by James Gardner

Department for Work and Pensions CTO reveals the power of crowd sourcing in government

Feb 21, 2010
GovernmentIT LeadershipIT Strategy

Can you imagine a technology leader without an appropriate IT strategy? Surely not: the increasingly complex web of dependencies between what technology can deliver and what the business wants don’t lend themselves to decision-making on the fly. It has now become so difficult to work out the best way forward that larger IT organisations have teams of dedicated strategists that are consulted on every question from the size and shape of organisations to specific technology acquisitions on matters of corporate culture. Everything which isn’t day-to-day operations is fair game if you’re an IT strategist. Increasingly, the work of strategy people is kept hidden: a result of the increasingly difficult and increasingly sensitive material they are being asked to work on. Hidden in the back room, the strategists chug away, thinking their deep thoughts until they come up with “the answer”. Eventually, after appropriate sanitisation by management, a document is rolled out and everyone else is expected to sign up to whatever transformation it describes. Not surprisingly, such an approach to IT strategy has a few problems. One problem is that technology strategists rarely have the deep expertise in all the sub-disciplines of IT and the business that they’d need if they were really going to create something actionable for the future. It is the perennial issue with doing strategy behind a closed door: you don’t know what you don’t know, so when you finally do release something, the real experts get up to tell you why “it can’t be done”. And quite often, those experts will be right. There’s another problem with closed-door strategy, and that’s that you never get the buy-in from those who have to turn the words into action. When people feel that IT strategy is done “to them”, they will often set up small, seemingly inconsequential roadblocks that either stop things getting done altogether, or at best slow things down to a significant degree. At the Department for Work and Pensions we’ve experienced all these challenges. So when we decided it was time to refresh our thinking about the way forwards, we decided to try an experiment. Instead of hiding our technology strategy team in the back room until we had a finished working product, we decided to get everyone that could possibly add value to the process in a room and crowdsource the detail we needed. We had experts in every discipline we could get our hands on, senior managers, and yes, even some junior people from close to the front line of service delivery.

We were inspired in this approach by attending various ‘unconferences’ – a way of planning meetings based on the principal that you don’t plan anything up front. In an unconference, you set up a broad topic for discussion, then you let the attendees organise themselves into breakout groups to drill into the detail. Anyone can propose sessions on anything, and attendees can move about between the sessions as their interest takes them. There are unconferences like this going on in every sector from banking to government, and attendees almost always say the format works much better than traditional PowerPoint shows where everyone is asleep by the end of the lunch-break. Our process was a bit more regimented than the approach followed in an ordinary unconference. At the start of each session, we posed the question “What do we have to design?” to the group of more than 100 individuals, and accepted proposals for sessions from the floor. We then called for a vote, and the top four topics were the subject of breakouts. In each breakout, we had people discuss the topic. We installed a chairperson in each room who had one function only: to call for a vote whenever anyone proposed a decision we might add to the strategy document. If the group majority came down in favour of the proposition, it was tentatively added. Then, at the end of each round, we had everyone return to the main plenary room to ratify each and every decision the group made, again by voting. We ran this methodology five times over a day and a half, and got almost 200 decisions from the attendees. The strategy team has since used all this material to construct what amounts to the most detailed technology strategy document I’ve ever seen. It includes significant details on how the organisation will approach everything from security to procurement, and all of it is actionable. It is also firmly based in reality, because we were able to use the brains of all our specialists in its construction. We didn’t know what would happen when we did this, and it had never been tried in a public sector organisation before. Could such a format actually work in the context of creating a working plan for the future? What if no-one contributed? What if everyone had such strong views that we couldn’t get to any decisions? Our experience, in the end, was good. There was plenty of debate in the breakouts, and though loud voices sometimes overwhelmed individual points, the fact that the whole group had to ratify everything at the end of each session seemed to impose a balance on the overall proceedings. We often had the situation, for example, when a particularly extreme point got passed in the breakouts, but was subsequently overturned by the group. The main thing everyone wants to know, though, is whether the work product we generated was any better than what a traditional approach might have done. The answer depends on who you ask. On the one hand, the document is substantially more detailed and broad-reaching than a traditional process might have achieved. It goes to places we would probably not have thought of going ourselves. It challenges us to face big technology problems in new ways. It is highly innovative, and not what we might have expected had the overarching perspective been that of what our technology leadership is comfortable with right now. It is also very well supported because everyone who could conceivably have added value has done so. Most of the experts we needed had their say. And, by contributing to the process, everyone has tacitly signed up to the strategy in advance. They helped build it, after all. But there is a key question everyone asks when they hear what we’ve been doing. How can you trust that so many decisions, made in such a short amount of time, are really the right ones? The answer to that, of course, is that you have to believe that 100 minds working together really are better than four or five working alone. The power of a crowd is that it enables you to process very complicated information efficiently. But even more importantly, each mind in the crowd has unique information that contributes to the whole. The data set from which the strategy is drawn is incredibly rich compared to what would be possible with a small strategy team. When you add all that together with some process structure you get a unique problem-solving machine, one that’s optimised for the specific strategy issues you’re attempting to resolve. There’s one final point that needs to be made. When you crowdsource IT strategy, you are unlikely to get the end product you expect. For us, there were some unusual twists that weren’t even considered before the event. The implication is you have to believe the people you put in the crowd have the best interests of your organisation at heart, and that they will collectively choose what is inherently right. Our experience suggests they will do so, given the opportunity.

About the author: James Gardner is the chief technology officer at the Department for Work and Pensions. He writes regularly on his blog at