Advanced Analytics for Agile Organization Design: 5 Principles

Organizations are more than a list of people. Agile Organization Design is much more than redrawing the org chart. Analytics are central to understanding the organization and designing its future. But organizations often end up with the wrong data and ineffective analysis in preparation for organization design. As a consequence they pay the costs of getting their design wrong. In this article, I propose a number of solutions, including: New data models, better analytics, non-defined space and feedback loops are needed for organizations to get organization design right.

The situation

Any large organization faces the need to change from time to time. Efficient routines are the heart of a good business, but when the outside world changes, organizations have to adapt. Customers demand new products, service offerings and activities.

The business has to deliver changes to routines, skills and people. Sometimes its people leap at the chance to take on new challenges and learn new skills. Other times, the organization must let people go & rehire for other skills.

The complication: Agile Organizational Design

Designing a future organization is challenging, for five good reasons:

• Imperfect information: You don’t know what you’ve got
• Complexity: The future organization is a complex system
• Bundling: People are a bundle of Skills; Roles are a bundle of Activities plus Behaviors
• Immeasurability: Costs are known; potential benefits are hard to estimate
• Change: Even the right answers are only temporary

What are these challenges, their consequences, and how should you address them? We will discuss them one by one.

1) Imperfect Information: You don’t know what you’ve got

First, how can this be? The organization’s managers and auditors are meant to control how physical and intangible resources are allocated, measured and tracked. Yet the reality is that most organizations have only an approximate idea of their resources.

Errors creep in from imperfect record keeping, and also from continuous evolution. As new demands emerge, new roles or departments serve them and the old definitions of cost, headcount or FTE by department are no longer accurate. We were not surprised when one major financial institution said its operational headcount was ‘between 60,000 and 85,000’ – because the definitions used in different parts of the business were continuously shifting. At the same time, the actual work that is required can change – and processes can evolve without a formal record being kept.

What’s needed here? Every organizational design project has to decide what data to extract from multiple existing systems, and what extra data to collect about its people. A pragmatic approach is recommended here to data collection: get the basics right, and only seek further information if it really adds value. By the basics, I mean, an accurate list of current people and an accurate list of future roles. Too many organization design projects are undertaken without even this basic data present.

2) Complexity: the future organization is a complex system

Every organization is a complex system. It adapts to its environment with multiple elements, multiple interconnections and feedback loops. The elements include:

• People – who is employed, for how long, at what pay rate, past performance
• Roles – job descriptions, grades, spans of control, responsibilities
• Activities – work definitions, level of detail, strategic importance, risk levels attached
• Skills – formal qualifications, technical, social, personality types
• Objectives – clusters of objectives, priority order
• Systems – especially IT systems

When you change an organization, we sometimes say, you intervene in one dimension, but have an effect on many. Removing a ‘layer of management’ seems attractive, but too often, people then have to be re-hired because their particular skills, or particular processes were not transferred. Or ‘redesigning our customer journey’ seems like a great change to make – as long as the implications for systems, roles, people and support functions are well worked out.

Systems thinking implies the analytics needed for organization design sometimes take place on one dimension (e.g. roles and their costs), sometimes on another (e.g. activities or skills) and sometimes on the interrelations between them – the fragmentation of work across roles, or the clusters of skills needed for activities. Sometimes not knowing how to analyze these complex relationships causes trouble.

In a voluntary redundancy situation, people suspect trouble ahead, but they usually have no method for calculating the service-level impact from the combination of skills of the people who choose to accept redundancy packages.

What’s needed here? A choice from the organisation as to whether it’s ok to design the future based on headline ‘roles’ only, or whether specific activities, systems and skills need to be thought about. You wouldn’t design a building without water pipes, or knock out a wall or two without considering the consequences. If activities, systems and skills are important, model them.

3) Bundling: People are a bundle of skills; roles are activities plus behaviours

People are the saving grace of modern society. The bodily engrained habits of activity, movement, location, and behaviour lend a layer of stability that would not be provided by rational economic entities.

As entities in organizations, people are not, thank goodness, fully divisible or interchangeable. You cannot swap in or out a pure ‘skill.’

The fact that multiple skills are bundled up in human beings has its counterpart in the bundled idea of a role. A role is also a most peculiar combination. It is more than a task; it is more than a set of tasks or even deliverables. Its very name is theatrical: to take on a role is to play a part. That’s why the role is a bundle of activities and behaviors. Behaviors are descriptions of the style of action, and are used when the precise nature of the action cannot be defined.

Why is this important? Understanding roles as ‘activities plus behaviors’ is important because it indicates that precise design of future work will not be possible. We need ‘behaviors’ precisely because the future work we would want cannot be defined now. We want people to fulfil a role – and that might require improvisation. A little or a lot, we can’t be sure. But it’s a role we want, not just activities and deliverables. We know that we don’t know. As Nordtsrom’s employee handbook used to say ‘Rule #1: Use best judgement in all situations. There will be no additional rules.’

What’s needed here?

An analytical approach can be applied at the stage of delivering a new organization design. At this stage, with well-defined roles, you can use recent developments in personality data collection to propose people for positions.

Historically, personality data has been slow to collect and unreliable to use. But now accelerated surveys are offered by companies like Ixly and Cut-e, and game-based profiling are offered by companies like Arctic Shores and Pymetrics. These generate a ‘fit profile’ of applicant, and claim to improve predictions of the suitability of each candidate to each role. This is especially appropriate for jobs with ‘discretion’ over delivery, where you might recruit for personality, and train for skills: such as service roles, sales roles and developers.

HR 2025
Competency Assessment

How future-proof are your HR skills? Take our 10 minute assessment to find out!

Start Free Assessment

A less analytical response may be more appropriate when roles are still to be defined, or are still evolving. In this situation, it makes sense to talk again and again about the culture of the organization and the behaviors that bring it to life. Question it. Burnish it. Make it your aim that people should talk proudly in their retirement about the magic of having worked together in your organization, and what they achieved for each other, their customers and their community. And – as no human organization lasts forever – be prepared for it to have its day and to hand over to the next organization, as long as its values are even better.

4) Immeasurability: Costs are known while potential benefits are hard to estimate

It is a lot easier to calculate recent costs than future potential benefits. This gives a dangerous inherent bias in organization re-design work in favor of cost reduction. Spans of control and layer reduction are classic design interventions which risk being over-used because they are easy to measure. There is still value in benchmarking and comparisons of organization’s structures, but the cost optimization story has its limitations: no organization started from nothing and then cut costs…

Certainly, you should use the data – there will be all kinds of nuggets hidden in the information. A great principle is ‘catching people doing things right.’ One of our clients identified teams half the size on one site of those on another but producing the same volumes of output, using better working methods.

At another client, high accident rates were correlated both with individuals working too many hours – and too few hours. Best practices can quickly be shared once identified. So analytics are vital, but it is also vital that data-led interventions are balanced with the need to create an incompletely-defined creative space.

We are all aware that the future is continuously being created – sometimes by an individual, but more often, by people working in groups. Michael Lewis’s book The Undoing Project describes beautifully how creative interaction drew more from two people than they ever knew they were capable of themselves. People working in groups define new products, new services, new concepts and new ways of working. By doing this they create new value, and that value itself cannot be fully known in advance, even by the people themselves.

The lack of control of one’s own creativity is of course, why it is vital for artists to have agents.

What’s needed here? For organization design, it is important to lay the foundations with core analytics, benchmarking and cost reduction, especially in well established, standardized processes.

On top of this, organizations must give people permission to explore, with processes and behaviors that protect them and believe in their ability to grow. The organization has to take the best view on benefits that are not yet known, while continuing to collect information and to adapt quickly to any signals received. It must liberate the entrepreneurial animal spirits to act, and manage the fine art of being wrong in an Agile fashion.

5) Change: Even the right answers are only temporary

First, even in a defined strategic landscape, other players may yet have to decide their plans. For example, European retailers now look at Walmart and Amazon and try to work out what is coming. What is coming will depend partly on what they expect. This is a recursive loop of feedback that cannot easily be modelled.

Second, value is determined interactively. Some goods’ value depends on them being relative goods. Some depend on whether they are perceived as being valuable. Larger and larger parts of the economy (arts, brands, luxury goods, consumer goods, service experiences) are created in this uncertain way. As Keynes said, in business life, the necessity for action compels us to behave ‘as if we had a good calculation of advantages and disadvantages,’ but the reality is often that “We simply do not know.”

Valuations are temporary, reflexive and may be unpredictable. There have been times when 20 bitcoins were needed to buy one Gucci handbag, and when one bitcoin could buy 20 Gucci handbags. It may not be meaningful to ask what is the ‘correct’ value for either.

What’s needed here?

Because right answers are only temporary, the organization has to get the customer’s best current view of the opportunity, and – in the Agile model – adapt the organization’s activities and roles continually (e.g. every two weeks). The great entrepreneurs are not the ones who bet on the ‘right’ vision, but those who build an organizational willingness to invest, monitor, identify issues fast, and adapt.

The last two points – immeasurable potential plus continuous change – naturally fit together. If you can create, but not know the value in advance, your org design needs an innovation selection mechanism to ‘support or shelve’ (but learn from!).[1]

The 5 Principles of Analytics for Agile Organization Design

So what kinds of analytics do you need for organization design? Taking the five challenges and analytical responses together, let’s try to gather some recommendations that can be included in your design strategy.

The analytics we need for great organization design are therefore data-led, multidimensional and adaptive.

You need (1) hard analytics of what is well understood and definable. (2) The analytics must cover many dimensions while understanding that (3) some aspects are undefined – which is why employees must keep talking about the vision and culture. (4) Your analytics must calculate some investments in time and space for the unplanned to emerge, while (5) analyzing the real time signals of which ones to back.

People sometimes say of a job definition ‘it’s not the name on the box, it’s what’s in the box that counts.’ I find this too ambitious (do you really imagine that you can define what’s in the box?) and too restrictive (are your definitions all that could be in there?). Yes, we should analyze everything we possibly can. At the same time, however, Agile Organization Design strategy has to be humbler in its analytics and more ambitious in its culture. It’s not just what’s in the box, it’s also what could be. How do we give it its chance?

And it comes from saying no to 1,000 things to make sure we don’t get on the wrong track or try to do too much. We’re always thinking about new markets we could enter, but it’s only by saying no that you can concentrate on the things that are really important.” Steve Jobs interview Business Week, 2004

[1] As Steve Jobs said:  “Innovation comes from people meeting up in the hallways or calling each other at 10.30 at night with a new idea, or because they realised something that shoots holes in how we’ve been thinking about a problem. It’s ad hoc meetings of six people called by someone who thinks he has figured out the coolest new thing ever and who wants to know what other people think of his idea.