On strategy

I am an enterprise architect in a shop that is actively embracing agile across the board. This has given a new urgency to my previously idle speculations on the role of an architect in an agile process. I have developed some thoughts on that role that I will share in my next rumination. But there is something to get out of the way first – I want to share a spike I needed to take en route.

I’ve always believed that architecture is a strategic function for an enterprise. But when I tried to uproot that belief from the big-up-front-design soil it was nurtured in and transplant it to newly-tilled agile soil I realized I believed in something I did not fully understand.

Today’s rumination is on strategy.

Two perspectives on strategy, both drawn in contrast to tactics, seem to form the basis of our conventional understanding – one in terms of time, the other magnitude.

‘Tactics are your plans for right now. Strategy is your plan for the future.’

This makes tactics and strategy seem part of some temporal planning continuum in which at some point in time, a year out, eighteen months out, a tactical plan morphs into a strategic one.

‘Strategy is a plan of action to achieve a major goal. Tactics are plans to achieve minor goals.’

This on the other hand makes tactics and strategy seem part of some magnitude-of-objective planning continuum where at some point a tactical objective is important enough to become a strategic one.

Neither one sounds all the way right, does it? My intuition is that there must be a qualitative difference between strategy and tactics, not just a quantitative one.

In previously trying to speak to that qualitative difference – in describing the strategic nature of the architect’s role to both business and IT people to whom it was unfamiliar – I have made frequent resort to two analogies.

The first is a sailing analogy. Strategy is picking a lighthouse as a desirable destination, and charting a course; tactics is dealing with sail and wind and water to stay the course as best you can and get there.

That’s better. Charting a course calls on boating knowledge, navigation skills, an understanding of weather, time, and tide, and experience blending them. But it is still not quite right. In the real world there are no fixed lighthouses by which to guide a business. There is very little not in some state of flux.

The second is a chess analogy based on the Silman method. To figure out your next move, first mentally remove all the remaining pieces from the board save the kings and pawns. Now replace your pieces on any squares you’d like, keeping bishops on their original colors. Can you position your pieces to create a marked advantage? Based on where your pieces actually are right now, can you get them to their desired destinations? How long will it take? Can you get them there while your opponent is both trying to prevent you as well as executing a plan of his own? If you can’t achieve the desired end state you must go back and choose a different one, perhaps a less ambitious one, and keep iterating in this manner until you find a realizable plan.

Better still, I think. It calls out the need for a strategy to be practical, to be realizable. It acknowledges that there are slow-to-change features of the environment to deal with – in your mind’s eye you leave the pawns on the board while you are rearranging your pieces. And it recognizes that you are not alone out there on the water – you have, in business or in war, competitors. But in the real world even the rules of the game can change over time.

The sailing analogy is probably more accessible than the chess one. I am generally unable to leave well enough alone and go on to belabor them, talking about tacking against the wind versus running with it, throwing that your boat needs to be ‘yar’ in for good measure, or describing the Stenitzian chess trade-offs of space for time and time for material. I will continue to exercise these analogies in what follows here.

Both simple definitions have something of the truth. And both analogies resonate. Let’s try to find an underlying structure that supports them all.

To get at the true nature of strategy let us imagine a world without it. Why do we need strategy? What would we lose if it didn’t exist?

Sun Tzu points the way when he says that ‘tactics without strategy is the noise before defeat.’

While we might make some short term tactical gains, giving us the illusion of progress, without strategy our path would be ultimately aimless and reactive. Dude, even surfers have to plan to catch a wave – without a strategy surfing is just floating on a board.

We use strategy because it is smart to do so.

Smart.

There is an interesting discussion of intelligence in Steven Pinker’s ‘How the Mind Works’ (mentioned previously in my rumination on BI). He quotes a passage about what makes good science fiction. Believable aliens seem intelligent because of their purposeful behavior.

Paraphrasing Pinker, one useful definition of intelligence is the ability to achieve goals in the face of obstacles. That is, intelligence is the ability to have goals, make a plan to achieve them, and change plans in the face of obstacles.

Sounds exactly like strategy, doesn’t it?

Let’s try a working definition of our own: strategy is the systematic application of intelligence by a volitional entity – person or group – to improve their lot over time.

Strategy helps us – individually or collectively – come up with the best answer to the question ‘What should you do next?’

The simple answer is ‘something to help you achieve your goal(s).’

So how do we pick the goals? What is a desirable end-state?

First, which end-state are we talking about? That is to say, end when? Next year? Five years from now? Fifty?

All of them of course. It is clear that at some level the process of creating a strategy must be ongoing, that we must redefine our strategy as we age.

Are there goals that do not change over time? Is there a fundament?

We can start with existence. The most obvious answer to ‘what is a desirable end-state’ is ‘one where we still exist.’ This must be the fundamental goal of any entity. Some argue this is why intelligence evolved, and so, according to our working definition, why we have strategy.

Beyond survival – even I suppose including it in some supererogatory circumstances – our values, vision and purpose provide structure for all our stratagems.

Our values are our moral primitives, our moral postulates.

Our visions are our values writ large. They are unrealizable goals projected into ideal resolutions. World Peace. An End to Poverty.

Visions perform the same function as vanishing points in perspective drawings. They are virtual elements that provide real organizing structure and consistency to all that follows.

Our purpose is our highest level potentially realizable goal beyond continued existence. As businesses or organizations it is what we were formed to accomplish. As individuals determining this is called ‘finding yourself,’ which illustrates that the font of our purpose flows from within and is ultimately an expression of our values.

Once we have our purpose, how do we get achieve it?

Generally speaking you can’t do it in one big bite.

You break it down into smaller bites – you create a series of sub-goals where the sub-goals lead or build up to the overall goal. So the answer to ‘what do we do next’ becomes ‘something to achieve one or more of our sub-goals.’

This implies a linear goal structure at its simplest, and a singly rooted lattice structure – assuming a single purpose – at its most complex. This also means that to some extent our goal structure is likely self-similar – that little branches look like big branches. Fancifully we may find it has fractal dimension (remember when fractals and chaos were sexy?).

But what is the structuring model for the lattice?

And how do we find the right sub-goals? How do we find any sub-goals for that matter?

Let’s look at an extension to get a sense of this.

If your vision is Elvis Forever, and your purpose is To Let The World Never Forget The King, you may make it a goal to visit Graceland to shoot some video for your blog. Natural sub-goals might be to save up enough money to get gas for the hooptie and a buy a big SD card for your sister’s camcorder. Then, after getting gas and borrowing the camcorder, you plan to get to a series of each-one-closer-to-Memphis rest stops where you can stop and rest. Eventually you will get to Graceland. Which you will find surprisingly small and unimpressive. But you will not let that get in the way of your purpose, so you will choose camera angles that make it look bigger.

It gets a lot harder to find sub-goals when the rules are more complex than the speed limit is 65, drive on the right and pass on the left (this means you left lane campers), and the domain is larger than I-55 or I-69 and their rest stops.

We can infer a number of things from this example.

A prerequisite for goal setting is to understand the comprehensive landscape in which a future state would exist. If Graceland is going to be closed for re-velveting by the time you get there you need to come up with a new desired future state, or at least a different time to achieve it.

You must predict the future at the nexus of the domains which form the arena in which your plans play out. Our arenas exist in a kind of Hilbert space, shaped by many dimensions of different type and different magnitude. While the dimension of time is one of those structuring our goal lattice, it is not the only one. Contracts, law, regulation, supply, demand, technology, infrastructure, public opinion, politics… And those domains may not be discrete, but may be cross-connected. A given sub-goal may not straddle all of the domains.

What can be easy to miss is that we are also agents of change. This is the lesson of Steve Jobs. We can shape our own future.

Change is neither necessarily continuous, nor as we find from the example of evolution necessarily toward increasing complexity or value. Change in different domains happens at different rates, with different periods and different punctuation.

External changes may impact near-term sub-goals. Some may strike higher in the goal lattice, up to and including survival.

There are external changes to which you must react. There are those things to which you may react. There are those things to which you should not react.

There are domains that evolve slowly or not at all – the laws of physics. Domains that evolve slowly – traditions and institutions. Domains that evolve more frequently – laws and regulations. Other that evolve even more frequently – the advance of technology, style, fashion. Others that advance moment to moment – stock prices.

And it is the evolution with respect to us that is important. Physics doesn’t change, but our understanding of it does. The invention of a new packing algorithm, while unlikely, would potentially revolutionize part of the arena for shippers.

Understanding the relative pace of change is even more important to the selection of lower-level near-term sub-goals.

Like weather forecasts (an analogy borrowed from my ‘role of EA in Agile’ research), sub-goals closer to the present are more reliably chosen than those further away. We have a better understanding of the near-term arena which gets rapidly fuzzier as we move out in time. We know where the hurricane will be in ten minutes. We have a good idea where it will be in an hour. We have only some idea of where it will be tomorrow.

So as we’ve seen the best uber-strategy is to frequently refresh our strategy.

The pace of change of business in this age of technology has tended to shrink the strategic horizon because of a perception of our inability to forecast the arena with any degree of accuracy even a few years out. While some companies still plan five to eight years out, and a few ten or longer, it seems that more have fallen back to a three to five year horizon. We need to remember though that technology, while important, is only one dimension of our arena.

But the most important takeaway for me from the weather analogy is this one: your near-term sub-goals are well-defined only if you understand where you are right now. In our chess analogy it is easy: you just look at the actual board. In business it is much harder. Very few enterprises have an accurate and comprehensive idea of where they are right now. Those that do have a leg up on strategy.

Similarly knowing when you have achieved your sub-goals is necessary to working your way up the lattice efficiently. Near-term goals in particular must have associated success metrics.

Finding sub-goals sounds like a daunting task.

But ultimately the question of how you find sub-goals has a simple answer.

Remember our working definition of strategy: the systematic application of intelligence to improving our lot over time. Finding sub-goals is part and parcel of being intelligent. How do you learn to find sub-goals? The same way you learn to speak. You are built to by evolution. Singly and collectively you are a goal-finding beast. Immerse yourself or your organization into the best information you have – the best research, the best forecasts – about the churning reaction in which you swim, and think about it, and you will precipitate out goals.

If you are insecure about trusting to your innate abilities, there are a number of formal methods that can help ensure you are being comprehensive in your strategic considerations, such as the wonderfully acronymed STEER, or its woefully acronymed cousin PEST.

‘What should you do next?’ also sounds like the question answered by tactics. I have been maintaining there is a qualitative difference between strategy and tactics. If they are both answering the same question how can there be a qualitative difference?

Let’s reuse our tactic of defining something by imagining the impact of its absence.

To get at the true nature of tactics let us imagine a world without them. Why do we need tactics? What would we lose if they didn’t exist?

We would still do something. And we could still follow our strategy. But we would make a lot more wrong turns. Without tactics we would not be able to recognize and take advantage of recurring patterns in our environment. It would take us longer to achieve our strategic objectives.

Without tactics we could not force things to happen. Tactics are not just single moves, nor are they plans without branches. In chess forced sequences that inevitably result in material gain are called ‘combinations.’ I do this, then my competitors may do x or y or z. If they do x I do p, if y q, etc. No matter my opponent’s response to each of my moves, there is an incontrovertible path that wins me material.

Forced sequences don’t have to end in material gain as combinations do. They can also be a single move or a series of moves that improves our control of space, or gives us an advantage in time, or improves the scope of our pieces.

So tactics are not just random acts. They are single acts or sequences of acts that improve the quality of your position. Recognizing what constitutes an improvement means understanding the nature of advantage. And understanding which advantages may be turned into other kinds of advantage. In chess you can generally turn an advantage in space to one in time, one in time to one in material, and one in material to checkmate.

We see this transformation of advantage in business: the Amazons of the world know that if you establish market dominance, profits will – likely – eventually come.

This is a place where strategy and tactics come together. A transmutable advantage as contributed to by a specific tactic is a valid and useful strategic sub-goal.

So a wide and deep knowledge of tactical patterns is a prerequisite to best executing strategy.

There is a skill involved in being able to find instances of those patterns in the real world. Not everyone can see a knight fork or the sting of a discovered check several moves down the road with equal facility. But everyone can learn the patterns, and improve their ability to see them with practice.

Every chess player knows the strength of ‘pigs on the seventh’ – having your rooks both occupying the opponents second rank. Every chess player has marveled at Fischer’s ‘windmill’, the sweeping scythe that Byrne graciously played out for posterity.

For a chess player the ability to succeed at deep tactics is partly a function of the ability to hold those possible futures in your head with clarity. This is what the non-chess player assumes is the fundament of chess strength as they ask ‘how many moves ahead can you see?’

But computers can do that better than humans – they can see many more moves into the future accurately. So how are we able to compete with them at the highest level?

A human player’s strength is in the ability to eliminate whole branches of moves based on a well-developed aesthetic – a sense of the beauty of a position. If a branch of moves is ugly it is generally safe to prune it.

Articulating an opinion on the relative aesthetics of alternate futures based on insufficient information with which to make a deterministic decision is a function of strategy. We can choose among alternate plans when the forced sequences have not been worked out in detail.

It is also one of the functions of an architect. They (we) are almost always very senior software and systems engineers with a well-developed sense of technical aesthetics. We can and do make or recommend good decisions based on what is ugly versus what is beautiful.

The challenge of playing to this strength in the corporate world is that it is ultimately an argument from authority – an argumentum ad verecundiam, aka ‘ipse dixit’ – and suffers from that argument’s failings, so well-known that it is on the short list of common logical fallacies. ‘Because I said so’ usually doesn’t get it done unless one is the king, or carries the king’s seal. (I actually said it once in a meeting, only partly in jest. It did not go over well.) So the architect’s contributions to strategy may be underutilized. (I know. Wahhhhh. /digression.)

Selecting the best tactic is not just a function of that which best advances us to the nearest sub-goal. There is another strategic expression in place, that, like the kiddie bumpers at the bowling alley, helps keep us out of the gutter. Those are our policies.

Policies are our bible. They codify our values into guidance for how we do the things we do. They help preserve our integrity by ensuring we don’t justify our means solely by our ends.

Policies are a mechanism for our higher-level goals to directly influence our actions without that influence having to trickle down the goal lattice. They help ensure that we don’t let the temporal dimension of our goal lattice dominate.

Sometimes we find that we need to perform the same actions repeatedly. Like building a pyramid we have to aggregate our achievements over time toward our goals. These repeated actions are our procedures. They take place in the lanes formed by our policies.

Procedures have specific objectives – measurable states of affair reached at a particular time or range of time. These objectives when accumulated reach a goal in our goal lattice.

Have we gotten anywhere? Let’s sum up.

Strategy is the systematic application of intelligence by an entity to to improve its circumstances over time. Tactics is the systematic application of learning by an entity to improve its circumstances right now.

Strategy deals with the novel. Tactics deals with the recurring.

Strategy and tactics both require an understanding of the environment as an n-dimensional evolving arena in which an entity’s actions play out.

Strategy requires the recognition of potential futures and the creative synthesis of desirable end-states in them.

Tactics require the recognition of recurring patterns and the successful execution of learned responses to them.

Strategy when best done evokes aesthetics. Strategy is an art.

Tactics when best done evokes calculation. Tactics is a science.

Policies are actionable rules encoding our values.

Procedures are rule-governed repeating activities, constrained by our policies, to achieve objectives we believe will aggregate into the achievement of higher-level objectives.

To make the most efficient progress toward our higher-level goals we must must be able to measure both where we are now and when we have achieved a goal.

Our near-term strategy is always more accurate and reliable than our long-term strategy. Therefore we must regularly renew our strategy to keep its edge sharp. As our higher-level objectives change we must evolve our procedures to meet them.

Our strategy and tactics must be consistent with our values. Selecting what to do next, and next, and next again is each and every time as much a moral action as is executing to our plan.

Strategy is the intersection of reason and righteousness.

Advertisement