This is a summary of a talk I gave this week to the Stuttgart Social Forum at the Bosch R&D Campus in Renningen, and you can view the accompanying slides here.

What is the use case for organisational structure?

The rise of social technology in the workplace creates new possibilities for how we organise work. The early use cases for Enterprise Social Networks (ESNs) and other social tools focused on improving collaboration, knowledge sharing and communication, but they have also brought some much needed lateral connectivity to previously silo-based organisations, and the question is how can we build on this to create more effective organisations?

Without changes to organisational structure, culture and practice, new tools cannot create better business on their own, but these platforms can be seen as a starting point for a future organisational operating system. In considering how to move towards the organisation as an OS – as an enabling platform for different business models, products and services – I think there are some useful lessons to be drawn from how software itself has changed in the past decade.

We consume a lot of research and analysis into the future of the firm in different sectors, and have never seen so much interest in how to reform slow bureaucratic company structures, and yet few companies understand where to begin. From professional services to industry and manufacturing, this is now an urgent issue for firms facing growing startup competition and fast changing market conditions. Anti-social organisational structures are adding huge cost to the management of modern organisations, producing a poor experience for employees and customers and standing in the way of the adaptation and evolution these organisations so badly need.

Even startups are not immune from creating these structures as they scale. As David Manheim points out in his excellent essay Go Corporate or Go Home, the need for legibility of structure also leads many fast-growing startups back towards classical hierarchy, although I would guess that this is often a response to investor preferences.

Do we even need traditional organisations when we have so many startups throwing themselves at business problems? Software is taking over the world and is expanding its domain to be everywhere, but it will never be everything for as long as we need gas turbines, electricity generation networks and the big industrial elements of the physical world. And that is where the great opportunity exists for traditional industrial firms – IoT and the Industrial Internet are creating amazing opportunities for traditional firms to augment ‘big iron’ offerings with software, data and intelligence. In theory, it should be easier for industrial firms to evolve software capabilities than for software startups to develop big industry and manufacturing abilities.

But one of the great challenges is the internal structure and culture of these industrial firms. Conway’s law suggests that companies are destined to produce systems whose internal structure mirrors that of their own internal communication processes. But also the internal divisions between competencies such as hardware, software, data and user experience emerge as visible seams on the products they produce. To make connected products, and to serve connected customers, it is essential to become a connected company. There really is no other way. But over-specialisation and optimisation for the divisional company have created a structural legacy that acts against connectivity.

The other obvious need is for more fluid sharing and co-creation across disciplines and product lines. Some of the greatest innovations in the history of manufacturing have been examples of combinatorial innovation rather than invention – putting existing components together in new and innovative ways. In a world of hardware, software, data and experience design, we have a much wider canvas to paint on when remixing and re-combining elements to make a greater whole.

But rather than just bolting on new skills and methods to existing out-dated company structures, I believe we need to think differently about the purpose and evolution of companies themselves.

If we go back to Ronald Coase’s theorem about why we aggregate economic actiivity in companies, the question is do we still need them? Do they still give us capabilities that could not be achieved by other means? What are the use cases for the organisation’s structure and practices? How do they add value to the work of employees?

If we begin to think about the organisation’s structure in this way, rather than accept it as given and work around its limitations, then I think we can start to build a clearer picture of how it should evolve.

Platforms as the new bureaucratic systems

It is important to acknowledge that there was originally a positive case to be made for bureaucracies, by people like Max Weber, in comparison to what came before. In a bureaucratic system, each person knows their place and they have certain defined roles and rights that protect them against the vagaries of powerful leaders. There is also a degree of legibility within the system, so the structure is more navigable than one where power structures are hidden. They were both human and inhuman at the same time. In other words, the system and its processes have primacy over people, but the whole thing is managed and populated by people, pushing paper around and enforcing rules. The problem today, however, is that in an age of networks, bureaucracies struggle to adapt and to compete.

We have written a lot about the need to humanise organisations, and others also talk about wanting to move from an organisation-as-a-machine philosophy towards a more organic view of the organisation. Perhaps counter-intuitively, I think organisations need to become more and less human at the same time in order to achieve this goal. In other words, replace bureaucratic functions with a core platform that uses technology, bots and even AI to create an operating system that enables people to focus on what matters and to work more freely and autonomously. If we consider that a key goal for smart technology is to support augmented human intelligence, rather than just to mimic human intelligence with AI, then I think platforms and the idea of an organisational operating system will be an important part of how we achieve this.

So, going back to the notion of an organisational operating system, what can we learn from software to help us chart a path for organisational improvement?

Service layers beat vertical integration

Software used to be a vertically integrated hot mess of hidden or opaque services within a single design and user interface, and every level. from data storage right up to the interface, was baked into the system. Today, software is built up in layers, from data and services all the way up to its interface(s), and we usually start by building the underlying services and then join them together to form a product. This makes changing things a lot easier, but it also means a single lower level service or process can serve multiple end user applications.

Platforms beat products

It makes more sense for an organisation to maintain a single, conservatively managed shared services platform to provide all the common services required, and then allow individual teams or business units freedom to create their own apps on top, than it does to maintain a one-size-fits-all vertically-integrated software stack that forces all parts of the organisation to tolerate a lowest-common-denominator inflexible product, which is traditionally how IT functions have tried to meet the many and varied “requirements” of their business.

As Dave Gray argues in his Connected Company book, this is a better way to organise previously powerful central services like HR, IT, etc., by turning them into part of a service-oriented architecture for the organisation. This is similar to how we understand the relationship between platform and apps in mobile ecosystems, where we cannot interfere with the base platform, but have a lot of freedom to build single-purpose, specific apps on top.

Algorithmic management: real-time data beats retrospective reporting

Perhaps we can go further and seek to embody the organisation’s wider ruleset in such a platform to reduce the need for conventional policy management and guidelines. I expect that we will see examples of algorithmic management emerge in the next 5-10 years, where the organisation’s policies and rules are embodied in the shared service platform that supports it, reducing the need for middle managers, who are often just water carriers and connectors, and whose reporting and decision-making role is ripe for automation, just as we used to have travel agents who would process bookings by hand, but now computerised fare rules and booking engines do the work.

Agile iteration beats top-down planning

One of the most visible changes in software development has been the shift from top-down waterfall planning to agile. Too many (arguably a majority of) major IT projects were seen to fail because of the polite fiction that you can lock down all requirements at the beginning and simply plan and execute your way to a perfect system usinng GANTT charts, project managers and brute force. In reality, of course, teams discover requirements and ideas as they move through the project, and their client’s understanding will also change over time, so rather than ignore this reality, agile embraces it. Also, by expressing target capabilities / features from the point of view of the user, in the form of user stories, everybody involved has a clearer idea of what they are doing and why, rather than just technically ticking off a feature on a long list. And by organising development into sprints, which are planned and reviewed according to emerging priorities, less effort is wasted.

Applying a similar approach to how we iterate on organisational structure and practices is likely to be a great deal more effective than consultant-led top-down change programmes. This will also change the way we perform leadership and management in important ways, as Steve Denning wrote yesterday in his commentary on HBR’s embrace of agile management:

Agile isn’t just a methodology to be implemented within the existing management framework. Agile is a dramatically different framework for management itself. In the community of Agile practitioners, which now numbers in hundreds of thousands, Agile begins with a different view of the goal of the organization.

Crowdsourcing beats surveys

The difference in usefulness between my car’s expensive “professional” navigation system and the free Waze app on my phone is extraordinary, especially when it comes to driving in London. The car’s navigation system always looks for main roads and its estimates of arrival times are always wrong, whereas Waze uses the behaviour and experience of other Waze users to give me the best routes in real-time. Sometimes these routes are crazily complicated and commuting feels more like rally driving, but it really works.

When we are doing way-finding for the organisation, we might undertake formal surveys of employee views and issues, but we don’t yet really give them tools to point out roadblocks, traffic or accidents in real-time. Some organisations have used crowdsourcing in a limited way to gather new ideas under the label of innovation, but I think we need to go a lot further in using this for organisational improvement, as I have outlined previously.

For me, this is about creating a more sentient, self-aware organisation that is able to use collective intelligence to find its way.

Conversational interfaces and chatbots (might) beat enterprise search

If we take a look at some near future development in software, I think these can also be mapped to the way we build and run organisations. For example, the rise of chat bots and AI suggest we might be able to operate aspects of the organisation through conversational interfaces that do the hard work of search, retrieval and joining the dots to support individual works and teams. So, if I am in my work chat channel and type a command to initiate a project or a task, it would be nice if a chatbot told me that somebody else in another department was doing something similar and offered to connect us, for example.

Getting the conversation started

Whether you work in a communications function, an embryonic digital transformation team or another function that involved with internal collaboration systems, then you can begin to use your Enterprise Social Network or Social Intranet to kick start a conversation on organisational improvement. For example, you could:

  • Discuss the use cases for organisational structure – how is it supposed to support the work of employees?
  • Define target capabilities that teams want from the organisation to be more agile and responsive.
  • Based on these, write agile user stories for the organisation to guide improvement.
  • Decide how to measure these target capabilities (crowdsourcing + data).
  • Encourage teams and departments to come up with ideas about how to create the target capabilities in a shared agile backlog.
  • Each month, ask each team to prioritise a small sprint of transformation actions (from small changes to behaviour or processes right through to new organisational features) and measure the results.
  • For new organisational feature requests, decide which are core and should be considered part of the platform, and which are specific ‘apps’ for individual departments or teams.

If you want to find out more about how we help internal communications and other teams get started with this kind of agile transformation, please contact us and we can share some real-world examples.