Tell me what you mean by “supply chain research”.

We cover telecommunications technologies, so for us supply chain research means eliminating the old distinctions among components, networking, and services analysts. Instead, in each report we look at the entire supply chain from chips and components to systems and services.

Why?

Because these distinctions among analysts and report categories don't make sense anymore. In the late 90s, when you had 20 components or services companies that had just received a VC round, it made sense to do a competitive analysis on who would survive. But as the telecommunications market has advanced, some of the most significant developments have occurred in chip and component technologies, while others have occurred in industry buying trends, and there have been some dramatic changes in how networking systems are sold. No analyst can ignore any one of these, so you have to look at the design, development, and sale of systems and services in their entirety.

Can you provide an example?

One of the most important microeconomic developments in the broadband industry has been the dramatic decline in DSLAM port prices, and increase in density from 1 port to 48 ports per card. Advances in DSL transceiver technology were an essential part of making this happen, and would have been missed if you looked at DSLAMs solely from a systems perspective, or DSL solely from a services perspective.

Then why not just look at components?

Because you also have to study the service provider's and network equipment manufacturer's end user markets to understand the demand side fully. Clients today want more depth than say, just the growth rate for corporate video, they need much more detail, something like the growth rate of videoconferencing among medical centers with more than 400 beds. Additionally, going to this level of detail allows you to create forecasts of broader markets than can be broken down in order for clients to target sales campaigns.

Where do you get the data to do this?

You need to find the metrics within major industries that determine spending, and then test them repeatedly. This requires analyzing the end user industries extensively. So if you're looking at shipping for example, you need to figure out the important metrics within that industry, and then most importantly, where the break points in spending on network services and equipment occur. In that case, it would be the TEU (container) capacity of the ship, not the revenue of the parent company. With hospitals, it's usually a certain bed count, with private jets, it's typically the takeoff weight of the plane, and so on.

Tell me a little more about your forecasting methodology.

An important aspect of forecasting is finding data sources that have demonstrated predictive value. Component price declines, for example, have a strong track record of increasing demand for network equipment ports. So they're one good source. Surveys are not. They work well when you're trying to determine which features to add to a mature product, so they're good for packaged consumer goods, but for developing technology markets, we've found that they're exceptionally weak at predicting demand.

Why?

Much easier to look at what determines usage than to have people tell you what they will or won't do hypothetically.

Do you forecast across a variety of scenarios?

No!  We forecast one scenario, and publish that. Saying outcome X has a 50% chance of occurring, outcome Y a 40% chance, and outcome Z a 10% chance makes it harder for clients to act on your on conclusions, and leads to even more “what if” scenarios that create more indecision. Market research is supposed to reduce uncertainty, not add to it.

But then how do you plan for unforeseen events?

We cover telecom technology markets, we're not forecasting the future Fed Funds rate, next year's GDP level, or what the global money supply will be in two years. So if you look at a major unforeseen event, like September 11^th, it had a major impact on these global and national macroeconomic measures, but it didn't do much to slow the decline of Ethernet port prices.

Also, you have to keep in mind the objective is not to forecast the future perfectly to the last dollar, but to do so within an acceptable range so that a client can rely on your conclusions to make sales, investment, or product development decisions now.

Forecasting is still seen by many people as an imprecise waste of time, why?

Many executives are compensated based on profit or revenue forecasts, business development managers are often paid based on sales forecasts, and forecasting is the foundation of most budget and planning activities. So it's very important, regardless of our view on the matter. But we've found the 12-24 month period is often the most important for our clients, and anything over five years is usually viewed as silly speculation. We've never had anyone tell us that the boss really needs a 2017 market outlook. Telling someone that you have one only adds to the skepticism some people have about forecasting.

Do you take vendor briefings?

We do on occasion, but we don't use them as a primary data source. We appreciate PR executives who facilitate good debates about markets and technologies instead of throwing a bunch of Powerpoint slides at us, but we get information on network configurations from end users.

Sounds like a lot of data collection and analysis, is it costly?

The reason the economics work for both us and our clients is that we don't do any custom research, everything goes into a report. This allows us to spread our costs over multiple clients, and gives our clients the sort of depth they normally expect from a $50,000 single client project in a $2,000 - $3,000 report.

You've talked a lot about numbers and forecasts, but how do you communicate your conclusions to clients?

In plain English. We cringe when we hear a series of verbs that end in “ize”, the word “value” used as an adjective instead of a noun, silly phrases like “best-of-breed”, “first mover”, or even just overused ones like “disruptive technology”, they're irritating. Our objective is to inspire firm decisions, so we try to stick with firm language.