Strategic framework: set the question, the people and the rules before you analyse

Norvane Building Services, a fictitious installer of heating, ventilation and electrical systems with 35 staff, had watched its operating margin slide for three years. The management team booked two afternoons and started where most teams start: a SWOT. By the end of the second session the flip charts held 60 points, the discussion went round in circles and nobody could say what had been decided. Things only moved once the managing director, Ines, wrote down the question the team actually had to answer and set a strategic framework around it.

That is the job of step 1. A strategic framework does not settle the strategy itself. It settles the question, the boundaries, who takes part, who decides and how. It takes a few hours, and it saves weeks of analysis that leads nowhere.

How to write a strategic question

Richard Rumelt argues that good strategy starts with a diagnosis: an explanation of the challenge you face. A strategic question is that diagnosis turned into a choice. It names a decision with consequences for money, people and customers, and it can be answered within a few months. Compare these pairs.

Work with two or three questions at most. Norvane's question became: "Should we keep bidding for new-build projects, or concentrate our scarce technicians on maintenance and retrofit of existing buildings, and what would that shift take?" Every later step was tested against that sentence.

How predictable is your environment, and can you shape it?

Martin Reeves, Claire Love and Philipp Tillmanns of BCG showed that the right approach to strategy depends on two things: how far ahead you can predict your market, and how much you or others can change it. BCG later added a fifth approach for companies under pressure. The approach you choose shapes the process itself.

Many small firms work in the classical way by default, even when their market is moving fast. Norvane chose a mix: classical for its core of maintenance and installation work, adaptive for anything that depends on changing energy rules and subsidies. Where the future is genuinely open, add scenario planning to the external analysis.

Set the scope and the room to manoeuvre

• Subject: the whole company, one business unit, one market or one product group.
• Horizon: usually three to five years; longer in capital-intensive industries, shorter with fixed review points in fast-moving markets.
• Fixed limits: what is not up for discussion, such as ownership, a maximum investment or the location.
• Free space: what is explicitly open, so that nobody holds back a good idea because they assume it is off limits.

Treat mission and vision as a working draft at this stage. The analysis may give you reason to sharpen them later.

Who takes part and who decides?

SWOT was designed as a participative method. In the original approach developed by Robert Stewart, all managers were asked to write down 8 to 10 key planning issues for their own unit (source: Long Range Planning, June 2023). Involving people improves the analysis and the support for the choice, as long as everyone knows in advance who makes the final decision.

Rules of play against thinking errors

Groups make predictable mistakes: they anchor on the first number mentioned, defer to seniority and look for confirmation. Dan Lovallo and Olivier Sibony call an absence of dissent a warning sign, not a sign of agreement. Agree these rules before session one.

• Criteria first: decide how you will judge the options before anyone has a favourite.
• Individual input first: everyone writes down their own view before the group discusses it.
• The decision-maker speaks last: so the room does not simply follow the boss.
• A premortem before the choice: Gary Klein's method, in which the team imagines the plan has already failed and lists the reasons why.
• Different perspectives at the table: at least one person who knows the customer, one who knows the numbers and one who will say what others think.

The strategic framework on one page: your process plan

Put the framework on a single page and share it with everyone involved. This is Norvane's.

Fictitious Firmfocus example. Amounts in euro; the logic works in any currency.

Pitfalls in step 1

• A question with no direction: "How do we become better?" cannot be answered, so every analysis seems relevant.
• Too many questions at once: five questions produce five half-finished analyses.
• Planning in a market you cannot predict: a detailed plan for several years is of little use when rules or technology change every year.
• Boundaries nobody stated: an option that is ruled out in session four, because the owner never wanted to sell, wastes everyone's time.
• Assuming a vote: if participants think the majority decides, they lobby instead of analysing.

With the framework in place, the analysis can start. Step 2, the external analysis, looks at what is changing outside your business. For an overview of all steps, see the strategy process in five steps; Rumelt's view on diagnosis is worked out further in what good strategy is, and every other topic is in Strategy.