Industry Analysis From Primary Sources: Put a Number on Each Force

Buyer power: HIGH

Picture the slide. Five boxes in a cross. Each box has one word and a color. Rivalry: HIGH, red. Threat of entry: MEDIUM, amber. Supplier power: LOW, green. Underneath, in small type: “Source: industry report; team analysis.” The steering committee nods. Nobody asks what HIGH means, because nobody on the slide could answer.

Michael Porter, who built the framework, would have asked. In his 2008 restatement of the five forces, he walks through how each force shows up in an industry's prices and costs, and after each one he asks the same two words: “How much?” ¹. Strong suppliers drive up input costs. How much? Buyer power lowers prices. How much? His instruction is plain: “analysts should look at industry structure quantitatively, rather than be satisfied with lists of qualitative factors” ¹.

Industry analysis is the work of explaining why the companies in an industry earn what they earn, and how that will change. Done for a decision, it answers a narrower question: what will the structure of this market do to the returns on the move we are considering? For the framework itself, see Porter's five forces.

The thesis of this page: an industry analysis is finished when each force that matters carries a number someone else can check. Most of those numbers are sitting in primary sources (government statistics and company filings) that are slower and uglier than a paid industry report and far harder to argue with.

Adjectives are free. A number has a source.

Draw the boundary around the decision

The first mistake happens before any data is gathered. Porter puts it bluntly: “Many strategy errors emanate from mistaking the relevant industry, defining it too broadly or too narrowly” ¹. Too broad, and you blur differences among products, customers or regions that decide profitability. Too narrow, and you miss the links across related products or places that drive advantage ¹.

He gives two dimensions to settle: the scope of products (is motor oil for cars the same industry as motor oil for heavy trucks?) and the scope of geography (is competition contained within each state, national, regional or global?) ¹. His test is structural: if two products have the same buyers, suppliers, barriers to entry and so on, treat them as one industry; if the structures differ a lot, treat them as separate ¹.

In practice, start from the decision. A company deciding whether to open three branches in one metro area needs a local industry. A company deciding whether to build a plant that will ship nationally needs a national one. Write the boundary in one sentence at the top of the analysis, and list the products and places it excludes.

Then expect a mismatch. Government statistics are organized by industry codes (in North America, the NAICS codes), and your boundary will rarely match a code exactly. Note the gap and decide which way it biases each number. That note is part of the analysis, not a footnote to it.

A map of the primary sources

These are US examples; most national statistics offices publish equivalents. Each answers a different question, and each has a catch.

Paid industry reports and databases are useful for speed and for a first hypothesis. The buyer guide to industry research databases covers the options. Treat them as secondary: ask for their method, and check their headline numbers against the primary sources before anything goes on a slide.

Turning the numbers into forces

Porter names several elements that can be quantified directly: the share of the buyer's total cost that the industry's product represents, which tells you about price sensitivity; the share of industry sales needed to fill a plant or run a logistics network of efficient scale, which tells you about entry barriers; and the buyer's switching cost, which sets what an entrant must offer to win a customer ¹. The table below extends the idea to each force.

A number alone is not the analysis. Porter's point is that each force works through the income statement: rivalry lowers prices or raises marketing costs, strong suppliers raise input costs, and so on ¹. The analysis is the chain from the number to the margin. “Four firms hold most of the sales” means little. “Four firms hold most of the sales, output prices have risen faster than input prices for five years, and none of them has added capacity” is a claim about why margins are high and what would end it.

A worked example

The company and every figure below are invented, to show the sequence. Real figures would carry a table reference for each.

Notice where the analysis ended: on a specific question that only fieldwork can answer. Public data narrowed the decision from “should we enter?” to “which of two independents has the denser routes?” That is what a good industry analysis does. The hypothesis-driven page explains why ending on a sharper question counts as progress.

Growth, technology and government: three things that look like structure

The most expensive errors in industry analysis come from mistaking a visible feature for the structure underneath. Porter singles out three ¹.

• Growth. Fast growth can mute rivalry, but it can also hand power to suppliers and draw in entrants. He notes that some fast-growing businesses, personal computers among them, had been among the least profitable industries, and calls a narrow focus on growth “one of the major causes of bad strategy decisions” ¹.
• Technology. Advanced technology does not by itself make an industry attractive. Mundane industries with price-insensitive buyers, high switching costs or scale-driven entry barriers are often far more profitable than glamorous ones that attract competitors ¹.
• Government. Porter does not treat government as a sixth force. He asks how each specific policy changes the five: patents raise entry barriers, for example ¹.

The stakes are large. Between 1992 and 2006, average return on invested capital across US industries ranged from zero or negative to more than 50% ¹. An analysis that gets the structure wrong is not off by a little.

One more trap belongs to senior teams. Porter warns against using the framework to declare an industry attractive or unattractive rather than to guide strategic choices ¹. The useful output is not a verdict on the industry. It is a statement of which force is setting the profit level, and what your company could do to shift it or position against it.

The case for the paid report, and its limits

The fair objection: primary sources are too slow and too coarse. The Economic Census describes one year in five and takes years to publish in full ². Government industry codes lump together products that compete with each other and split products that do. A paid industry report is current, specific to the market you care about, and ready by Friday.

All true. The answer is to use each for what it does well. Use the report to form a hypothesis and to find the names. Use primary data for the numbers that carry the recommendation, because those are the numbers a skeptical executive, an investment committee or a regulator will check. Where the report and the primary data disagree, do not average them. The disagreement is a finding: either the boundary differs, or one of the sources is wrong, and you need to know which before you commit capital.

There is also a question of what you are paying for. A report's market size is only as good as its method, and many do not show one. A census table comes with definitions, coverage notes and a release schedule. When the two are close, cite the census. When they are far apart, the TAM, SAM, SOM page shows how to rebuild the number from parts a skeptic can test.

Monday: replace one adjective

Open your team's latest five-forces slide. Pick the force that matters most to the decision on the table, and replace its adjective with a number: a concentration share, a price trend, a share of the buyer's cost. Cite the table. Then write one sentence on how that number reaches the margin.

Before the slide goes out, run it through the analysis QA checklist. Then show it to the most skeptical person you know. If they argue with the number, you have started a real conversation about the industry. If they argue with the adjective, you were never having one.

Questions

What is industry analysis?

Industry analysis explains why companies in an industry earn what they earn and how that is likely to change, usually by examining the competitive forces: rivalry, threat of entry, buyer power, supplier power and substitutes. Done for a decision, it asks what the market's structure will do to the returns on a specific move, which means defining the industry around that decision.

What primary sources can I use for an industry analysis?

In the US: the Economic Census for size and concentration, County Business Patterns for local establishments and employment, the Producer Price Index for output and input prices, the BEA input-output accounts for who buys from whom, and company filings on SEC EDGAR for what rivals say about competition. Most countries' statistics offices publish equivalents.

How do you quantify Porter's five forces?

Attach a number to each force that matters: firm counts and concentration shares for rivalry, the share of industry sales an efficient plant needs for entry, the product's share of the buyer's total cost and buyer concentration for buyer power, input prices against output prices for supplier power, and relative price and share trends for substitutes. Then trace each number to margins.

Are paid industry reports good enough?

They are useful for a fast first view and for finding names, but treat them as secondary. Ask for their method, and check the numbers that carry your recommendation against primary sources. When a report and government data disagree, find out why before relying on either; the gap often reveals a different industry boundary.

Where to go next

• Competitive Analysis Template: A Rival Profile You Can Score
• TAM, SAM, SOM, Done Honestly
• Porter's Five Forces: How to Apply the Framework
• Barriers to Entry
• Industry research databases: buyer guide