The Journal.

A technology blog run by Rowden team members

Taking matters into our own hands

We know the UK MOD struggles to support rapid technology and system development, so how do we respond?

I got the feeling some time ago that people wanted change in public procurement:

‘Why would we pay [insert massive company] to build our idea and let them keep the Intellectual Property?’

‘I’ve just been quoted these from the OEM for half the price, why can’t I buy direct from them?’

The answer was always similar. ‘It’s easier’, or, ‘that’s the way it’s always been done’. When response 1 was likely to garner a heated debate, the answerer would deploy response 3. This was a mixture of vague references to EU law – guaranteed to bamboozle even the most determined end user.

Why Now?

Discontentment will have existed for decades with the Governments [in]ability to rapidly test concepts and procure innovate solutions. So why is now any different? Put simply, the supplier base has changed, and knowledge is more accessible for the inquisitive.

The former is both traditional and non-traditional defence suppliers, with a higher risk appetite, who are utilising contemporary engineering and recruitment approaches.

The latter is the internet. It’s hard for the detractor to deflect questions which relate to more efficient but more difficult procurement approaches when evidence of their effectiveness is more accessible.

Frustratingly though, nothing substantial seems to have changed in the big delivery agencies. And why should it? It’s a contentious subject but without amendments to the incentivisation model, we will continue to be left with stats for ‘successful deliveries’ of capabilities that the user never used. Because the level of change needed feels monumental, think 40 years of ingrained thinking, we have ended up with a middle ground. Lots of new UK Gov teams and departments with works/werx/[insert works spelt differently] on the end of their names. I joke but it is progress.

Whilst they have differing scopes of responsibility, the mission statements all indicate one thing, a desire for change, mostly manifesting itself as pace, better performance and cost savings.

Bureaucratically developing a solution is frustrating when a one week ‘hack’ with the right skills would have provided the data required for a go/no go. The issue is that these departments have budgetary constraints greater than their well trusted Equipment Programme brethren and must still abide by slow procurement processes – by the time we’ve won some work, someone else is building it (and lets face it, it is someone in the US).

Elsewhere, competition for prototyping and pilots is good if the customer can articulate their intent clearly and they have access to technically competent markers. Rarely do both co-exist. Tendering is referred to as ‘best liar’ competitions for a reason.

What is an SME with a desire to build and test concepts in a two-year horizon to do?

It sounds easy:

  1. Remove third party dependencies – namely, a reliance on our customers
  2. Triage and continually foster ideas
  3. Develop concepts rapidly and test viability (see how I avoided ‘fail fast’ there!)

With a commitment to 100% reinvestment of profit into growth, the ingredients for a skunkworks existed in Rowden:

  • A high-risk appetite
  • Industry leading technical skills & domain knowledge
  • Resource – people, facilities and environments
  • Reusable, lightweight processes

Part way through a 10,000sq ft. facility expansion for concept development – due for completion in July 2020

After a short meeting, the Rowden Foundry was formed – no fanfare, no marketing, just focus.

  • We dedicated 20% of everyone’s time to ‘The Foundry’ which sounds minimal, but that’s 25 days of dedicated effort a week based on our current headcount
  • A £3500/month hardware budget. Why? There is very little that can’t be tested with that money and it forces thinking outside the box

3 months in, what’s it looking like?

The Good:

  • In the space of one day we were flying autonomous systems via voice control using machine learning. We have now invested a full funding line into the development of this capability under six separate work packages
  • We are embarking on a collaboration which will utilise free space optics in congested and contested RF environments. We are aiming for three months from initial phone call to demonstration. It’s on track
  • It diversifies the day-to-day work of the engineering teams and builds new relationships within the company

The Bad:

  • The ultimate indicator of success is technology exploitation (I am personally obsessed with it) and we won’t find out if we have managed that for at least 12 months
  • Phase 2 of a concept always feels like £30 – £80k of our own dollar, which is significant in a cash flow funded company. Maybe it’s not all bad, it makes us focus on exploitation chances, something grant money does less well
  • A lack of customer involvement. That’s not their fault, we can’t afford minibuses full of Business Development folk to lay siege to their cafes and let them know

The Ugly:

  • Giving up a day a week when everything else feels more important is hard. Paying work dominates. We are yet to find a viable solution

It would be tempting to believe the model can be ‘cookie cutter’d’ but the devil is in the detail. Why is it weighted to software engineering over electronics? What is ‘just enough’ documentation to exploit the technology? There is still a shed load of learning to do and it will be different for everyone who implements it, from culture to available skills.

What we are learning from other areas of the company is that Government approaches to innovation can be tweaked and tested but the results are largely superficial. There are limited viable options for scaling the current approaches without delivery agency buy in and it does not suit a number of the incumbent providers. Put bluntly, it needs some mavericks with central Government support.

We hear a lot of groaning in the SME community and that’s fine, it’s going to be harder in this economic climate than before but don’t expect that negativity to change anything. What we can do in Rowden is focus on what we control, most notably our ideas, our risk appetite and our work ethic. The best way to deliver lifesaving technology is to remove dependencies and plough on – plus, failing fast feels easy when no one external knows what failed!

I’ll leave any future Foundry related posts to the engineering teams who can chat through successes and failures. We question everything and so should you. Watch this space….