Our recent blog ‘How to pace yourself in the race to the cloud’ talks about the need for strategic and purposeful cloud adoption. This is important, because cloud is not just an environment, it’s a mindset. To derive benefits from the move, you need to consider cultural factors as well as technical capabilities.
In our experience, implementation typically follows a four-phase process, from opportunistic adoption to full cloud-native status. And it’s the transition from cloud-first to all-in that causes problems for many enterprise organisations.
With early, opportunistic projects, engineers tend to work meticulously to embrace the full benefits of cloud, advancing architectural patterns and principles. But this becomes less feasible when you’re faced with the mammoth task of relocating the bulk of the IT estate. You need to alter your expectations and treat short-term decisions as steps towards longer-term goals.
We’ve worked with many large enterprises, bringing lucidity to complex challenges that threaten to derail migration efforts. Here are four key learnings that could help prevent your own migration going beyond the deadline or breaking the budget:
1. Minimise your migration bubble
In the context of a typical large-scale migration, the goal is to exit your datacentres in as short a time as possible. During the migration, dual running costs create a ‘migration bubble’, driving up short-term spend. So suddenly, speed is of the essence and you no longer have the luxury of time to fully optimise every application moving to the cloud.
Evaluate the situation objectively, decide where you can make compromises, what you can come back to later and what, if anything, truly justifies a full rewrite during the move. Provided you continue optimising applications once they’re on the cloud platform, the maths stacks up in favour of this approach. It’s important to be ambitious, but you also need to be realistic. You want to stretch your capabilities, not snap them.
2. Plan it, but don’t overthink it
Naturally, you need to plan how you’re going to get to the cloud. But avoid overthinking it, and stay clear of the analysis paralysis trap.
You’ll need to fully catalogue your application portfolio, making sure you understand technology, dependencies, criticality, ownership and other factors that may be relevant to your business. We’d recommend devising a straightforward set of criteria that allows you to rank applications in terms of value and complexity. Apply this consistently across the portfolio to help establish a priority order.
Consider which of the three paths for mass migration each application will take. And think about the target platform for applications that will follow the ‘Evolve’ (replatforming) route.
It’s important to resist the temptation of going into the minutiae of the move at this stage. Even a high-level design is probably too much detail.
The planning process should take weeks of concentrated effort, but not months. Otherwise the plan will be out-of-date before you start, and could be all but obsolete by the time you reach the applications at the end of the list.
In short, do enough planning to get started, have a route in mind, but expect the plan to change as you progress.
3. Start small, then build on skills
Many organisations struggle to work out where to begin, but the advice here is relatively straightforward. Start with something significant enough to deliver benefits, but simple enough to get done quickly.
Applications in a typical portfolio have a correlation between the complexity and value of moving, i.e. the bigger the benefit, the harder it is to realise. Look for your outliers; applications that are simple to move but offer relatively high benefit. Prioritising these means you’ll enjoy some quick wins, giving the team a boost and delivering tangible gains.
Many businesses are tempted to tackle the most complex and/or valuable opportunities first. But this should be avoided, especially during an all-in migration. Start simple and build on the team’s growing confidence and capabilities by moving on to similar application types that will reinforce recent learnings. As you progress through the plan, some of the risks associated with your more complex moves will be reduced by learnings and assets developed along the way. Taking small steps which accumulate and gather momentum is the most effective way to achieve migration at scale.
4. Make sacrifices
Some applications simply won’t make the cut for migration.
It’s hard to see the wisdom of migrating a legacy CRM package when you could take the opportunity to move to Salesforce, Dynamics or a similar SaaS offering. Some older mainframes simply can’t be migrated, and a rebuild may not be viable. As much as 10 to 20% of the estate can often be retired, with some moving to SaaS and others potentially staying on-premise for the time being.
Clearly, if your all-in migration is going to have some stragglers, this needs to be considered at the outset. Provision will have to be made for a much-reduced datacentre footprint to accommodate them.
Maximise the migration opportunity
Are you looking for guidance on your cloud migration journey, or making a business case for the best way forward to present to colleagues? Our whitepaper ‘Focus on the ‘how’ for large-scale cloud migration success’ could help. Download a free copy here.
We’ve written a series of blogs on enterprise cloud adoption and migration. The following might give you some inspiration:
Or you can contact us on firstname.lastname@example.org if you’d like some specific advice for your own circumstances.