Getting everyone on board for your data platform cloudification project: 4 major hurdles to take
5 min.
Your data platform has to move to the cloud: you yourself are fully and absolutely convinced of the ever more urgent need to make that happen. More importantly, you see the many potential benefits of such a move.
This is the easy part, however, for now you still have to convince all other internal stakeholders involved, if you want your cloudification project actually to succeed. So this means selling your project internally, to your colleagues or at least to those colleagues who have the authority and necessary leverage to get your entire organisation on board.
All in all, there are four major hurdles to clear before you can really kick off such an ambitious, strategic project. In what follows, we will explain how you can best prepare your organisation to take those hurdles. In other words: which checks you need to perform and tasks to do, to gain a better insight in your cloudification initiative and turn all the stakeholders in your organisation into co-promoters.
Technical hurdles: some basics to check with your CIO
Check your bandwidth
You will need to check and test the effective bandwidth and quality of the network that connects your data centre with your cloud provider, enabling the up- and downloading of data between both.
Check your data transfer volumes
In your network analysis, you also need to take into account your current way of working with your data platform and the resulting data volumes that will be transferred between your local and cloud environment. If that turns out to be a potential bottleneck, you should re-evaluate it and seriously consider adopting new and more appropriate techniques and potential design changes. Otherwise this could turn out to be a complete showstopper!
Check your load strategy
Talking about showstoppers, do not forget to implement the basics of data management in your ingestion strategy. Check for instance where a full load strategy or a delta load strategy, using a change data capture (CDC) solution, would be appropriate or even required.
Legal hurdles: check the regulatory constraints with your CISO, DPO or CDO
This ought to be a no-brainer, really, but do not forget to check the data security and privacy regulations with your Legal department and your Chief Information Security Officer (CISO), Data Protection Officer (DPO) and/or Chief Data Officer (CDO). Make sure you have these colleagues on board from the outset, so you can avoid unpleasant surprises or even showstoppers along the way. Getting upfront security clearance will allow you to proceed without interruption with the execution of the cloud roadmap for your data platform.
This is also where you decide which data to move to the cloud and which to keep on-premises. As a rule, non-sensitve data tend to be moved more easily to the cloud, whereas for example personal or financial data tend to rather remain on-premises.
You can read more about the legal and regulatory hurdles in our blog about the business lessons learned when moving to the cloud.
Financial hurdles: keeping costs under control (or is there truly no limit on your credit card? 😉)
When taking on a strategic project of this kind, you most certainly want to avoid financial surprises as much as technical surprises. That’s why you better engage with your CFO and/or your financial department.
Check your business case
First and foremost, consider what you could do more or better in the cloud. Migrating your data platform to the cloud could present itself, for instance, as an opportunity to implement advanced analytics, data science or self-service on a broader scale. You can read more about the importance of finding the right business drivers for your cloud migration project here.
Discuss your data governance strategy
Don’t hesitate to use your cloud initiative to correctly implement your data governance strategy. Also, do not forget or avoid to discuss and clear out some important governance issues with data user community. Here are just two examples:
- Align with your data scientists what freedom they can obtain to ingest and process data.
- Check with your reporting user community how far they want to go in self-servicing for the data ingestion/connection part of your platform.
Calculate your TCO
As with any architecture exercise, do not forget to do a TCO calculation at the start of your assessment. You can make convenient use of the built-in cloud cost calculators and advisors of the different cloud platforms.
Monitor continuously
Even more importantly: do not forget to do some financial monitoring during the execution of your cloud migration project. Experience has shown that this needs to be a continuous exercise in order to avoid unpleasant events, such as the sudden explosion of your monthly cloud invoice.
You also need to continuously track your background processes as well as the usage of your reporting/analytics environment, to be sure that appropriate actions can be taken to counter the highly intensive and therefore costly usage of your data platform.
‘Human’ hurdles: technology without people won’t work
Communication and training – two cornerstones of the change management process – are also key to get your colleagues and, more importantly, your end users and managers on board. Remember above all to keep it real, and find the balance between your long- and short-term goals.
Stay realistic: go for quick wins but also work on longer-term benefits
One of the worst things that can go wrong in any cloud project, is that you managed to oversell the cloud to your colleagues in and outside IT. There is nothing that undermines and endangers the successful execution of a cloud roadmap so much as the mere idea or suggestion that the cloud will solve all issues and problems.
The cloud is not some kind of technological Walhalla, so remain pragmatic using it. Try to find some quick wins instead, this to help you get the buy-in from your user community. Here’s just a handful of quick examples:
- Set up a proof-of-concept (POC) to showcase how easy it is to upscale and downscale your IT resources in the cloud. This can show flexibility in performance and speed-to-delivery but also from a cost perspective.
- Demonstrate the strong disaster recovery functionalities the cloud has to offer.
- Optimize cost and flexibility on your DEV & QA environment.
- Find the use cases where a cloud solution is explicitly showing its strengths in flexibility and scalability. Take a case, for instance, where you have a combination of structured and non-structured data, the majority of which is already in the cloud, and where some users are already experimenting with cloud solutions.
Most of these arguments will also have a positive contribution in the business case if identified and addressed well.
Develop the specific cloud competencies you require
As with any change in existing technology or adoption of new technologies, you need to invest in a training plan or programme to develop those specific cloud competences that can assure the successful execution of your cloud roadmap. In other words: be prepared to make the necessary investments to define and execute a detailed training plan for your internal staff,or find an experienced and competent partner to help you get started with it. Outsourcing the service of the new environment is of course also a valid option.
In any case, do not underestimate the mental and cultural changes that are required from your organisation in general and some of the users of your data platform in particular, whether they are working in an IT or business environment.
And last but not least, should you get confronted with an acute lack of in-house expertise, don’t hesitate to use some plug-and-play PaaS components to make your life – or rather your work – somewhat easier. Finally always keep in mind that Rome too wasn’t built in a day!
Assess for success! That is our motto at LACO. But maybe you would like to assess us for a change?
Samuel De Klerck
CDO at LACO
