Porter without a strategy department: three tools, one choice

Three customers in one week ask for a discount because they hold a quote from somewhere else. The owner of a small plant has no strategy department, only a few free afternoons. Where should the work start, and how far must it go before a decision is possible? This post chains Porter's five forces, generic strategies and value chain into four working sessions, using a hypothetical packaging-film printer as the example: its name and every number are invented by me.

How do Porter's three tools fit together?

Each tool answers one question, and the output of one is the input of the next. The library places five forces at industry level and the other two at firm level. The chain below is my arrangement.

The step from 1 to 2 rests on the summary of the 2008 article: strategy can mean finding a position where the forces are weaker (Porter, 2008). The step from 2 to 3 is mine.

Four sessions: who does what, and when to stop?

Each session takes a few hours and has one stopping condition.

How to score each force low, mid or high is worked through in the post on five forces in the sports-shoe market. Here I deal only with evidence and loss.

Session 1: which force takes the most profit from plant A?

Plant A is hypothetical: it prints laminated packaging film for about 40 small food makers in southern Vietnam, with revenue of 60 billion dong a year and operating profit of 3.6 billion. Resin and base film make up 45 per cent of revenue, 27 billion; long-run film accounts for 36 billion.

My evidence rule has two halves. Outside evidence shows whether the force belongs to the industry's structure or is just plant A's own problem. Internal figures only estimate how much the force takes.

The buyers row uses one internal number: twelve contracts worth 30 billion dong all closed after the customer showed another quote, and the final price averaged 7 per cent below the opening price. The opening price was 30 / 0.93 ≈ 32.3 billion, so the concession is about 2.3 billion, nearly two-thirds of operating profit.

Buyers and rivals partly overlap because both show up in price, so the ranking picks a direction rather than closing the books. Plant A buys base film from one supplier only (hypothetical), which is single sourcing: a dependence risk, but the estimated loss is only 0.36 billion. Buyers come first.

Session 2: which cell weakens the buyers' power?

Pick a position where customers find it hard to compare quotes as they would for a commodity. The library poses the question to a small-business owner whose bigger rival is cutting prices: compete on cost or on differentiation, across the whole market or in one narrow segment. The library attributes the matrix to two books by Porter, which I have not read (see the last section).

For plant A, three cells are hard to hold. Cost leadership needs the lowest cost in the industry, and the plant has one line. Broad differentiation needs a reputation it does not have. Cost focus works only if its short-run cost beats rivals', which nobody has measured. That leaves differentiation focus: serve small food makers who need runs under 5,000 metres, quick artwork proofs and delivery in five days. In my analysis, customers who need speed and help with artwork will compare every dong per metre less.

On "stuck in the middle", the library records that Porter's 1980 book argues that a firm that does not choose a clear direction will perform poorly, and asks you to treat that as a hypothesis, not a law. Plant A tests it with its own figures: of the 12 contracts squeezed on price, 10 were runs over 20,000 metres and 2 were under 5,000. That hints the plant serves both ends on one price list, but with 12 contracts it proves nothing.

Session 3: can plant A perform the activities that position needs?

Not necessarily, and this is the session most likely to surprise you. The library's value chain entry warns against drawing a value chain without comparing it with rivals, and says it is built for a business unit rather than an industry and that accounting categories such as direct labour or overhead rarely match value activities. Here the "unit" is the short-run line.

Take changeover. Plant A's books show a 3,000-metre run as profitable, because overhead is allocated by revenue. A week of machine log says otherwise: at 150 metres a minute the run takes 20 minutes, but changeover takes 3.5 hours, so 91 per cent of machine time is set-up. Assume a machine hour costs 3 million dong, the run brings in 12 million and other variable cost is 5.5 million:

• Changeover of 3.5 hours: 12 − 5.5 − 3 × (3.5 + 0.33) ≈ −5 million per run.
• Changeover of 1 hour: 12 − 5.5 − 3 × (1 + 0.33) = 2.5 million, about 21 per cent.
• Break-even: changeover of about 1.8 hours.

The nearest rival, plant B (also hypothetical), does not reveal its changeover time. Only what an ordinary customer sees can be observed:

The sales and delivery gaps can be closed with working rules. Changeover is the real gap. If it cannot be brought below 1.8 hours, drop this position and repeat session 2 with another cell.

Session 4: how do you decide, and when do you drop it?

Fix a time-limited test rather than a plan. Plant A takes short runs from 5 customers for 3 months, charges set-up separately and logs changeover on every run. Drop the position if changeover is still above 1.8 hours at the end of month three, or if fewer than 3 of the 5 customers place short runs without asking for another quote. I set both thresholds for the example.

The procedure fits if the industry can be defined, there is at least one rival to compare with, and you can collect outside evidence yourself. It does not fit if the firm runs several unrelated businesses, because each needs its own pass.

What can five forces not do?

It says nothing about one particular firm, and it does not tell you what to do. The library entry says the model describes industry structure, not one firm's performance. The summary of the 2008 article adds that a fast-growing industry is not always profitable (Porter, 2008).

For plant A, if its customers' food industry is growing fast, do not conclude that the packaging industry serving it is profitable. That needs the five forces of the packaging industry itself.

Going deeper: where do the three tools come from, and how far can they be trusted?

The summary of the 2008 article says that Porter's first article for the Harvard Business Review, in 1979, was "How Competitive Forces Shape Strategy", and that the 2008 article reaffirms and extends that work, with new sections on putting the five forces into practice (Porter, 2008). The library lists the 1980 book as another origin of the model, credits the generic strategies matrix to the 1980 and 1985 books and the value chain to the 1985 book.

I read only the summary and the opening paragraph of the 2008 article, shown free on the Harvard Business Review site, not the full text. I have not read the 1979 article or either book. Everything above about a chapter or an argument in the books is what the library records, not something I checked in the books.

Three limits follow. First, the industry boundary is drawn by the user, and a small plant tends to draw it too narrowly. Second, the "stuck in the middle" hypothesis is not tested in this post; the example only shows how to test it. Third, I have not read any source showing that the three-step chain gives better results than the tools used separately. That is why each step has a stopping rule and a time-limited test.