In a previous tutorial (How to Size a Market - Part 1), we showed you two methods for sizing a market. Here, we continue with two other methods that can be used to size a market for Business-to-Business (B-to-B) and Business to Consumer (B-to-C) companies.

MARKET BUILDUP

This method is most suitable to B-to-B companies. You will see this method used in many business plans presented to venture capitalists.

Here's a step-by-step approach to the Market Buildup method. Notice how we first break down a market into smaller units, estimate the size of the smaller units, and build the estimates up into a larger estimate. That's the idea behind this method.

Step 1

First you want to take the market, broadly construed, and break the market up into reasonably defined and identifiable categories.

For example, if you were trying to size the B-to-B market for general business services, you might use the government statistics known as the SIC (Standard Industrial Classification, now known as the North American Industrial Classification System) that you can find in any library or on their website. You could use categories such as manufacturing, construction, retail, etc, or the sub-classifications (e.g., construction is sub-classified into heavy construction, special trade contractors, etc.).

If you were trying to size the market for Internet measurement services, for example, you might use such categories as E-commerce and content providers, Enterprises, ISPs (Internet Service Providers), etc.

Companies selling to consumers (so-called B-to-C) can also identify categories, such as demographics (income, zip code, etc), but these are a rather arbitrary. There is a better way to handle consumers that we discuss below.

Step 2

With the categories well defined, the next step is to break down the broad categories into finer detail, say by the size of the company (or the number of employees in the companies) in the various categories.

For example, there are small, medium and large ISPs, or manufacturing companies with under 100, 100-1000, and over 1000 employees. Precisely how you break down the various categories is dependent on what makes sense (that is, there is no right way, but there are ways that are more sensible than others).

Next, with the finer categories determined, you would then identify how many companies fall into the various sub-categories. For example, determine the number of small, medium and large ISPs, or manufacturing companies with varying levels of employees.

Step 3

Determine the average potential demand for your product or service for each firm in the sub-categories. For example, you might determine that the average small ISP would purchase only 40 units of what you sell. By multiplying the number of ISP's in that sub-category by the average demand in that sub-category, you can determine the unit market potential for that sub-category.

Finally, you simply aggregate up to determine total market size. If you can estimate the average price for what you sell, you can apply this to determine that dollar market size.

That's all there is to this method...the difficult, as usual, is getting the details. But that's where the library comes in. A good reference librarian can help you get any of the information you need, whether it's about old or new economy companies.

MARKET FACTOR INDEX METHOD

Selling to consumers requires a slightly different approach to this problem. This is partly because it is often difficult, if not impractical, to categorize consumers into easily identifiable ways. Of course, if you segmented the market correctly, you could use the segment descriptors as a basis for classifying consumers, but often this gets messy very quickly when you are asked to get an overall estimate of the market size.

The approach often used with consumers is to identify market factors that are highly related to market size, and then combine these market factors into a weighted index. The method is often called the market factor index method.

While the idea of factors and weighted indices may sound complicated, they are easy to understand.

Step 1

The basic idea is to identify any observable factors that may be highly correlated with the market size. Such factors are typically of a demographic nature, like age, income, etc. So, for example, if you plan to have a web site that sells luxury items, you might say (and you'd be right) that the sales of such items are highly correlated with the income that people have.

Or, if you sell boats, the sales of boats are highly correlated with the percentage of the population that live near water.

In these two examples we have a single factor that is presumed highly correlated with market size. Let's continue to explain this by just focusing on a single factor.

Having determined that sales is highly correlated with this other factor (high income or live near water), you can then estimate sales by determining how many people high income or live near water.

Notice what you're doing. You're estimating sales via an observable factor that is assessed as highly correlated with sales. This is the key to this method.

Of course, often there are several factors that may be highly correlated with sales. In this case, a weighted average of these various factors is used to estimate sales.

So what are the other factors that may be related to sales? Well, as we said before, that depends on what you're selling. But with data in abundance these days, you can find tons of things that might be related to market size (zip codes, size of neighborhood, etc.). These may be related to market size, but often the correlation is very weak. Remember when you're using this method the key is to first identify factors that are demonstrably highly related to market size.

So when some company is trying to convince you that thesy can put together a model that has lots of factors in it and that together they will tell you the size of your market remember they are simply using the market factor index method. Make sure all the factors that go into their model are truly highly correlated with sales, else they are just selling you a "bill of goods".

FINAL COMMENTS

First, remember that all these methods are methods of estimation, so there are no "Right" answers, just ways of approach the problem that is more logical and defensible than others.

Second, many methods found in our Hitchhiking Guide to Forecasting can also be used to estimate the future size of a market, including time-series analysis and leading indicators. These are the methods used by big Internet research firms like Gartner, Jupiter and Forrester. You can be a knowledgeable consumer of their research by going through our guide.