Business Design for an Agile world
Transforming complex organisations so that continuous change becomes a way of life
Nobody in Silicon Valley ever said that a digital service was “done”. (What would it mean for ‘Facebook to be finished’?)
Today’s digital world emerged in stark contrast to the “carrier-grade” mindset which prevailed in the established IT/telecoms industry that came before it, where systems are engineered and tested to meet high availability standards (such as “five nines”), and provide fast fault recovery through redundancy.
Services like Twitter often became overloaded due to their new popularity and suffered outages, even leading to a whole new vocabulary such as the “Fail Whale”, in reference to the screen Twitter showed when it was over capacity. In a 2013 interview we found out that the company had taken the whale out of production, as the platform was now more stable. However:
“If you come to Twitter, there are always gonna be problems, no service is ever perfect.” Chris Fry, VP Engineering, Twitter, 2013
Twitter’s response to its issues was both scaling out the infrastructure and scaling out the organisation at the same time. They improved the reliability and efficiency of the service at the same time as improving the productivity of their developers.
The digital companies whose services we all use every day are constantly reinventing themselves. Some of that we see from the ways they innovate their product, iterating and building great user experiences. Some we may never see from the outside, as they are fighting reliability fires, or putting the core infrastructure in place to become more efficient and reliable.
The digital organisations that are getting it right deeply understand the actual user experience they are offering. They fully understand their operating model, down to the nuts and bolts. And they’re organised to continuously make adaptations in response to customer needs.
The Operating Model
In government, we’re finally starting to put user needs first in the same way. We want to design (or re-design) services so that we can rapidly react to user feedback. We want to operate as a living system, which continuously reacts to all user needs, including those of our policy colleagues.
You could describe that as an Operating Model. An operating model is just the description of the way that we operate a business. And we would like the business to operate very differently from the way it does today.
The Target Operating Model (TOM)
That leads us to the idea of the ‘Target’ Operating Model, or TOM, which is simply a way to refer to the desired state of a business. A TOM is a way to describe how the business (or part of it) should operate to best deliver on its strategy and objectives. This normally includes the people and technical capabilities the business will need, how and where these will be delivered, and who will deliver them in an organisation design. It often comes with a roadmap that describes the steps to change from the “as-is” operating model to the TOM over time.
So far so good?
TOMs are one way to provide a vision for organisations undergoing change. So why are they so unpopular amongst those who are leading the government’s digital transformation? Here are a few of the reactions to Matt Edgar’s recent post on the subject:
Are TOMs intrinsically bad? One of the challenges that’s widely recognised even by their proponents is how to make sure that the operating model becomes a reality, and isn’t just a document which collects dust while the business moves in another direction.
Critically, the notion of a TOM as a target in some way suggests that the change is a finite and one-off activity. A couple of years ago, Mike Bracken explained what agile means for government and its services, highlighting the downside of one-off policy-led change as a five-step model.
He went on to make the case that in a digital world, delivery informs policy (not the other way around), and that the ramifications for government are profound.
Perhaps it is the way TOMs get used that implies they are following this old process, especially in the context of government services. Some people do value having a shared vision of the future, but have an aversion to TOM as a consulting product that later gets abandoned or doggedly adhered to. Both the consultancies and the buyers should share the blame when this happens. Some people see TOMs as a manifestation of bureaucratic, linear, inflexible change. At the same time, many do recognise a need for some sort of shared understanding of the current state of an organisation and its aspirations.
Operating models in digital businesses
So how much do leading digital organisations like Facebook, Twitter and Amazon need to define their operating models? Can Agile and TOM coexist?
Some digital role-model organisations have very small teams and little physical infrastructure of their own, whereas others have a significant operation behind them. Lifting and shifting the infrastructure and people embedded in an operational function tends to be costly and time-consuming, and often has a significant human impact. So that’s the kind of activity that organisations want to avoid, and which leads to the desire for a long-range plan, and, inevitably, a description of the desired outcomes, whether or not that’s labelled as the TOM.
What size and scale are digital organisations?
Some truly digital organisations are tiny in terms of their headcount. Facebook acquired the mobile messenger service WhatsApp for $19 billion in 2014. It was launched in 2009 by two former Yahoo employees, and in just over four years had grown to 420 million monthly users. The company had only 55 employees. WhatsApp may have been tiny, but it turned the telecoms industry upside-down — its users now send 30bn messages a day, while the global SMS system handles 20bn messages a day. I doubt they had much of a TOM.
In contrast, the big players now have big headcounts to match. Twitter has 3,600 employees in offices around the world, and 50% of those employees are engineers. Facebook has 8,000 employees.
The digital organisation that’s most relevant to my work is Amazon, because it’s not just an online operation, but also has a substantial headcount and a major physical presence, though of course as a customer you don’t normally get to see them.
Amazon now has over 130,000 employees
Amazon used to run their on-line bookstore from one warehouse in Seattle. In the UK alone, Amazon now employs more than 6,000 permanent workers and took on 13,000 seasonal workers in the run up to Christmas 2014. In the previous year’s busiest day, Amazon’s UK customers ordered 4.1m items, or about 47 items a second. Orders are processed by workers at eight giant fulfilment centres and network of regional delivery hubs, which service highly populated regions of the country that can be reached in short travel distances.
To support its business growth, Amazon had to create a radical distribution strategy. The location and size of Amazon’s warehouses have a direct bearing on how responsive and efficient the company can be, for example shipping from closer to customers for a higher service level at a reasonable cost. Import duties, exchange rates and transport costs all greatly impact the profitability of their supply chain.
Amazon has spent at least $14 billion building warehouses since 2010. It used to take two years to open a new one, but now they’ve got that time down to under a year. Or, to put it another way, they have continuously improved their best-in-class building process and it still takes a whole year to open a warehouse.
One day Amazon expect their Cloud-based revenues to overtake their ‘traditional’ retail revenues. They have datacentres across at least 11 regions and an aspiration to build many more.
Building all of this infrastructure has multi-year lead times to plan and deliver, and pouring concrete into the ground is inherently non-agile. But still, Amazon is a very agile organisation. This week they axed their digital Wallet product after only 6 months (saying, as is traditional, that they’d “learned a great deal”). Amazon evolves their product offering and the technology platform continuously. Its competitive advantage depends on it being able to iterate, and on having a clear long-term design for its organisation.
That’s the critical difference between a pure digital service like WhatsApp, and a company with a geographically-distributed operation that is working at scale like Amazon.
The Business Design we need is an ‘Evolving’ Operating Model
Should complex digital organisations invest in defining a TOM, or should they be testing and learning?
There are some government departments that may be more like WhatsApp for the purpose of the TOM debate. They should be, and in many cases now are, testing and learning as they go along.
Department for Work and Pensions
My focus is on the Department for Work and Pensions (DWP), which has 90,000 staff working across 850 buildings, 720 of which are JobCentres across the whole of the UK. DWP pays out £650 million a day in 2.8 million payments, across over 100 products and services. Whilst becoming more digital, DWP has to keep the critical day job in operation, which depends on 1,000 IT applications, some of which contain elements dating back to 1974.
Our customers depend on our services and don’t have a choice which organisation supports them. That’s a stark difference from Twitter, Amazon, WhatsApp and Facebook. Whilst it might be conceivable to have a “Fail Whale” for some of our services, other components of our service do require “carrier-grade”, and we need to recognise these in our design.
There are areas of DWP that can make great progress by taking a pure agile approach, testing and learning as they go. We’d like to see more of that, and we’re investing in the skills to help our people make that a reality. But, like Amazon, some other parts of our business are necessarily slower to change because they rely on bespoke technology, large-scale physical infrastructure and the people who operate the services.
If we were to embark on transformation across the department without some sort of joined-up vision for what our future looks like, we’d end up in chaos. That’s the user need our Business Design is there to address. To break up the overall transformation activity into manageable chunks, it’s also helpful to have a view of where we are today (though we don’t want to invest more than we have to in documenting things that will change anyway).
Our vision for DWP
Our vision for DWP is to put user needs at the heart of our thinking, delivering the policy intent through responsive digital services. We recognise that many parts of our organisation do work which is fundamentally face-to-face in nature, which makes us a truly multi-channel business. But at the same time we recognise there is an opportunity for us to deliver some of our services in a much more modern and efficient way, by collecting and using data in a more joined-up way, automating wherever it’s safe to do so, constantly improving our services and reacting rapidly to user feedback. As we become a more digital organisation, we won’t be delivering all of our change through finite programmes as we mostly do today — and there are already some shining examples of where this new way of delivering change is starting to work for us.
In our future world, the change is never finished. Our Business Design is for an organisation that continuously iterates to improve itself. Is our Business Design a TOM? If you think of a TOM simply as the description of the way that we want to operate a business, then it certainly is.
But if you see a TOM as a prescriptive, absolutist, linear description of exactly how the organisation will work in 2020, then that’s not for us. It’s critical we don’t imply we can have a level of certainty now, which we’ll only build up through the many years of testing and learning which lie ahead.
I think it’s helpful to call it an “Evolving” Operating Model, rather than a Target Operating Model.
We might one day refer to it as the “Minimum Viable Business Design”.
The Business Design will drive the desired outcomes at DWP by creating a continuous feedback loop and making it real throughout the organisation, not by planning.
It’s guiding our decision-making already
We’ve worked across the department to build up a community to work on the Business Design. We’re working on the design itself in an iterative way.
We’re still on the journey, but there are areas of our business design that we know will be critical to its success — and we’re finding ways to draw these together into a single consistent story.
One of the main purposes of our work is to address the need for a shared vision of the department. Now that we’ve achieved consensus around that, we’re starting to be able to make statements about the future which help guide our decision-making today. Recognising that DWP, like any organisation, is a complex system of people, we’ve found one of the most useful tools to share more widely has been a rich picture. It’s been well received so far, and the team’s writing a blog to share more about what we learned.
A valuable asset for the whole organisation
With good “design thinking” we believe this work can be an asset which helps us make the right decisions and get there faster.
We are already achieving a shared understanding across our change programmes and beyond. DWP is delivering the largest reform programme for a generation, and we’re starting to be able to join-up across that work and start to eliminate the multiple overlapping systems which would have resulted otherwise.
We don’t want to fossilise our costly paper-based processes, so that they live on forever, by simply building digital front ends around our old ways of working. Our vision has to be more ambitious. We’re working towards a cost-effective and achievable way to run the business, which will help us meet taxpayers’ expectations and give clarity on how we will operate in a modern and efficient way. This will be critical for us as we move towards the next Spending Review, which will require robust information to explain our operating costs and the investment required.
And our design is not just for the future of DWP, but for the future of government, as it will be critical to us defining and deivering our role in new era of Government as a Platform.
Join the Business Design conversation
From our work over the last few months, we already have a clearer picture of the future of the department and it’s starting to make a real difference — we’re starting to do the right work in the right places.
We’ll continue to share about how we’re going about it, the ideas we’re working on, and the results of our experiments, through our GOV.UK blog.
