Chapter 7

Winning the arguments

It is the Balkanisation of authority across government that is responsible for dysfunctional decision-making.

— Nick Clegg, Deputy Prime Minister (2010–15)

The focus of a digital team shifts over time. Sometimes the primary objective should be creating the conditions for the team to do the right thing. At other times, it will be using those conditions to actually deliver improved experiences for users. Once you’ve created space for a team to succeed, they can get on with shipping small, fast and bold products and services. The more ambitious the aims of a digital organisation become, the more conditions they need to put in place to have a chance of releasing greater benefits.

These two tasks – building things and creating the space to build things – can run happily in parallel once a team is up and running. The whole digital team should not be pivoting from one to another; your product teams will focus on the delivery, the bureaucratic hackers on clearing a path to advance. What will change is the issue that is front and centre for the digital institutions’ leadership. Calibrating the balance between healthy delivery and clearing space for moving on to bigger things at any given moment is one of the trickiest judgement calls. Stay too attached to delivery, and you risk curtailing your ability to build a truly digital organisation. Fly too far off into the distance, and you may lose track of actually delivering improvements to both users and the bottom line. As they are the benefits of creating a better organisation in the first place, you forget them at your peril.

The temptation that new digital teams encounter is for them to try and line up a perfect scenario for a successful digital organisation to take flight in one go. You might come away from reading this book thinking, ‘What’s the point in beginning anything without having first sorted out the best staff, the perfect political conditions, the right organisational culture, and had all the most obvious blockers swept aside?’ If all that were ticked off, you could surely then accelerate as fast as you can into the space you’ve created, delivering more impressive and impactful things for users, and saving money more quickly. This sounds very sensible. The only problem is it will never happen. The chances of creating a perfect environment for changing an old, legacy-driven organisation – and holding that scenario steady in an ever-changing, event-driven world – are zero. One of the biggest mistakes we have seen new digital institutions make is waiting until they can see the very bottom of the pool before diving in from the highest board. Taking a shallow dive into murkier waters is the wiser way to go.

Digital teams should feel comfortable (or at least, get used to feeling uncomfortable) with working things out as they go along. Win arguments as they arise. There is no perfect end point for a digital organisation to aspire to. In the UK government, the digital team tried to win the most important arguments that were in front of it at the time. At many times, the strategic question confronting us was very simple. Mat Wall, a GDS technical architect, summed it up: ‘What can we fix to help our teams ship better products this Friday than last week?’ Having your strategic priorities led by what is blocking delivery and meeting user needs right now (rather than some unspecified point in the future) is a good way to maintain focus. This week’s delivery niggles are a valuable source of suggestions for where to invest the efforts of the bureaucratic hackers that can fix them.

The first set of conditions that need putting in place for a new digital institution to work are largely about making sure that a new kind of organisation, capable of agile, user-­centred delivery, can be transplanted into a large bureaucracy without tissue rejection. We explained these in chapter 1: these are the four things that give a digital team the licence to simply make a start.

After putting those in place, most digital teams will then go through a period of successfully delivering a certain kind of output; small, low-risk and greenfield projects of the type we spoke about in chapter 5. These are projects that can exist and thrive independently from the legacies – technological and cultural – that are attached to the wider organisation. However, as the digital team turns to look at redesigning services that are deeply embedded within the various tendrils of the existing organisations – the brownfield sites – it quickly becomes obvious that the conditions that allowed a digital team to exist are not enough to transform the whole business. Creating the environment for this kind of change to start happening required the digital team to acquire a new mandate, and win a new set of arguments.

From the centre and here to help

You will not be able to effect change within a government machine or large corporate body if you cannot operate levers of influence from a central position. This leverage over the various parts that make up the whole organisation, be they government departments or subsidiary businesses, is critical. The levers must allow the digital team to consistently modify behaviour and overcome inertia. If your digital institution has reached the stage of delivering new products and services relatively easily but has no leverage over the existing legacy, it is probably time to prioritise creating a stronger cross-­organisational mandate.

Mandate is one of those slippery, dangerous terms. It means different things to different people. From a department’s perspective, one reading of the paragraph above is that a central body – such as a finance ministry or strategy team – should be handed the power to dictate terms to the rest of the organisation about anything they feel like. That is not what we’re saying.

Mandates vary in two ways. They can operate through a different mix of powers; some combination of soft (via influence, personal relationships, exchanges of favours and shared best practice) and hard levers (laws or decrees, rules, spending controls). They can also vary according to the range of issues that the mandate covers; recruitment, money, technology choices, legal frameworks, and so on. The mandate your digital institution needs depends on the organisation you are working in.

In the UK, we took the decision to create a relatively broad mandate with a set of hard levers for the GDS. The digital team needed to be able to administer stiff medicine in order to overcome entrenched interests within and outside government. These interests prevented parts of the civil service from delivering services that either met user needs or represented a reasonable investment of taxpayers’ money.

That decision was made partly based on the experience of others. Previous attempts at reform with a powerless or non-existent centre struggled to make any sustained progress. This is a reform problem, not a digital problem. Other attempts to change the rhythm and shape of institutions have faced similar problems; it took 26 years, two major ministerial reviews and an intervention from Winston Churchill to give statistics its rightful place in government.45

Not everyone agreed with our judgement. Some complained that the GDS’s central mandate unreasonably diminished their department’s organisational power. However, whatever the exact nature of a digital’s team influence over the collective organisation, having some form of central mandate appears to be essential in driving savings and improvements to the digital experience for users. We have yet to find a compelling example of an organisation that has successfully transformed itself from a legacy-driven business without some form of central push.

Other organisations and governments – particularly those that are younger and carry less historical and cultural baggage than Britain’s bureaucracy, 150 years old in its current form – will not require so adversarial a central mandate to drive change. We have seen large, legacy-ridden corporate conglomerates achieve a huge amount of digital change largely through having trusted digital leaders working in the centre who win the trust of their fellow executives. The mandate the central team has in this case is largely implicit in their soft influence; departments follow a central team if and when they recognise it is the expert. It also helps if the chief executive publicly makes it very clear that she believes the centre is the expert too. Giving a digital team a central mandate does not mean it has to be combative.

It is easy to ascribe the inertia of bureaucracies to natural forces, uncontrollable by individuals. This is not true. It takes surprisingly few individuals to completely gum up a huge government machine. Not many people are obstructive for the sheer hell of it. In most cases they will have very good personal or professional reasons to maintain the status quo. When your digital team turns up to upend their position, they will hide, delay or fight. Faced with this, influence, charm and friendly wheedling may only get you so far. Teams trying to start digital institutions (especially in governments) therefore shouldn’t underestimate the value of acquiring hard powers. You don’t get many chances to ask for them, and they are much more easily removed than conferred.

To get that hold of an expanded mandate, you need to make the most of your opportunities. After two years, the GDS was invited to present what had been achieved in digital government to a Cabinet meeting – an extremely rare privilege. Demonstrating new digital services meant getting a TV into the Cabinet room, and persuading Number 10 that this didn’t represent a threat to national security. Having managed that, we took the chance to make some very clear asks of the country’s most senior politicians. By presenting the proposed mandate alongside the progress already made, no one was about to get in the way.

Deciding the right balance of hard and soft power is a choice that you can shape according to the organisation around you. The scope of your mandate – the areas of the business that the digital institution gains the right to shape and judge – should be determined by what is blocking delivery. Again, this will vary from place to place. However, there are some issues where a cross-organisation mandate is always useful in managing the hard conversations to come.

Wrangling IT

Digital and IT often have a troubled relationship. IT, in the form it tends to take in large organisations, is technology that has been conceived with a mindset that predates or ignores the open internet. This sets it up in direct opposition to digital.

IT, like laws and regulations, is used by organisations as a sweeping excuse for why the user experience of online services is so poor, and why an organisation can’t behave as the digital institution it apparently wants to be. Those who are most often guilty of this dissembling tend to be the people who have the weakest understanding of IT. In some institutions, pinning the woes on the IT department is an article of faith. Relying on ancient back-end systems does naturally make things difficult, but nobody is forcing the organisation to use them. Treating IT as a fixed constraint, as opposed to something that the organisation can make an informed strategic choice about ignoring or prioritising, allows too many senior managers to blame the IT when the real problem is much closer to home.

Regardless of the personalities involved, there are three common reasons for the relationship between IT and digital being rocky: misunderstanding, mythology and contracts.

Misunderstanding may creep in as a consequence of your organisation believing digital is just another way of doing IT. ‘You make websites don’t you? You must be another IT team.’ The existing IT team doesn’t want another group muscling in on their patch. A depressingly high number of managers, public and private, are obsessed with ‘turf’ and their sphere of influence; IT managers are no different. They are generally not inclined to welcome a bunch of upstarts who have turned up uninvited to criticise what they’re doing.

Many IT teams in big organisations have got used to nobody really understanding what they do, especially at a senior level. They are unloved – which is unfair – but they are also essentially unaccountable. This can lead to some unhealthy complacency. By having the technical skills to be able to ask the right questions of IT colleagues, digital teams pose a threat to a quiet life. This is not a good place from which to begin a healthy working relationship.

IT security offers another seam of mythology, providing a rich seam of questionable reasons for why things can’t possibly be changed. At their worst, security myths actually lead to organisations taking on bigger risks – forced to rely on unusable old technology at work for fear of being hacked, staff eventually turn to unsecured personal devices to get things done.

Retrospective: secure statistics

The UK Office of National Statistics (ONS) was long an organisation stuck under a digital cloud. Despite having some very able people working in it, the ONS was running an online presence described as ‘the worst website in the world’ by the Financial Times. For most organisations, this would be an embarrassment. For the ONS, a public body with publications that moved markets, it was a deeper worry.

Security mythology contributed to ONS’s woes. ONS was obliged to publish certain key economic statistics, such as inflation numbers, at a certain precise time. Many numbers needed to be published at 9 a.m. sharp; not a minute later, as the market demanded them on the dot, and not a minute earlier, as they were embargoed until that moment. Built on creaking old IT, the website’s frailties meant that statistics uploaded at 9 a.m. – as the rules decreed – were often not viewable until several minutes afterwards. These delays were earning the ONS some very powerful critics. Fear of security breaches, ill-informed box ticking and lack of basic technology capability had prevented any solutions coming forward.

To unblock the impasse, ONS brought in a national security expert. He gently but firmly explained that provided the numbers were securely encrypted, they could be uploaded to the site hours before. On the stroke of 9, the encryption could be removed – the job of a second. It was a simple solution, but one that could only be reached by framing technology as an enabler of meeting user needs, not as something fixed and unchanging.

As a consequence of outsourcing, many IT teams in big organisations have been effectively captured by suppliers. Denuded of their own technical capabilities, they have been reduced to the role of contract managers – buying things in the hope it will fix the problems caused by the last order of stuff they bought. Without the skills needed to properly interrogate suppliers’ offerings, organisations buy the wrong things on lengthy contracts, leaving minimal room for them to respond when circumstances change. All this is anathema to designing and running decent digital services or meet user needs.

It is often difficult to redesign or transform a digital service without tying it back in some way to the legacy IT. To have a chance of success, a digital mandate must make it possible to stop poorly conceived, hugely expensive and long IT contracts from being let. It must also ensure people with genuine technical knowledge are given the opportunity to interrogate new investments in IT; and ask the basic questions that may not have been raised in decades. Good IT managers will embrace the chance to bring new skills into their teams and have the business pay them some proper attention for once. Bad ones will resent the oversight and stick doggedly to ploughing their own furrows.

All of this requires an organisation to stop looking at IT as the dull, frustrating systems that constrain the decisions they can make, and instead consider what technology is needed to support the ever-changing needs of their customers, employees or citizens.

One of the most important technology concepts to embed within your organisation is the idea of technology as a commodity. In simple terms, this is the idea that many building blocks of digital services are relatively inexpensive, widely available and more or less interchangeable with other components of the same type. This simple idea has three major implications. First, there is no real need to build or minutely specify all these components from scratch; things that were once built to custom specifications no longer need to be. Second, there is even less need to get locked-in to long-term contractual arrangements with one supplier; you should be able to swap between trusted services relatively easily. The hosting provider for GOV.UK was changed on several occasions to get a cheaper service without any visitors to the site noticing. Third, with technology getting cheaper and smarter as Moore’s Law continues to hold, increasingly complex components will become commoditised as time passes.

In the UK, GOV.UK came in for criticism from people who thought the team wrote too much bespoke code. It was argued that the GDS could have deployed more commodity technology rather than trying to build our own answers to well-understood problems. Some of that criticism was fair. Part of the reason for the GDS writing much of GOV.UK was that doing so offered a learning opportunity to many developers in a way that simply implementing commodity technology would not. Should the organisation have effectively chosen to spend more money on GOV.UK to improve the technical ability of the civil servants who built it? There is an argument on both sides. Either way, there are always trade-offs to be made between flexibility and capability, and between commodity and customisation.

For governments, adopting open standards can play a huge role in driving up the take-up of commodity technology. Heroic work by officials and experts in the UK government helped break our bureaucracy’s default dependency to expensive, proprietary technology choices. This saved money and helped thaw out a frozen government technology market. Having spent decades making contracting choices largely based on the supplier’s name and ability to deliver things on a huge, lumbering scale, open standards made the government’s technology buying decisions more about how to ensure flexibility and provide scale when it is needed, rather than going big from the very beginning. A change in philosophy like this doesn’t create success overnight, but it does get an organisation to think much harder about technology in the context of its overall strategy – and it certainly wakes up the suppliers.

If the senior leadership of an organisation has little awareness of what is going on in the world of technology, they will be left to guess what to buy, build or hire, how much they should pay it, and what the strategic consequences of those choices might be. Too few executives ask to see the full wiring diagram for how their organisation’s technology is set up. Being confronted with this horror is often reason enough for even the most incurious technophobe to wonder whether something might be amiss. Wardley mapping offers an excellent toolbox for helping those managers develop greater situational awareness,46 but getting to that point requires an certain openness at an executive level to listening. It also means hiring technologists who know when to code, and when not to code, in positions that have a voice in making strategic decisions.

Wrangling people

Most large organisations have set themselves up to bring in the same sort of people on an industrial scale. This partly happens out of necessity; as people change jobs or leave, ready-made replacements have to be ready to step into the breach. That is fair enough, but as an unintended consequence, it also logically dictates the creation of standard, template recruitment processes and rules. There’s a reason why big companies call the departments who run these systems ‘human resources’ – the philosophy sitting behind it is that the people are largely exchangeable, and easily replaced by others with similar skills and backgrounds. This undermines an organisation’s diversity of appearance and diversity of mind.

Trying to transform a large organisation is therefore pretty much impossible without disrupting the norms applied to hiring people. When the GDS was starting out, the way government officials were hired in Britain was very similar in nearly every case. Applicants filled out a long form, writing lengthy answers to provide evidence and experience against certain competencies. If they passed this stage, they would be invited for an interview, where they will again be asked to articulate evidence for meeting a particular competency. This spiel didn’t need to be any different from what they said on the form – candidates could sit and read it out loud, if they so wished. If their answers proved the most convincing, they were in. This process is clearly biased towards certain kinds of people. It works especially strongly against people who aren’t strong writers or plausible when sitting in front of interrogative committees.

Digital teams need to bring in skills that can’t be tested through this kind of process, or indeed, any single process. You can’t draft your way to proving yourself a great designer or coder. Nor can a non-technical interviewer make a reasoned assessment about whether a technical architect is well qualified or not. Being able to change the typical way an organisation assesses applicants’ employability and adapting the recruitment method for different types of skills are essential parts of a digital mandate.

The hacking of HR shouldn’t stop after getting people through the front door. To build a digital institution capable of transforming the wider organisation you will have to break the other shackles that keep an organisation from hiring in its own image. That means looking at pay and introducing more options within the standard balance of rewards (not everyone will value a good pension or longer holidays over having more cash, for example). It means reviewing performance management systems that have few options for recognising excellence in career paths not expected to scale the senior leadership ladder. It means breaking grade structures that are designed to signpost the upward trajectory for a certain mixture of skills, and create perverse and dead-end promotions for specialists. The old trick of promoting the specialist into a management role, used as a last resort when it is the only way to pay specialists enough to stay, often lands those experts with responsibilities they don’t want and are ill-qualified to handle.

The digital team’s mandate must include permission to bend and test HR rules. This doesn’t mean building a bonfire of procedures and process, no matter how tempting that might seem. Many of the existing principles that govern how an organisation handles its people will be perfectly sensible, but interpreted into uselessness by HR staff who lack any empowerment to take sensible and proportionate risks.

The quid pro quo for giving the digital team freedom to find better ways of hiring and looking after its people is that it must share the benefits of a better way with everyone else. There shouldn’t be special dispensations for digital. If the digital team comes up with better job descriptions, innovative interview methods, or a more flexible pay framework, everyone else should share in the benefit.

Wrangling money

How an organisation chooses to invest money is a good test of its health and personality. In bureaucracies, especially those of a certain age, the business case process designed to appraise investments generally manages to pull off the worst of both worlds – it is both slow and arbitrary. Slow, because following the rules laid down by finance and procurement teams to the letter is a task that necessitates multiple people, several months and many thousands of largely redundant words. Some people view the time it takes bureaucracies to ruminate over investments as a source of strength. It combats hastiness, and allows the supporting logic for doing something to mature like a good wine. Unfortunately, in the majority of cases, the maturing process more closely resembles the effect time has on a good peach.

Lengthy business case processes add delay without adding more certainty. As a consequence, they fail to protect organisations from making poor choices. This is where the arbitrariness creeps back in. When confronted with a default route to getting things done that lasts longer than their likely term in post, legitimately impatient ministers or chief executives will simply resort to shouting loudly or conducting back-room favour exchanges to get what they want, without any recourse to analysis at all.

The traditional business case process typically applied by governments and large organisations is a neat example of a one-size-fits-all process that only properly serves a certain kind of project. This does not mean that it is useless. Through standardising certain processes, many governments have got much better at delivering certain kinds of project on time and within budget. Let’s say a department is trying to build a large piece of infrastructure in a relatively controlled environment – a major new highway tunnel, for example. This is a well-­understood problem, tackled in various forms before. The materials, behaviours and challenges at play are largely well-known. No major innovations or social changes are expected to dramatically alter the need for a tunnel. In this case, doing lots of upfront thinking in preparation for releasing one substantial chunk of money to get the work done is sensible.

Some tasks that governments and large organisations take on do occur within these relatively controlled environments. A great many do not. It is in these cases that the templates fall apart. The logic of being able to predict how future investments will turn out does not apply to projects and programmes with a large technology component. The market for new technology moves too fast for the business case process, as does the digital society in which the new policy or service is supposed to flourish. User expectations of what is possible, or even what is expected as a basic level of functionality, are accelerating all the time. Spend two years economically justifying all the requirements you demand of your new employee communications system, for example, and you’ll find the world has changed in the meantime. This is why organisations end up being forced to defend the idea that in 2017 pagers are a perfectly good way of meeting your employees’ communication needs.47

Creating a cumbersome process for releasing even small amounts of money is not a good use of time or brainpower, and reflects a very bureaucratic belief that terror of risk can be made to go away provided one simply writes everything down. If it takes you a year to write a business case, you want the investment it supports to last a lot longer than that – five or ten years, at least. Again, this is not a motivation well suited to the rapidly evolving world of digital technology. Nobody has a 10-year mobile phone contract. There’s a reason for that.

Fixing the choice architecture for how money is spent in large organisations is no small task, especially those under the heavy scrutiny experienced by governments. The battle to win as a digital team is two-fold. You need a process that allows digital delivery teams to spend small amounts of money, quickly, in exchange for those teams demonstrating that the cash they’ve been given has allowed them to reduce the risk of scaling up their service for more people to use. Spending £50,000 to find out in four weeks if any of your 20 trial users will use a mobile app to log their weekly sugar intake is a lot safer than spending £10 million over two years on building an app, launching it to the world, and crossing your fingers. Traditional business cases push teams to gamble on the latter; if you have to write 200 pages of nonsense either way, why not go big?

Another, more insidious problem some very big organisations suffer from is the need to choose options that are reassuringly expensive. Large organisations have grown so used to receiving huge bills for their IT systems they cannot take the dramatically lower costs of commoditised technology seriously.

Money wrangles kill off many digital teams, because most processes for getting the money needed to get started on redesigning services either take too far long or are unable to release a small enough amount of money. Worse, they insist on teams being able to deliver many pages of fiction about how certain they are about the assumptions they make for their project’s success. The truth is that many finance ministries or heads of finance would prefer to see a complete lie about the lifetime cost of a project than a relatively certain estimate of how much the next three months will cost – that is what their spreadsheet demands. This is cognitive dissonance operating on a grand scale.

Unpicking all of this will take a long time. In the UK, it took more than a year to put in place a business case process more suited to agile projects than the Treasury’s waterfall-friendly Green Book guidance. As a digital team, your focus – beyond challenging and adapting default processes to stop them from breaking agile projects before they begin – is to help make sure that the people making investment decisions in your finance ministry or elsewhere are properly qualified to opine about technology. At the centre of most finance departments in government around the world is a cadre of young, intelligent and gifted amateurs. They know little or nothing about the area of spending they oversee, but they know the spending process inside out. Getting some people with internet-era technology knowledge, rather than good, generalist guessers, can completely transform the ability of a large organisation to invest wisely in technology-led change. A better process won’t fix everything; you need different people too.

Summary


45 Reg Ward and Ted Doggett. 1991. Keeping Score: The First Fifty Years of the Central Statistical Office. Central Statistical Office.

46 A very helpful place to start is: http://www.wardleymaps.com/

47 Pagers are still relatively common devices in parts of the NHS (the organisation uses 10% of the world’s remaining supply), with 130,000 nestling in pockets alongside nurses’ personal smartphones. https://www.theguardian.com/society/2017/sep/09/old-technology-nhs-uses-10-of-worlds-pagers-at-annual-cost-of-66m