Strategy and structure: squaring a necessary circle

Strategic decision making – is it history?

In 1977, the historian Alfred Chandler of Harvard Business School published a seminal book on the history of strategic decision making at the highest levels of American firms, including General Motors, DuPont, Standard Oil and Sears Roebuck.  Of these, GM and DuPont remain strong businesses.  Standard was broken up as in illegal monopoly in early 20th century although its progeny,  Exxon-Mobil and Chevron, continue to thrive.   Sears struggles as do so many retailers in the face of the storm called Amazon.  The book is called “The Visible Hand:  The Managerial Revolution in American Business”.  From the book comes a maxim that I believe still rings true.  It is attributed to Alfred Sloane, one of GM’s founders.  The maxim is, “Structure follows strategy”.

In this fast-moving age, many commentators argue that being too strategic renders an organisation sclerotic.  Google “is strategy dead?” and you get millions of search results.  Articles claim formal strategizing subdues innovation and responsiveness.  However, some sense of purpose, direction and objectivity is required otherwise the business and, more particularly, its people are operating in a vacuum.  As the founder of Ritz-Carlton Hotels, Horst Schulze, remarks, there needs to be sense of “true north”.

If strategy is alive and kicking, is its sibling structure still breathing?

So, let’s assume and accept that devising strategy remains necessary.  To execute it, resources are needed and the humans ones need some co-ordination.  What structure works in the second decade of the 21st century and will continue to work in future decades?  There are countless options; from pyramidal, bureaucratic hierarchies, e.g. when I started work in 1977, Barclays Bank had 21 managerial layers(!), to the constantly changing shape of holacracies (see Tony Hsieh, CEO of Zappos, now part of Amazon), via a vast array of hybrids in between.  I’m not a fan of matrix structures; dogs with two masters are confused too easily.

Structures should achieve two things.  They describe the organisation’s shape from end-user customer through to the C-suite, and the relationships between functions, operations and divisions.  Structures that don’t provide a clear “line of sight” to the end-user can easily wane into dysfunctional silos; perceived sovereign entities bounded by distinct functional walls.

Secondly, any structure should contain concisely and precisely defined roles.  Role-profiles that wax-lyrical over a dozen pages tend to be wrought with ambiguity, contradiction and imprecision.  My favourite was from an NHS Hospital in southern England for its Director of Patient Experience.  Apart from the word “patient” appearing in the title of the role in the header at the top of each page, the word wasn’t used in the main narrative until page 11.  So, what was this role about; it certainly wasn’t the patient?!

Are structure’s children, role profiles, healthy?

These “profiles” should set out the role’s purpose, its lifespan (because no role is for ever – Tuckman has added disbandment to his team formation steps of forming, storming, norming and performing ), its goals and objectives, which must relate to corporate intent otherwise what value is it adding?

A role can be defined to be entirely disruptive, i.e. it has been created to innovate, challenge, provoke and stimulate discussion about foundational principles.  Consider how Walgreens started out as US restaurant business and now operates as one of the world’s biggest pharmacy businesses owning Boots in the UK.  Or that Nokia went from manufacturing Wellington boots to becoming the world’s largest mobile telephone manufacturer (before suffering a Sears-like fate at the hands of Apple and Samsung).  It now seems to be doing okay, see this article from The Economist,

Who reports to who?

Additionally, and critically, the role should set out the impact in time of the decisions expected of the role-holder.  This keys into the important but now, sadly, somewhat disregarded work of Elliott Jaques.  He defined the reporting relationship between two roles should be determined by that time factor.  The CEO makes decisions that have impact running out many years (although reward systems do not necessarily reflect that and the reducing tenure of CEOs in their roles suggest they are increasingly regarded as tacticians rather than strategic leaders).  A team leader whose team members coordinate AI to devise a sequence of marketing initiatives makes decisions that have impact of far shorter duration (although, of course, a wrong decision with a marketing campaign can have massive long-term damage on a brand; who remembers Hoover’s free flights offer?).

Structurally, the CEO shouldn’t report to the team leader.  However, at the human level of the respective role-holders, isn’t the CEO be duty-bound to take heed of her team leader’s counsel when they express viable and credible opinions or ideas?  Isn’t that democratic or servant leadership at play?

Lastly, a role profile must describe the qualities required of the role-holder in terms of skills, knowledge, experience and behaviours (more so than personality; organisations pay for the former not the latter).  In aggregate, all should reflect the organisation’s values.  Lack of cohesion between an organisation’s values and an individual’s is a recipe for sub-optimal performance.

All words but no action

And all of this can be boiled down to mere words on paper.  What really matters are those behaviours that encapsulate how people think, engage with each other, achieve things and manage their emotions (back to Walgreens and Jim Collins concept of Level 5 leadership – humility and fierce resolve – which he applied to Walgreens founder as he navigated its move away from the hospitality sector).

Conceiving and executing strategy remains important.  Conception isn’t just about setting a desired destination, it also establishes a firm’s current “position”.  Without accurately knowing that the first step on the “arc of travel” could be in entirely the wrong direction.

Some structural arrangement is necessary so that the people occupying the roles in that arrangement know what is expected of them.  But this isn’t a fixed mechanical or physical set-up; it is a flexible and adaptable biological one, see

Even in that team running AI systems mentioned earlier, biology prevails because the AI code has been created by the human team members, so is probably imbued with all their complex biases of which there are many, see illustration below.

Perhaps Sloan’s maxim should now read, “Structure follows strategy, and behaviours rule them both” (with apologies to JRR Tolkien’s “ring verse”).


This site uses Akismet to reduce spam. Learn how your comment data is processed.