Riki Williscroft, Cloud Product Owner, Auckland Transport
The consumption model and the billing API’s now available allow us to directly assign costs through show back or charge back in real time. Having a unified view across different hyperscalers means finding somewhere for your new app has become an open and dynamic market. It is a small thing but it will turn the IT department business model on its’ head allowing becoming a true services organisation seamlessly integrating with your wider digital program. Before you can do that though you need get to the Cloud and that can be painful.
If you have tried a move to the Cloud and not seen the promised savings and now have a growing OpEx problem you are not alone. Many organisations have been encouraged into the cloud with the promise of savings that have not been realised.
Cloud economics argues that direct costs are only a part of the overall picture and when balanced by organisation agility and reduction in overheads the savings and advantages become readily apparent. This is little comfort to a small/medium business owner faced with a growing quarterly invoice for cloud services.
Can you save money going to the cloud? Yes. How? Think differently. Ignore business cases that do not have directly attributable savings; make money, save money, risk, or compliance the only four cases that should fly. If productivity gains are cited then there should either be roles disestablished or clear links to revenue growth programs. Agility is a valid case but only if time to market will give you a competitive advantage or the reduced development/implementation time results in a lower cost of deployment.
It makes sense to time your migration around the asset lifecycle of the hardware. You can minimise the effect on the balance sheet and spread the workload over a period that will have less impact on your team. The downside is it will take longer to realise cost benefits. These will only kick in once you start reshaping your organisation to match an easier to support and maintain cloud portfolio. In a large enterprise the complexity of the technical landscape means you will likely have to stage the migration over a longer period. In this case it is critical to work with the finance team and senior leadership so they understand there will be a cost rise as services and licenses are duplicated in the cloud before there are savings. This may mean the usual three year return is pushed out to five or longer.
Time is not on our side. If you have recently bought hardware or added an on-premise application you can expect to be stuck with it for at least three to five years depending on depreciation. Now is the time to start planning for a cloud migration. If you are not ready to migrate when it comes time to renew, the path of least resistance is to re-platform on new hardware committing the organisation to another 3-5 on premise. If we think about that for a moment what will the cloud look like in ten years’ time and what will we have missed out on?