The solution I believe is in the bit of 1970s thinking that has got lost.  Kenneth Andrews talked about “matching” opportunities and capabilities.  While we have done lots of thinking about both opportunities and capabilities, strategists have done little on “matching”.   This is implementation and change management and … the sorts of things that strategists do not get very involved in.

But what if “matching” is the key ingredient? How does matching occur?   Simple observation suggests that matching is achieved by the entrepreneur or manager or intrapreneur.  Matching requires someone who understands the opportunity and has influence over the capabilities and can bring the two together.  With this view, strategy development could be more about identifying individuals capable of doing the matching, than about exploring opportunities and capabilities.  These people are likely to be much rarer and harder to find than opportunities or access to capabilities.

This suggests a different focus for strategic planning processes.  The new focus would be on identifying managers inside or outside the organisation who have the ability, energy and insight to find some new or improved match.   Strategic plans would then be focused more on defining the “strategy experiments” that are being sponsored than about laying out the five priorities for the next five years.

The typical organisation chart does not tell you much about how the organisation is supposed to work.  Normally, charts are drawn in layers: all the people reporting in to a boss are on the same level and so on down the structure.  Because there are multiple layers, the chart often extends to multiple pages.

If an attempt is made to distinguish between people reporting in to the same boss, it is normally done using pay grades or titles.  So that those with an SVP title are higher up the chart than those with a VP title and so on.

This simple format is partly a result of the simplicity of the computer programs used to construct the charts.  But, it is also a result not knowing a better way.

A good organisation chart can achieve the following:

-       explain who reports to whom

-       communicate information about the business model of the organisation and

-       define the relationships that should exist between boxes in the chart.

The operating core

The first step in drawing an organization chart is to define the operating core; and to decide how this operating core is divided up.  There are three ways to divide up the operating core: by function, making each box a cost centre; or by business unit, such as product, market segment or geography, where each box is a profit centre; or by matrix, where profit reporting is made on more than one dimension.

The operating core of a single restaurant would be structured as functions: for example buying, cooking, serving and marketing functions.   The operating core of a chain of similar restaurants would be structured into business units, with each restaurant as a profit centre.  The operating core of a group of restaurant brands would also be structured into business units.  Each brand would be a separate business. The operating core, therefore, may be different at different levels in the structure.

If the operating core consists of functions the relationship between these functions is that of a “tight team”.  Profit can only be calculated once the activities of all the functions are combined.  Hence the functions need to work closely together to ensure their combined activities are achieving the objective. Monday morning meetings or something equivalent are needed to ensure regular coordination between the functions.  This also has implications for the role of the leader of these functions.  He or she will be “hands-on”: frequently involved in easing tensions, deciding priorities and guiding action.

If the operating core consists of business units, the relationship between the units will be much looser.  Coordination does not need to be as close: each can operate with a good degree of independence.  The relationships between business units can be quite distant.  Monthly or quarterly meetings may be sufficient to achieve needed coordination.   Also the leader of the collection of business units is likely to be “hands-off”.   Because he or she is not needed to ensure coordination on a daily or weekly basis, much more can be delegated to each business unit within the constraints of defined policies and strategies.

Above the operating core

Once you have clarity about the operating core at any level, the next challenge is to define any parts of the structure ‘above the operating core’, yet still at the same level.

Obviously the boss or leader sits above the operating core in the organization chart.   But it is also helpful to position some other activities ‘above the operating core’.

The way to do this visually is to have the operating core boxes all on the same level of the chart.  Then draw a horizontal connecting line just above the operating core boxes.  Then draw a vertical line, from the horizontal line, to the boss or leader.  Make this vertical line long enough to allow additional boxes to be added to left and right of this line .

So what boxes should hang off to the left or right of this vertical line?  Four kinds of support functions:

1. Policy functions: these are functions with policy roles, whose job is to set and monitor policies for the operating core.  Examples are accounting rules from finance, people management rules from human resources and IT rules.
2. Shared services: these are activities that could be carried out within each part of the operating core, but have been centralized for reasons of economy of scale or skill.  The relationship is a service one: the operating core is the customer and the shared service is the supplier.
3. Partner functions: these are also functions that could be embedded in each part of the operating core, but have been centralized for competence or control reasons.   The relationship is that of a partner rather than a supplier.  The representative from the function has a seat at the table of the relevant part of the operating core.  Examples are HR business partners, central research, IT business partners, finance business partners, etc.
4. Lobby functions:  these are units whose role is to influence the operating core to give more attention to some area of importance.  If the operating units all involve manufacturing, the lobby function might be lean manufacturing.  If the operating units are structured by product, the lobby function might be global accounts.  Often one of the roles of a lobby function is to facilitate coordination.  So, where a number of business units operate in one country but report in to their global divisions, there is often a country management structure whose job is to coordinate the business units in the country, but whose power is limited.  One of the features of the lobby function is that it has responsibility that is greater than its authority.   It is expected to influence, but does not have the power or resources to execute.

The final part of drawing the chart is to place the “bossy” functions on the left and the “service and influencing” functions on the right of the vertical line.  So that policy functions and partner functions go on the left and shared services and lobby units go on the right.

In practice, finance, HR and IT typically contain, within their empires, at least three or sometimes all four of these support functions.  This can make it difficult to decide which side of the line to place them.

There are two responses to this.  First, place them in which ever of the four roles is dominant.  Second, consider splitting the function into its separate roles.   For example, many companies are achieving significant improvements from separating shared services from policy functions.  The reason for the success is that the skills needed to manage these different support roles are very different.  The same is true for lobby functions.  They are often more successful with a separate reporting line to the overall leader.  There is, however, less evidence about the value of separating policy from partner roles, and these are often combined in functions where they both exist.

One of the entries explained the the origins of the scorecard came from Analog Devices, whose approach to measurement was reproduced in the famous HBR article.   While appropriate for Analog Devices, it is probably not universally appropriate, which explained why I have always struggled with the four dimensions.

My work on measurement has led me to conclude that the stakeholder approach is the only theoretically sound way of tackling the measurement and targeting challenge.   

The reason why the stakeholder approach is superior is because the success of the organisation depends on winning and retaining the loyalty of a group of stakeholders, such as financiers, employees, customers and suppliers, or, in my case, business school leadership, member companies, course participants, fellow directors, fellow academics, clients, Harvard Business Review Readers, etc.  Hence measurement should start with the degree of loyalty or satisfaction of each important stakeholder group.

All other measures should be treated with caution unless we are certain that the measure correlates with success.  Take waiting time for medical treatment – topical in the UK right now.  If we place this on our balanced score card, it may or may not lead to greater success.  Or take market share.  In financial service markets half the market is often unprofitable – so a market share measure can lead to less success.

Of course our strategy should guide us to the measures that matter – but this assumes that our strategy is wise.  If we do not have some measures that are independent of the strategy, how will we know when the strategy is wrong or needs changing?

So, start by defining all the “active” stakeholders (those who can influence success).  Then look for measures of satisfaction for each stakeholder.   Some stakeholders will not be good at giving accurate feedback (think of subordinates with regard to a boss) – so you may need to find surrogates for “satisfaction” (such as employee engagement).  Then scrutinise the strategy and look for additional measures that will record the progress of the strategy.   Then stand back and consider a universal measure “productivity” or “resource untilisation efficiency”.  More success will always come from achieving more with fewer resources – so consider ways of measuring either the amount of value created for a given input of resources or the amount of resource used for a given output (e.g. cost per unit or profit per employee).

These steps will give you a “Comprehensive Scorecard”.  Measures of success;  measures of strategy implementation; and measures of resource utilisation efficiency.

Strategy involves answering six basic questions:

1. What is the External environment? Strategy involves aligning the organisation’s capabilities and assets with external opportunities and threats –so it makes sense to first get a grip on what that environment is like
2. What is the Internal situation? Two aspects of the internal situation must be understood: the objectives of the organisation and its capabilities. The tricky part is that both have to be viewed in the context of the external environment. It is not enough to know that you have capabilities in R&D; you must understand how valuable they could be in generating superior products for customers and how these capabilities compare with those of your competitors.
3. How might the situation Evolve? Your strategy will play out in an uncertain future. Understanding the current external environment and internal situation is an important start, but you need to project how things might evolve and identify the greatest sources of uncertainty.
4. What is the primary Issue? Understanding the situation and how it will evolve allows you to frame the issue (or perhaps issues) facing the organization. Is it how to grow – or how to survive?
5. What are the Options? Once the issue has been properly defined, options need to be identified.
6. Which Option is Best? The final step is to pick an option – although it may be one that permits you to take different pathways in the future.

Since these questions are so important, it may be helpful to use the acronym EIEIO as an easy way to remember them. External Environment, Internal situation, how the situation will Evolve, the primary Issue, and what the Options are.

You can answer the questions using a mix of strategy concepts and frameworks and people and processes (see blog at

http://ashridgeonstrategy.wordpress.com/2010/07/06/how-do-you-create-a-good-strategy/ )

You will not need to spend equal time on every question. If the situation is familiar you may be able to jump to question 4 or 5. Or, if you are considering entering a new market you may need to spend a lot of time on question 1. You need answers to every question, but you will start with differing levels of knowledge about each. Some will require a lot of analysis – some you can jump to an intuitive answer. So, it is important to know where to spend your time, and when you can trust your gut feelings ( see

http://ashridgeonstrategy.wordpress.com/2010/05/05/when-to-trust-your-gut/ ).

Focusing strategy making down to these six questions helps demystify it and turn it into a manageable activity. Decide where you need to spend your time, and step back occasionally to consider if you need to reprioritise. At the end, of course, you still have to implement and adapt your strategy – but that is the topic for another blog!

However, an HR group-level function is difficult to hold to account.  There is no bottom line for the CEO to look at.  The customers (the managers in business units) have no ability to discipline HR if it is not delivering.  Internal suppliers of information or IT have no easy redress if the function become tyranical.   Only the managers and employees within HR have any ability to hold the function to account.  For this reason, over time the function at the group-level often becomes self-serving rather than business-serving.

The solution is to disect the group-level activities in these functions and subject them to rigorous challenge.

1. Is the activity required for governance or compliance reasons?

2. If not, is it a necessary part of some larger initiative or influence that is expected to raise performance by at least 10% above what it would have been with out the group involvement.

3. If not, is it an obvious (ie non contentious) source of extra value and low risk to implement.  Risk is assessed in three ways. Is there little risk of the activity being done badly?  Is there little risk that the activity will constrain the commercial endeavours of the business units?  Is there little risk that the activity will absorb the time of any  top managers?

4. If it fails to win a place on any of the above, the last challenge is about whether the activity is necessary because it is part of a legacy that is being wound down or part of an effort to change the status quo so that future initiatives to add value will be easier.

If it can’t be justified agains any of these four challenges, the activity should be stopped. The underlying logic is that the time of top managers is scarce.  They should not spend any time on activities unless they expect to improve performance by at least 10 percent.  If they can find three or four such activities, they are doing a great job.  Everything else should be stopped unless it is governance; low hanging, non-contentious fruit; or a legacy that is being worked out.

What seemed to be causing most pain in their thinking was the difficulty of setting priorities between different stakeholders. Yet for me the issue is pretty straight forward.

1. Some stakeholders – shareholders, bankers, customers, suppliers and employees – operating in a market and the company needs to win their support in order to be able to operate. If the company does not offer value at least as good as other companies in the market place, it will start to lose the attention of these stakeholders. So it must provide a certain miminum value to each – “value for money” to customers, returns equal to the cost of capital for shareholders, etc.

2. Some stakeholders do not operate in a market – pensioners, local community, the environment, etc. These stakeholders need legal protection – as pension law provides for pensioners. The company must do what the law says.

3. If the company does not have enough resources to meet the obligations of 1 and 2 above, it normally has to provide less to shareholders in order to given enough to the other stakeholders. So the shareholders suffer is there is not enough to go round. Of course the company may also squeeze suppliers and cut back benefits to employees and even give less good value to customers, but this is a dangerous strategy because it can undermine the company’s long term prospects.

4. It stands to reason therefore that if the company has some surplus after it has met all its obligations to stakeholders (as Microsoft has), then it is not unreasonable that this surplus should go to shareholders to compensate them for taking the extra risks. Of course companies usually also pay slightly higher than necessary management salaries, may pay their suppliers a little earlier and negotiate less strongly on prices and even give their customers slightly better value, not to mention investing in community relations, endowing a university chair and supporting a charity. However the prime beneficiary of surplus is and should be the shareholders who are taking the risk when there is not enough to go round.

So stakeholder theory, to me, seems to be a lot less complex than is often made out.

With regard to his main product line he is running a business unit. His job is to make as much profit as possible from the US market with this product. He draws on central support for product development and marketing and global account management. But he controls pretty much everything else – manufacturing, local product development, logistics, sales, local marketing. There is some confusion about product management, which is supposed to be a global function. But as he says – “they just fill out paper work. We make all the important decisions for our market in the US.”

For another product line, this manager is acting as a sales agent or distributor. He only controls sales and marketing and local logistics. All the decisions are taken at a global level and all the product is manufactured outside the US. The product is transferred with a transfer price, and he is expected to sell as much as possible at a good margin. He sets his own targets and draws on global support when it is available. He is little different from a third party distributor – which the global product line uses in some other markets.

For his third product line, this manager has two roles. He is acting as a sales agent as above. But this product line has a factory and some product development located in the US. However, the factory and product development are run by the head of the global product line on all strategic issues, even though, for pay and rations purposes, the staff report into the US. The head of the US’s role here is one of running a shared service (providing management services) for the global product line. He provides HR services and recruitment services and factory loading services, etc. So he has two different roles with regard to two parts of this product line.

This manager’s job is complex because he as different role for each product line and even for some different functions within each product line. Maybe the job is too complex. But, he is coping – at least partly because he is pretty clear about these different roles. So long as the manager is clear about the different roles and his counter parties are also clear, it is possible to make this work.

Conclusion – clarity about roles and a language for discribing the different roles is vital to make complex matrix structures work.

In another example, a central product line depends on regionally based sales.  The product line has dedicated sales people, but they reported to the regional sales team that also has responsibility for other product lines.   The head of the region changed, because the region was underperforming.  But the central product line never thought to contact the new head of the region to discuss whether he would be likely to make any changes that could affect the product.  In fact there was no contact until the product line discovered that the sales people devoted to its product were going to be cut back and reorganised.  It was also surprising that the new regional head did not bother to contact the product line before deciding what changes to make.

The explanation for these weird behaviour patterns is that people are busy.  Because they have so much to do, they fail to invest in building networks, establishing processes and anticipating relationship issues.  They work without taking stock of the bigger picture and, as a consequence, do not give attention to “matrix” issues until they surface as problems.  Unfortunately, once a problem has emerged it is normally accompanied by bad feeling, turf issues and mistrust.  it can then be ten times more difficult to resolve.  

The conclusion I have come to is that top managers should not set up a matrix strutures unless they are personally prepared to spend a lot of time coaching and helping the participants in the matrix to learn to work together.  In fact, it might even be wise to appoint a senior manager as a full time “matrix facilitator” in the same way that it is often wise to have a full time person leading an acquisition integration process.   The challenge of making a matrix work is just as demanding and time consuming as the challenge of post merger integration.  A full time “matrix facilitator” would have time to anticipate where the flash points are likely to be and make sure that relationships, processes and structures are put in place to avert the tensions. 

There will still be individuals who find living in a matrix difficult.  But my experience is that, once they know how they are supposed to behave and they are shown how it can help them get the job done, there are very few managers who cannot rise to the challenge.

1. People. Involve the right mix of experiences, perspectives and skills. Include someone who knows the market, someone who knows the company, someone who is a wild thinker and the pragmatist who is responsible for implementation
2. Process. Customised, of course, to the people and the situation. Do you need to manage conflict – or encourage it? Do you need to push people to make a decision - or hold them back before they rush to implement a half baked plan?
3. Concepts. A powerful concept – such as “competitive advantage” or “attractive markets” or “re-framing” can drive your thinking. Make sure you are using appropriate concepts. For example, “creating value for customers” could  lead a retailer such as Marks and Spencer to one set of ideas. ”Beating Waitrose” might lead to another. Both are probably useful concepts to use!
4. Frameworks. Frameworks come last as they are largely complements to the above – helping focus a process, or capture a concept.  Organisations favour very simple and communicable frameworks such as SWOT analysis – although such frameworks often lack analytical power. There is a compromise to be struck between insight and ease of use.

Have I missed something? Let me know – I would be interested in what other levers should be on the list, and what is good practice for each lever.

PS. I am only talking here about COMING UP with the strategy. Of course, IMPLEMENTING a strategy is a different matter!

In the end he had to go the Tony Blair to get permission.  He then intervened successfully, and now Sierra Leone is a peaceful, success story.

Cut to Tony Blair saying yes the mission did change;  yes it was opportunistic;  and yes he was ‘up for intervening’ because you need to stand up to blood thirsty mobs. 

Then cut to the Brigadier saying that this experience together with the success in Kosovo made the military top brass over confident about Afganistan and Iraq.   “We did not stop sufficiently to consider the differences in these other interverntions, and how they might affect the outcome”, he said – or something to that effect.

This is a classic example of past experience providing misleading emotional tags for future decisions.  On the surface it would seem incredulous that the top military strategists would not think deeply about the differences between Iraq and Sierra Leone.  But at the level of  “are we up for this”, detailed assessment is not the main input.  It is more about confidence and will power .. and these judgments are ones that emanate from the subconscious triggered by the emotional tags stored there. 

The ex-journalist who wrote the first draft of the Iraq dossier could not recall at any point doubting the wisdom of the Iraq adventure prior to the invasion.  Why?  Even he could not understand fully how a cynical hack could ever not question things.  The answer:  Tony Blair had been right many times before on big issues such as this – hence the journalist, then communications officer for the Foreign Office, had developed positive emotional tags towards situations where Blair was leading from the front.  Hence his subconscious never stimulated his amygdala to send any warning chemicals into the brain system (even when Robin Cook, the Foreign Secretary, resigned over the issue).  Interestingly he could not even recall have a dinner party conversation or argument with his wife about whether going to war was a good thing.  The top military were presumably in the same postion – just as the Brigadier in Sierra Leone implied.  They had “good feelings” about these initiatives and focused on how rather than whether.

The same apparently happened in the board of RBS over the ABN Amro deal.  One board member said that the deal was discussed at more than 10 board meetings. But the discussion was always about how and not about whether.

Moral of this blog -  we should learn when not to trust our gut.   It will always influence us, so we should always listen to it.  But in certain situations we should recognise that it may well be giving us a bum steer.