In my work, I regularly find myself discussing transformation with customers who are thinking of moving some or all of their IT services to “the cloud”. Previously, I’ve talked about a project where a phased approach was taken because of a hard deadline that was driving the whole programme:
- Lift and shift to infrastructure-as-a-service (IaaS) and software-as-a-service (SaaS).
- Look for service enhancements (transform) – for example re-architect using platform-as-a-service (PaaS).
- Iterate/align with sector-wide strategy for the vertical market.
The trouble with this approach is that, once phase 1 is over, the impetus to execute on later phases is less apparent. Organisations change, people move on, priorities shift. And that’s one reason why I now firmly believe that transformation has to happen throughout the project, in parallel with any migration to the cloud – not at the end.
My colleague Colin Hughes (@colinp_hughes) represented this in diagrammatical form in a recent presentation (unfortunately I can’t reproduce it on my personal blog) but it was interesting to listen to episode 6 of Matt Ballantine and Chris Weston’s WB-40 podcast when they were discussing a very similar topic.
In the podcast, Matt and Chris reinforced my view that just moving to the cloud is unlikely to save costs (independently of course – they’re probably not at all bothered about whether I agree or not!). Even if on the surface it appears that there are some savings, the costs may just have been moved elsewhere. Of course, there may be other advantages – like a better service, improved resilience, or other benefits (like reduced technical debt) – but just moving to IaaS is unlikely to be significantly less expensive.
Sure, we can move commodity services (email, etc.) to services like Office 365 but there’s limited advantage to be gained from just moving file servers, web servers, application servers, database servers, etc. from one datacentre to another (virtual) datacentre!
Instead, take the time to think about what applications need; how they could work differently; what would be the impact of using platform services; making use of a microservices-based approach*; could you even go further and re-architect to use so-called “serverless” computing* (e.g. Azure Functions or AWS Lambda)
But perhaps the most important point: digital transformation is not just about the IT – we need to re-design the business processes too if we’re really going to make a difference!
* I plan to explore these concepts in more detail in future blog posts.