10 Rules Of Market Sizing

Published by TomFrancis on

10 Rules Of Market Sizing

The demand for market-sizing work keeps growing.  So much unstructured, “flat” information is published online that clients have almost come to expect a readily available answer to any question they may have about their business environment.  But, except for a few well-tended areas of human activity, there is actually no one out there systematically noting, counting and reporting on what is going on.  In this mismatch between expectation and reality comes the growing demand for market sizing.

As an exercise, it takes many forms.  At the simplest end you have clients who don’t know how to use basic reference tools or how to interrogate a database.  These situations are simple: you give them the answer, hopefully teach them a trick or two, and charge little or nothing for the little or no effort you have made.

At the most complex end you have clients who want you to perform Holmesian miracles of deduction and masterpieces of spycraft – that’s after you have mastered the particular technologies of bioengineering or quantum computing that they are concerned about.  With clients like these, you can hope to rein in the fantasy and scope out a workable project – or of course, charge a fee that is high enough to make the whole adventure worthwhile.

But it’s in the middle that the real work lies.  Typically, the client will want to get a better fix on some commercial area that is a little too small or specialist for published sources to throw direct light on.  For example, if you want to know how much concrete is used across Europe per year, you can go to Eurostat.  But if you want to know how much goes into building schools across Europe, you have some work to do.

Market sizing is a craft, not a science.  When properly conducted, it is a dialectical process through which, by the end, you have built a model which resolves the apparent contradictions in the mass of disparate data that you started from.  Since every exercise presents its own features, and since it’s an inherently open-ended process of conjecture and refutation, there can be no neat formulation that spells out how to do it.

But here, for your consideration, are ten rules that we have found useful:

1.   It’s a campaign, not a stroll

You are not going to get straight from an initial set of data to an answer.  Some data will be wrong; some causal processes will be unknown; your actual understanding of the question will contain some errors or at least imprecisions.  So keep documenting where you think you are and where your line of reasoning is taking you.  You are almost certainly going to have to retrace your steps several times.  You’ll probably need to move around the existing information until you have a better sense of the topography.  All of that is impossible unless you leave a retraceable trail.  You can draw diagrams, or write it all down, or model your reasoning in annotated spreadsheets – it doesn’t matter how you do it as long as you have usable records.

2.  Triangulate

This follows on from point 1.  You should start from extremely disparate sources of data.  You should find at least two routes to your target estimate (and hope that the two answers are not out by an order of magnitude).  By definition, you are working in an area full of uncertainties: one single line of argument backed up by a narrow base of evidence is simply not adequate.

3.  Trust no one

Companies lie about their market share.  They have numbers to prove that their products are the best-value, greenest, highest-performance.  Consumers boast about how well they chose or how badly they were treated.  Even governments lie.

It’s vital to ask yourself why each particular entity makes its particular claims, because you are going to have to judge between conflicting claims from different entities.  You will hope to find a way of weighting for bias rather than outright rejecting everything that a source has to say.  But at all events you can’t just pass other people’s claims through on the nod.

4.  Nobody’s perfect

The demand for numbers outpaces the supply of number-makers.  Journalists make mistakes.  Government officials make mistakes.  Everyone makes mistakes.  However good the source may look, if you really, deeply can’t accept one particular number from that source then consider the possibility that you should throw that one number away (or find a plausible correction).

5.  Everybody’s lazy

It’s very, very easy to copy things out and repeat them.  Just because the internet says it 500 times over, that doesn’t mean it’s true.  Check your sources.

6.  Ask “what is it?” before you count it

You’re running around, picking up nuggets of information from here and there, poring through large datasets from institutions.  And it’s all covering the overlapping topics that are your concern.  Well, obviously, you need to coordinate all of these sources.  Stick all of those numbers together.

But before you do, you must check who is counting what.  When Source A says “10% of UK companies”, do they mean exactly the same by “company” as Source B does?  Chase the source down, read the methodological annex: it is absolutely vital to know exactly what the units are that each source is actually giving numbers for.

7.  Trash can be valuable

You usually need to do more than just look for published information and ask people questions.  You need to look at the unconsidered traces people leave, and draw conclusions from what they do rather than what they say.  That is all free information, and is often very valuable.

8.  When you can’t find it, build it

You are going to come across some intermediate questions that simply don’t have readymade answers waiting for you to discover them.   Well, if you have to have some number as an input for the next stage, just go ahead and build that number.  Start with ceiling and floor values that you are certain the number sits between.  Work by analogy, keep looking for nuggets of information.  You will find the floor rising and the ceiling getting lower.  Don’t just sit there looking for exactly what you wanted.  It isn’t there.

9.  Anything might come in useful

Market sizing is open ended.  What you personally once saw on the street could be useful.  An economic analysis of eighteenth-century China could be useful.  Simple microeconomic common sense is almost always useful.

10.  Show your working

You are highly unlikely to end the project by just delivering your estimates to a grateful client.  Something will be surprising, or unwelcome, or even wrong.  As long as you have a model that makes causal sense, and that has its sources properly documented, you are sure to be able to get to a successful final result.  You just need to sit down for the final stage of the dialectic – the client review.  And for that to work you must provide the inputs and the process as well as the outputs, in a properly organised and accessible form.