Organisation design:  data and complexity

The UK’s Confederation of British Industry (CBI) boss, Tony Danker noted in early September “Labour shortages are biting right across the economy. These shortages are already affecting business operations and will have a negative impact on the UK’s economic recovery.”  He said that the UK the UK needed to simultaneously address short-term economic needs and long-term economic reform.

As I read this, I thought a similar statement could be made about any large organisation.  They are likely to have labour shortages, they are addressing short term viability and performance, they are also developing longer term strategies that may take them in somewhat different direction and they are responding to a national and international context.  

To keep the organisation performing in response to context changes, the participants in it are making small adaptive changes all the time to stay ‘on course’.    Curiously though, as Sharon Varney points out in her new book Leadership in Complexity and Change, ‘In the working world, change is often described in terms of static states (‘as is’ and ‘to be’) … We have artificially separated ‘change’ from ‘no change’ and assumed that ‘no change’ is the norm.’   She rightly says, ‘organisational life is in constant motion. The bigger patterns are continuously created from numerous small changes that we may not even notice under normal circumstances.’

She asks us to ‘Imagine, for example, that we make a structure change by creating a new position

in a team and bringing someone in to fill that position. Inevitably, there will be some procedural actions required for that to happen. Yet nothing has actually changed. The dynamic process of changing begins as people make adjustments in anticipation of the new person joining the team. The new person and existing team members are then involved in the process of changing as they adapt and respond to one another in the course of their work.’  

She makes a good case for asking us to notice (devoting a whole chapter to the topic), these processes of changing, saying ‘noticing and noting is an important practice for leadership in complexity and change. It sounds easy, but it takes some skill to do it well.’  What she is asking us to notice is the small scale, human, experiential, weak signals that are often easy to ignore, but could have high import.  (I have on my wall the Jon Kabat-Zinn quote, ‘The little things? The little moments? They aren’t little.’)

At the same time numeric data of all types is collected in organisations.  Numeric data is the bread and butter of organisational life – think of all the charts, dashboards, graphs, tables, excel spreadsheets – that you come across in your organisation.  People are described by their grade number in some organisations, e.g. ‘He’s a grade 6. (See my blog At Sixes and Sevens).

The numeric data gives comfort that we know something – how many people, their skills, their ethnicity, the activities they do, their capability, how engaged they are … ‘   But, as Nate Silver says (in his book The Signal and the Noise), “The numbers have no way of speaking for themselves. We speak for them, we imbue them with meaning.”

Statistician David Spiegelhalter take this further, saying ‘We can’t just collect some data and it’ll tell you the answer. There is an art to trying to extract information, knowledge and understanding from data, and even in choosing what data to collect.

The data-centric world view and the complexity world view are usually both at play in organisation design work, and often don’t sit comfortably with each other.   Yet they each have a part to play.

Taking the data angle first.  AHIR has a cheat-sheet of 51 metrics that could be collected and McBassi has a list of 100 questions typical workforce analytics,  of the type AHIR lists, can answer, for example:

  • Our recent employee survey highlighted our lowest scores; are these, in fact, the most important areas for us to focus on?
  • Some locations get new employees up to speed much more quickly; what are they doing differently?
  • Why is employee engagement higher for some job functions than for others?

orgvue takes a different approach, ‘using data points to deconstruct people, roles, and positions … then breaking down the roles into the processes and activities – in other words, the work – alongside the skills and competencies needed to do that work.’ They say, ‘by attaching accountability metrics to each role, you can compare how effectively the work is organized.’  (Note:  I am discussing an orgvue hosted event Bridging the gap between strategy and execution, with Giles Slinger on 18 November).

But the data is not the story and neither does it answer important questions – see the wonderful piece Data Will Help Us a brief manifesto about the promise and perils of data.  It begins ‘Data will help us remember, but will it let us forget?  It will help politicians get elected, but will it help them lead?’.

Of course, numeric data collection, analysis and interpretation is useful not on its own, but in combination with qualitative data, critical reflection, complexity science and open-minded curiosity.   Sharon Varney is firm on this: ‘Charts, trends, statistics, and dashboards are always wrong, even when rigorous procedures have been applied. They are wrong because they are simplifications of a more complex reality. We can never know in advance whether we have captured the aspects of complexity from the specific situation that will turn out to really count in what happens.  We must be provisional about what we know and understand that it will never be fully right. ‘

Returning to the first paras – about labour shortages, short term viability and performance, and longer-term strategies, you can see the need for both numeric data and complexity thinking.  For example.  We know that nationwide, ‘The shortage of HGV drivers is estimated to have grown from 60,000 to over 100,000’ during 2020/2021 but locally the availability of HGV drivers varies.  Data for October 2021 saw the largest increase in Norfolk, with figures going up more than three times.  While, in Wales, the number of jobs posted nearly tripled over the period. But in contrast, North Yorkshire and Hertfordshire saw a decrease in the number of vacancies for HGV drivers.

Suppose you are a nationwide employer of HGV drivers.  You may have a long-term strategy that you would, for example:  Open up routes into HGV driving via’s a new apprentice scheme, a traineeship, a kickstart placement.  Or, change transport method to train or waterways so fewer HGV drivers were needed.  Or recruit from a wider talent pool, considering those who might otherwise be overlooked in HGV driving based on their gender, ethnicity, or background. 

You would still need a short-term strategy to overcome the specific shortages, your own numeric data revealed – likely to be in some regions/localities and not others.  You would have to take into account personal understanding – the real people with context (aptitude, attitude, career interests, background etc) not fully represented in the data. People have lives, families etc. and manager/colleague understanding of people is a critical consideration when it comes to delivering strategy.  (For example, you couldn’t simply decide to relocate HGV drivers to a shortage region).   NOTE: Both short and long term strategies involve redesigning.  The aim is not to compromise the long-term strategy in favour of the short-term one.

Sharon Varney discusses three types of data that help with designing in complexity. (See image above).   One of these is traditional numeric data.  What types of data do you use for working in complexity?  Let me know.