I went into how it was achieved here, but the short story is less, better quality information, and more, good quality digital services supported on devices right down to 320px wide using RESS.
There's still much more to do though. If you stand still online, then you're less likely to continue to meet the evolving needs of customer expectations and technology.
Whilst the project to launch our new platform might have finished, we're conducting continuous improvement, prioritised by actual numbers for services used on a Transactions Dashboard.
If I had one thing to pass on from this it's about minimum viable products, or perhaps a better phrase is a minimum usable service. What I mean by this is get your digital service to a usable state, test it in alpha, launch it in beta, gather data and improve it until you're happy to launch it live, then continue to improve.
Perhaps you want your service to do ten things. Launch it with two and build on it.
It becomes much easier to make a case with actual usage data and customer feedback rather than just a projected channel shift. The minimum usable service route allows you to do this which can be useful in an environment where many groups are competing for the same funds.
Just one word of caution; be sure that your service is good enough to move to the next stage by assessing user testing, before you do. This should be standard practice though. not just for this but for anything that involves any sort of user acceptance testing.
If you've got a view on this, follow a different path or have tried similar I'd love to hear from you.