."

At first glance such a forecast would normally be enough to make any manager’s heart jump for joy. On a second glance, however, one should be suspicious of such statements pending review of the underlying assumptions and possible vested interests in pushing such numbers. What should you look for?

• A demand growth pattern that is hard to believe: Five-fold growth of any market is pretty outstanding — unless the market is just beginning to evolve —and with the dampened expectations for service provider spending in both Europe and North America, such numbers should strain the imagination.

• Ignoring historical economic realities: It is often times hard to predict when the big crash will happen or even when economies will begin to slow. However, to expect that markets will not contract or that CAPEX will continue to grow forever at some increasing or even stable rate is a freshman business student’s mistake.

• A supply pattern that is hard to believe. Even if the demand was there, one has to question whether the supply is adequate. Clearly the optical components industry has been a victim of shortages and some recent research on the sexier components — such as optical switching chips and tunable lasers — suggests that these shortages aren’t going away. Unless we are talking about major price increases, which is always unrealistic to expect in any industry (save for energy), or some misleading definitions, it is questionable that there is enough capacity in the industry to produce such numbers.

• Technology adoption never keeps pace with technology speculation. How many technologies have burst on the scene in our lifetime that were supposed to take over the world — or at least some part of the network — and actually did?

So where are such numbers drawn from? Usually from the manufacturers of optical components and optical equipment, whose understandably bullish expectations serve as the basis for the aforementioned forecast. On one level, of course, there is nothing wrong with the expectation that a company will succeed. After all, if you are not ambitious in your plans to be a market leader, why bother? There is nothing fundamentally evil or dishonest about manufacturers being optimistic and a little optimism can get one through the dark times. However, there is a bill that needs to be paid for excessive optimism and the market corrections that we have experienced are clear evidence of that.

So far we have painted a pretty gloomy picture of the optical industry forecasting business. Based on what we have said so far, the reader of this White Paper is probably left with the impression that objectively forecasting the industry is hard, if not impossible, and that — at the risk of making industry forecasting sound like an episode of the X-Files — even if objective forecasts could be achieved, there are powerful and shadowy forces at work that would seek to suppress them! This is a gloomy picture indeed for the absence of historical data, the forecasting methodologies employed and corporate self-interest are forces that are surely difficult to overcome.

But all is not lost. Take the point about self-interest first. It is clearly true that there are many individuals within an industry who have a strong interest in seeing market forecasts exaggerated. But, it is also important to remember that there are powerful interests that favor – or at least should favor – objective forecasting. Hype-based forecasts are unlikely to impress those managers whose careers are dependent on constructing corporate plans or who are told that they must gain a market share of say 10 percent in a particular market which is supposedly $1 billion, but is actually only $350 million. When such managers turn in annual sales of $50 million, they should actually be commended for doing better than the 10 percent asked for. Instead they are fired for only achieving 5 percent of a $1 billion market that never existed in the first place. Hype-based forecasts should also fail to impress smart investment types who are held accountable for where their dollars go. (They should even be questioned by sell-side analysts who appear on the evening financial television programs, but that remains a point outside of this paper)

We therefore believe that a revival of objective, or what we at CIR call "reality-based" forecasting, will require the support and practice of those within the industry whose vested interests lie in longer-term success. However, this will not occur unless people can be convinced that there really is a way of forecasting the optical networking and components markets that is far more credible than what they have seen to date.

Reality-based forecasts are based on end user needs and "hard data." The differences between hype-based forecasts and reality-based forecasts are summarized in Exhibit I. But the main way that hype-based forecasting and reality based forecasting differ is that reality-based forecasting demands that forecasts are related back to real market needs.

The best approach to achieving this goal is to talk to end users – for example talk to service providers, if the product you are trying to forecast is optical networking equipment; or to optical networking vendors if the product you are trying to forecast is optical components. This is the only way to understand real needs and the real pace of change. By talking with the customers, we get a better idea of when things will become real – which is extremely important in coming up with realistic forecasts.

When one takes this approach one often finds some important distinctions between buyers and sellers. Two almost inevitable trends are the following:

• Buyers are less optimistic than sellers about future purchases. For example, while some equipment vendors are telling us that mass deployments of all-optical switches or metro DWDM are just around the corner, the service providers continue to be more cautious. Not taking this caution and proven conservatism into consideration when forecasting often leads to inflated numbers, as does believing everything that vendors say about expected sales.

• Buyers are less impressed with "technology stories" than sellers. While technological innovation forms the heart and soul of the optical industry, it can be overstressed. CIR research shows over and over again that what buyers want often differs from what vendor products can actually do. Vendors stressing that they have a new technology, without really explaining what this can do is, of course, a classic high technology marketing error and tends to lead to lost sales opportunities. From the perspective of forecasting, there is a danger for forecasters of being seduced by vendors into believing that a new technology will be rapidly deployed because it is new. In fact, buyers are often very skeptical of new technologies and many new technologies simply disappear within a year or two. For rapid deployment of a technology to occur, it must usually provide an order of magnitude improvement in price/performance ratios – or better, not to mention bringing down operational cost in a proven and quantifiable manner.

Understanding these eternal verities of forecasting is only part of the reality-based forecasting process. The other aspect of reality that needs to be brought to bear in reality-based forecasting is the use of hard data. And – excuse the pun – such data is hard to find. In the end, forecasters may have to resort to indirect methods such as talking to as many end-users and purchasers as possible -- provided of course they have relationships that extend beyond PR or Investor Relations departments. Researching census data and FCC filings can also be a useful source of data. But as a matter of course, researchers, should automatically deem any vendor claims of a customer suspect unless the PO is placed on the table!

Credible forecasters can also rely upon past history of the industry, proven knowledge of the market and a track record for being reasonable in their assumptions.

The reality-based forecasting process is summarized in Exhibit II. However, the point of this White Paper is to provide less the definitive guide to reality-based forecasting in the optical industry, but more of a statement that this type of forecasting is even possible.

Even the best reality-based forecasts are not certainties. At their very best, they are built on a platform of conjectures, estimates and prior experience. However, contrast this with the hype-based approach in which – according to the forecaster, anyway -- every market is certain to be a multi-billion dollar market regardless of how many times in the past it didn’t happen or how outrageous the future looks to be. We believe that it would be an interesting exercise to add up all of the supposed market projections around our industry and see how close we would get to a ridiculous trillion-dollar telecom market. Yet, for some reason, no one seems to have bothered to carry out this calculation.

Reality-based forecasters inevitably set themselves up for criticisms from those who fundamentally disagree with the assumptions on which the forecasts are built. True, it always looks nice to toss out numbers that measure markets in the tens of billions and until recently, the market accepted them as credible or worthy of quoting. And, as a marketer, one should always attempt to make the world look as grand as possible and talk up the available, yet unattainable, opportunities. However, living for today without having to be accountable for tomorrow is the hangover that we all have to endure.

CIR does not purport to have all the answers or that we are experts at predicting the exact future. What we do, however, want to emphasize here is that by building up forecasts from an understanding of the realities of the telecommunications industry, rather than by summarizing hype and customer expectations, one will end up with forecasts that are actually useful for planning, investing or selling.

Reality-based forecasts require among many things, industry experience, good sources for finding out what is true, and sound judgment. But, we believe that judgment of another kind is also required by forecasters and that is the kind of judgment that is synonymous both with ethical and responsible behavior. Hype based forecasts might look good in press releases, short term stock valuations and business opportunities but in the end, they are tremendously damaging not just to the firms that make or rely upon them in terms of reputation, but to the industry as whole.