Your Engineering Team Does Not Need More Priorities. It Needs a True North

Ask an engineering team what matters most and you may receive six perfectly reasonable answers: ship faster, reduce incidents, improve accessibility, pay down technical debt, modernise the stack and support the next commercial launch.

The problem is not that any of these goals is wrong. It is that a list of good intentions does not help when they compete for the same people.

A useful engineering strategy needs a direction that survives individual projects. That is the practical value of True North: a description of the condition you are trying to move towards, rather than another target to squeeze into the quarter.

A direction is not a dashboard

"Increase deployment frequency" is a measure. "Move to a new framework" is an initiative. Neither explains what better looks like for customers and the people maintaining the service.

A web engineering team might choose a direction such as: customers can complete important journeys reliably, and engineers can change those journeys safely without depending on a single specialist.

That statement is deliberately broader than a quarterly objective. It connects user experience, delivery capability and organisational resilience. More importantly, it provides a basis for choosing between competing investments.

Make the direction concrete enough to disagree with

Vague aspirations such as "engineering excellence" allow every proposal to claim alignment. A useful direction excludes something.

If safe, independent change matters, a feature that introduces another undocumented dependency deserves scrutiny. If reliable customer journeys matter, a polished internal dashboard may be less urgent than a recurring checkout failure.

Ask the team what it would stop doing if the direction were taken seriously. If nobody can identify a trade-off, the statement is probably too comfortable.

Separate the destination from the next experiment

Long-term direction should not become a distant ideal that excuses present problems. Translate it into the most important current obstacle.

For example, suppose releases routinely wait for one engineer. The obstacle is not a lack of ambition. It is concentrated knowledge and deployment authority. A sensible experiment might pair a second engineer through several releases, improve the runbook and test whether the dependency falls.

That is a more credible strategy than announcing a complete platform transformation. It creates evidence before increasing the commitment.

Use measures as tests of your reasoning

Choose a small set of observations that reveal whether the obstacle is changing. For the release example, track how many people can deploy independently, how long changes wait and how often a release requires emergency help.

Include customer evidence. Faster deployment is not progress if broken journeys increase. Equally, a low incident count may reflect a team avoiding changes rather than operating a healthy system.

Measures should challenge the story leadership wants to tell. Their purpose is to help decide what to do next, not to decorate an update.

Give everyday decisions a common language

The real test arrives during planning. A product request, security update and test improvement all need the same engineer. True North should help the team describe the consequences without pretending one category always wins.

A time-critical commercial feature may be the correct choice. Record the risk being accepted and when it will be revisited. Strategic alignment does not mean maintenance always outranks product work.

It means trade-offs become explicit rather than being settled by whoever applies the most pressure.

Keep improvement connected to real work

Small teams rarely have a separate transformation department. The same people maintain the application, investigate production and deliver features. A strategy that assumes unlimited improvement capacity will fail before the architecture does.

Attach improvement to work already happening. When changing an important route, add the missing operational signals. When onboarding an engineer, test the setup documentation. When resolving an incident, remove one repeatable source of recovery delay.

This is not an excuse to avoid dedicated investment. Some problems need protected capacity. But improvement should not depend entirely on a future quiet quarter.

Review the direction without rewriting it every sprint

Check the current obstacle regularly and the direction less frequently. If every new stakeholder request changes True North, it is not a direction. It is a live transcript of organisational anxiety.

Equally, do not protect a statement that no longer fits the business. A major product change, operating-model shift or new customer obligation may justify revisiting it.

A practical starting point is one page: the desired condition, the largest present obstacle, the next experiment, the evidence to review and the person responsible. Discuss it with the team before turning it into a leadership slide.

You do not need Toyota's scale to benefit from clearer direction. A small web team in Dundee faces the same basic problem as a much larger organisation: more worthwhile work than available attention. True North earns its place when it helps people choose, learn and improve without needing a new strategy every Monday.

Comments

Popular posts from this blog

The Concept of True North in Lean Methodology: A Toyota Perspective

The Archetypes of Staff Engineers: How to Excel and When Your Business Needs One

Serverless Architectures: Simplifying Backend Development for Scalable Apps in 2025