Startup Whitehall: Is it doable, and how?

The Minister for Intergovernmental Relations, Pat McFadden, gave a speech recently arguing that government needs to act more like a startup, to foster a culture of innovation. I think this could indeed help to shake Britain out of its current torpor by radically and rapidly improving state capacity. From my experience in government, it is sorely needed, but it won’t be easy. Managerial process-ism is chronic in the civil service.

I’ve worked in data science, digital transformation and innovation projects for clients that range from government to large corporates, and startups to small business. Below are five quick thoughts on how a culture of innovation could actually be embedded across the civil service.

1. Recruitment and promotion in any senior civil service role needs to be tied to innovation: 

Most startups fail, that is simply a fact. Thinking and acting like a startup will only work if ministers are willing to accept risk and failure as a normal part of doing business. This attitude then needs to filter down into the senior civil service before it can permeate the rest of the organisation.

Innovation needs to be a core component of SCS interviews. When did you try something that failed? What did you learn? How did you improve next time? How did you de-risk an innovation project? Managing a process shouldn’t be enough – senior leaders should have to show iteration and improvement (if not reimagining) of that process.

2. Agility is not optional

Agile methods are one of the best ideas around innovation methodology yet devised. Sadly, we live in a world where cynicism has developed around Agile innovation practices, largely because of gatekeeping and bureaucratisation by those wishing to monetise the idea. For me, Agile has only two essential components:

i) Iterative delivery. Iterative delivery reduces risk. It means prioritising ruthlessly and getting rid of all non-essential components to delivery something, fast. Once you have a basic working version, you can build on it. This is true at every level – for components, features and whole products. In his speech, Pat McFadden indicated this idea (which he called a test-and-learn culture) was at the core of his approach to innovation, which is encouraging.

ii) Staying close to customers (users). Get feedback, constantly. Like all of us, users are unsure, they change their mind, have a better idea, or forget things. Regular user touchpoints are essential to effective innovation. Ensure you have a user representative in your standups/daily meetings. Ensure that you have regular reviews where your product is presented to a range of users (‘sprints’ of 1-4 weeks are useful here).

3. The civil service needs to learn to accept risk without bureaucracy.

A large part of ‘being more innovative’ is just ‘innovating faster’. If we have the same number of good ideas, but I test twice as many ideas as you, I will be twice as innovative. over time. The problem is that, while the civil service does on occasion accept risk, it replaces that risk with bureaucracy. Every conceivable risk is studied, from every conceivable angle, memos are written, long reports are produced, small expenditures are escalated for approval far above any rational level.

If you want to do innovation, financial control can’t be your main concern. It just can’t. You need to accept you are spending a chunk of money and you may not get any return. The way you mitigate risk is to make sure that this money itself is limited. You aren’t trying to ‘boil the ocean’ with a massive multi-billion pound ‘change’ programme. You are spending a few hundred thousand pounds to test an idea, prototype it and get feedback. If you can do this for 5 different promising ideas, perhaps one or two will prove both viability and value.

As the saying goes, ‘move fast and break stuff’. If the Permanent Secretary has to sign off on every idea of a few million pounds, it won’t work.

4. Standardisation is essential

Innovation is all about reducing ‘time to value’. Cross-departmental working is not exactly a new idea. The current government’s mission boards may prove an effective strategic solution to the problem. But on the tactical (i.e. programmes and projects) level, many innovative solutions to intractable problems will require intense cross-departmental working. I am a big fan of ad hoc cross-departmental project teams. However, these are made much more difficult by technical and administrative differences, which in turn cause problems sharing data, creating teams and working effectively together.

While this is a problem, it could be turned into a major advantage. Startups don’t have the problem of effective co-ordination, because they don’t have scale. However, if everything is done differently across government, this scale advantage is lost. Economies of scale become diseconomies of scale. Two of the most essential areas to standardise are tooling and professional development standards. Technical standards for data access and storage – e.g. around APIs, add huge value. A standard suite of tools and a common data platform will be even more impactful. A secure, cross-government platform of pre-approved tools for various kinds of analytics, ETL, etc, would encourage standardisation and code reuse, and make inter-departmental collaboration far easier. Similarly, a common process of professional development, involving common certifications and learning pathways, could both ease the flow of people around government and smooth the institutional blockers to collaboration.

5. A 10x mindset.

This is perhaps the most vague but also the most essential idea here. The point is that startups and not trying to be 10% better than the incumbent. They are trying to offer a solution that is 1,000% better. Thinking about how to change an established pattern of working so as to get results that are 10x better leads to a fundamentally different mentality. This kind of mindset tends toward radicalism, and that often doesn’t go down well in government. If we really want a culture of innovation, longstanding ideas are going to be challenged, and unpalatable solutions proposed. Civil service empire-building will have to come to an end. Performance and impact will need to be the criteria by which success is measured.

Conclusion

These five points are by no means comprehensive, but I think whether or not the government achieve them will do much to determine the success of this latest push into innovation. If nothing else, it’s encouraging that both Pat McFadden and Morgan McSweeney have indicated that innovation, data and technology is such a critical part of their thinking. In the past, Dominic Cummings indicated similar inclinations, but his methodology was so chaotic and his tenure so brief, that he seems to have had little impact on the broad swath of the civil service.

Hasta la victoria siempre!