In software development a project that is driven on, even though those working on it fear it will never work, is known as a “death march” and right now that is what the UK’s “Universal Credit” project is looking like.
Claimed to be the world’s biggest “agile” software development project it has now been admitted that more conservative “waterfall” methods are being used to manage the back end of the system. That is actually a good thing in my book – because the backend manages the payment of essential benefits to millions of the poorest and most vulnerable people. Agile is not what the way to do this – you cannot tell the customer “sorry you didn’t get paid this month, we’ll tweek the system and fix the problem”.
The front end, the bit that manages people’s claims, is still being delivered in an “agile” way and that might make sense: this is about interaction between users and the system, but even here the risks are enormous and where is the testing? Testing is supposed to be what agile is all about yet all we are promised is a few small short pilots, beginning tomorrow – in other words small-scale testing with a live system. You wouldn’t do that with air traffic control or nuclear power safety so why do it with Universal Credit: and UC ought to be seen as every bit as mission critical for the UK government as those two systems.
The project is on it’s third leader in a year and more and more bits of it are being thrown overboard in a desperate effort to meet the deadlines set for political reasons.
- Universal credit has great potential for positive social change (guardian.co.uk)
- Universal credit pilot project to start tomorrow (itv.com)
- Universal credit pilot to launch with only a few dozen claimants (guardian.co.uk)
- Universal Credit pilots scaled back (bbc.co.uk)
- Universal credit: the essential guide (guardian.co.uk)
- Universal Credit ‘fraud risk’ fear (bbc.co.uk)