3 steps of agile sense-making

This is a brief post to remind myself of ideas from “Agile Sense-Making in the Battlespace”.  So what relevance do weary soldiers have for those of us back home?

Getting started: the jargon

Agile

We spot some technical jargon immediately.  Computer people like “agile”.  What they mean is doing just as much as we can to be able to get some feedback.

Imagine it this way, when we begin a journey from London to Edinburgh, we ask the SatNav for a route and then we tend to assume the journey will then be pain free.  Often the journey is not and the real outcome involves looking for another route in a mild panic when the inevitable happens and we are diverted.

The alternative is a SatNav that works like this.

  • It finds the best route and first junction about 15-20 minutes away
  • As we reach that junction, it quietly rechecks the route taking into account any weather and traffic information that has arrived since then

Agile is simple getting on with the first task but allowing that the overall route to the destination may change.

Sense-making

Sense-making is best understood through the SIR COPE acronym of Karl Weick.

Sense is not about truth.  Sense is about piecing together whatever information we have so that there are no discrepancies and so that we are willing to ‘stay in the game’.

Sense-making is an ongoing process; it is a confusing process; and it is ultimately a social process because a key factor in our decision to stay or go is our judgement of the people around us and their loyalty and commitment.  In military terms, it is ‘morale’ – do I even want to belong to ‘this man’s army’?

What can we learn from weary soldiers managing the battlespace?

So jargon aside, what new does William Mitchell add in his description of thinking clearly in the battle space?  I will be using my words now because these are my notes.  I hope you find them useful but if you do, check back to the original article.

#1 Think imaginatively

Technically, we call imagination in “thinking about the systems of systems” or in Mitchell’s words “network philosophy”.

In practice, we think like this:  I want to attract more customers to my business.  They either don’t know I exist, or barely pay me any attention, and when they notice me, don’t trust me.  I want to win their trust.

Of course, I can woo them directly and sometimes I will.  But they already have relations among themselves.  So when I woo the fellows who, say, wear hats, the fellows who don’t wear hats don’t want to take part.  That second level effect is systems thinking.

When we are busy, or in goal mode, our systems thinking tends to get turned off.  Let’s go back to driving from London to Edinburgh.  When I set my SatNav and I head out onto the motorway, I know the trip will be boring, so I don’t want to know about all the wonderful places I could visit just 5 miles off the motorway, or I will not stick to my task.

But I also don’t know about the inter-schools football championship that is about to disgorge a flood of cars into the junction ahead of me.  That’s what management intelligence is for. To make a system that scans for the opportunities or threats that we aren’t scanning for, and should not be scanning for, because we are in executive-mode and concentrating on something else.

But the key takeaway is not that we have lookouts.  The key takeaway is that we have lookouts how understand second order effects what causes what.  And for there to be any point to having intelligent lookouts, we need managers who understand the messages from lookouts.  That’s why managers must be fluent in systems of systems thinking.  They must be able to follow the briefings and ask the right questions.

 #2 Write things down

Technically, we call this state “iterative modelling”. We write down what we think to build a bridge from our brainstorming to our action.

In practice, we log our interactions with potential customers and we see how well we are doing.   We calculate our open rates and click through rates and sales.  We use numbers to focus our attention on what must be done and to learn how to do what we do even better.

Very simply, when we drive from London to Edinburgh, part of the system is written down for us. The SatNav is doing the map calculations for us using a straightforward A* algorithm and some detailed information from maps.  Then it presents it on a map annotated with voice commands.

We do the rest. We look at our clock.  We note the time to destination on the SatNav.   And we note what time we ‘must’ arrive and make our decisions accordingly.  We can see immediately that SatNavs are going to become much, much better at learning.

There are several skills involved in modelling dynamic information.  We have to know what to model. We have to capture data.  We have to write programs of very many sorts.  We have to lay out information.  And we have to learn, a lot, about how to make the whole system better.

And in that morass of work, we might forget what all this is about: to bridge the dynamism of systems about systems thinking with action that has to be taken in some instances, in a split second.   This is what we are doing this for!

#3 Look at alternatives

Technically, the third stage is called “hypothesis generation and testing” or “scenario planning”.  Oh my, how we hate to do this when we are in the thick of action!  To be goal-oriented means to be confident of what we are doing.  And we resist any undermining of our confidence including thinking about what else might be a good idea!

But snap decisions are dangerous and unwise.  A good MI system delivers the correct information to make choice at the right time.  We slow down thinking to speed up work – or avoid false starts and over commitment to unwise courses of action.

Let’s imagine, for example, that we are very attracted to selling big ticket items to wealthy customers.  And that we are reasonably successful.   But that our smaller items fly off the shelves in our ‘outlet’ shop around the corner.   Now imagine we have a choice: spend the next hour serving the high value customer, or spend the next hour helping move the queue around the corner.  It’s helpful to have a display that shows our two choices and their consequences so we can make the choice in terms of what we will achieve and not simply our personal preference.

Equally, when we are driving from London to Edinburgh and we are diverted, in the time we have to reroute, it would be helpful to have a display that shows the best 5 choices rather than requires us to step through them painfully – a task that cannot be done until we find somewhere to pull over.

Every MI system has assumptions built into them. And though we use the systems in a very trusting way on a day-to-day basis, we should know what those assumptions are and what information we are not seeing.  Yes, the data must come packaged ready for action. But we must have people in the background looking at alternatives and produces displays for those too.  Caveat emptor: If we rely on computer systems that we don’t understand and don’t insist on getting better and better, then we only have ourselves to blame.

The three steps of Agile Sense-Making

So this is it agile sense-making –

  • Think imaginatively (imagine the side-effects)
  • Write things down to bridge imagination to action (a computer program counts as writing things down)
  • Have alternative programs that bring together analyses in different ways (our methods must learn)

This is the new world of management consulting folks – data driven.  Now let’s find the clients to match!

CHECK OUT SIMILAR POSTS

Rational data analysts have to consider motive first

Winners by Ann Morgan via FlickrDo the leaders of the world consider hard cold data when they make decisions?

I was sitting on aeroplane, catching up with the executive of an houshold name mulitinational, when he said something that was either the befuddled product of a jet-lagged brain, or something quite profound.  I know he is capable of both, so see what you think.

He said that he had learned, that “at this level  [buying and selling business units across borders]” the final decision just depends on the person making the decision.

Is there a general truth in this observation?

It is true, that like the son of a rich man, a firm can seem to “do well” just because of inherited wealth.  If we compare like-with-like and mentally subract the advantage of licenses and capital, is it true that at some point, organizations stop making rational decisions, and revert to the prejudices, preferences and whims of decision-makers?

Is the world ultimately run on whim, concealed by the size of major organizations?

Does data and rational decision making really matter in this complex and fast moving world?

I doubt we would ever have the hard data to systemaically explore this speculation. Wee may as well put it aside.  But another question does suggest itself.

If the world is puttering along on the basis of whim and prejudice, maybe we should stop worrying about being rational?  Maybe we should just suck-our-thumbs and join in? Put two stones in our pocket and say yes or no depending on which one we pull out?  Should we worry about having data at all?

What is the purpose of data?

It seems to me that we data pundits may have created a “straw man”.

Skilled leaders know that every decision is, at the last, personal.  Their decisions are their choice and the choice of the people they represent.  For them, the important order of events is not data and then preference.  It is preference then data.

Data is not there to tell people what is rational.  Data is there to follow through personal choices rationally.

Do we always use data sensibly?

There are, of course, foolish leaders who discard data when it does not endorse their naïve preferences.  The are inept leaders who discard data when it is ‘politically impossible’ to convey bad news to the people they represent.

More enlightened leaders, and I’ve known a few, use data to learn more about the group’s preferred possibilities and more about ways of achieving their preferences.  The enlightened leader becomes more conscious about what they need to negotiate, with whom, to secure their preferences.  The enlightened leader becomes more aware of what to look out for as they enact their preferences.  The enlightened leader becomes aware of side-effects that might undermine their preferences once the achieve them.

Data is useful to them but not when we try to tell them what they and their followers want out of life.  Data is useful to them when we elaborate their values and help to understand more fully what they want to do.

What is the role of a data pundit?

In short, I suspect data is always welcomed by enlightened leaders when it helps their mental model become

  • fuller
  • more flexible
  • more oriented to the outside world
  • and more supportive of informed, sensitive engagement.

A concrete example

I think Iit’s pretty much like taking a long haul flight with several segments across the world.  If I need to be in a certain city by a certain time, a rational analysis of options is pretty useless.  I am only interested in my options for achieving that goal.  And if that goal becomes uncertain or impossible, then I want data that allows me to re-formulate my goal and communicate to all the people who will be affected.

Leaders are rational; they are motivated

This seems obvious but we expect leaders to be more rational.  They aren’t rational.  They are motivated.

Sometimes we think that is wrong.  But following that reasoning,  it would be equally wrong to want to be in a certain place by a certain time.

The questions facing data pundits are

  • which motives do we enjoy serving?
  • and which data puzzles  can help to solve?
  • Executives often seem satisfied to support massively important decisions with sub-standard data.
  • Our data often seems good.  Yet they throw it out.  We jump to the conclusion that they are too motivated to follow a course of action without a rational analysis.
  • The reality is that we may have done the wrong task.  We may have interrogated a question that interests us rather than people we serve.

So why did I bother to write this post?

We shouldn’t be dismissing leaders as being willing to proceed without data.  They’ve been forced to proceed without data because we didn’t make data available to address the problem facing them today.   We might think they could make better use  of their time and effort but it is likely they would use data that elaborates what they value should we make it easily available.

Motive first. Analysis second?  First, be clear about the motives of the leaders and people we serve?