Implementing a Cloud Data Platform: The Do’s and Don’t

3 June 2021 Updated at 16 June 2021

The Cloud is presenting great opportunities in terms of performance and scalability for companies. However, considering some golden rules as a starting point to mitigate risks and ensure your company is ready for a successful cloud strategy implementation is highly advisable.   

We already stated how powerful Cloud Data Platforms are for companies, enabling new use cases and offering great scalability and flexibility at the same time. As your company is now ready to move to the Cloud, you should be aware that cloud migration can be challenging as some internal prerequisites including your companies’ objectives and strategy are key to initiate a successful migration.

Understand your objectives

Cloud Data Platforms offer strong advantages in terms of flexibility, agility, IT costs reduction and have serious competitive advantages in the market. We understood how harnessing the power of data for a company’s digital transformation is important, but a clear strategy vision is needed for such a move. Therefore, as a company, you must address the reasons why you want to implement a Cloud Data Platform, meaning you must clearly define the challenges this platform will resolve for you.

It is important to identify what the performance expectations are and what the expected ROI would be by moving to the Cloud. The best indicator for this part is probably your use cases which are usually the drivers of change. This is what we call a top-down approach that goes from company strategy, to business goals, all the way down to building competencies in the Cloud.

At Business & Decision we notice that companies usually want to switch to a Cloud Data Platform solution because their existing (legacy) data platforms don’t provide sufficient agility nor are they able to respond to modern needs usually related to the AI revolution (advanced analytics use cases like machine or deep learning, real-time use cases for fraud, predictive maintenance or text-mining for further knowledge discovery etc.).

Companies realize that moving to the Cloud can be done quickly in an agile and easy way and present competitive advantages in delivering more performance and better results while keeping the incremental costs low.

How to implement a Cloud Platform smoothly?

You do not need to think BIG for your data

For this part, you don’t necessary need to start big. Cloud enables us to begin with a use case in smaller increments (or iterations) which can be scaled up with time (from a POC to MVP and finally a Scale Up approach). As we like to say: “think small, fail small”.


It is important to avoid the confusion between the Big Data Movement and the Cloud Data Platform. Cloud Data Platforms are fit for different projects regardless of their size (big, small, wide, deep data) and nature (descriptive, predictive, prescriptive or cognitive analytics). They can easily combine different patterns at different times: you can opt for a small platform or a bigger one or even both at the same time while still optimizing costs by paying what you truly consume. This is one of the main differences between a Cloud Data Platform compared to the classic and traditional Data Platforms, like Hadoop or BI, which cannot offer the same flexibility to our clients.

It is not because your company is in the Cloud that it means you do not have any infrastructure to manage. Companies still have the choice to manage nothing, only core application or the whole infrastructure depending on the setup they choose.


It is important to avoid the confusion between the Big Data Movement and the Cloud Data Platform. Cloud Data Platforms are fit for different projects regardless of their size (big, small, wide, deep data) and nature (descriptive, predictive, prescriptive or cognitive analytics). They can easily combine different patterns at different times: you can opt for a small platform or a bigger one or even both at the same time while still optimizing costs by paying what you truly consume.

This is one of the main differences between a Cloud Data Platform compared to the classic and traditional Data Platforms, like Hadoop or BI, which cannot offer the same flexibility to our clients.

Check your internal requirements

Your company must ensure a validation from the organization to initiate a Cloud Data Platform solution implementation. We often see enterprises who already made some pre-requisites regarding the skills and solutions they need whilst other companies can be at a different stage regarding a Cloud Platform implementation (“Should I move to the Cloud or not?”).

Ensure your company infrastructure readiness!

It is not because your company is in the Cloud that it means you do not have any infrastructure to manage. Companies still have the choice to manage nothing, only core application or the whole infrastructure depending on the setup they choose (IaaS, PaaS, SaaS). In any case, you must ensure you are ready to manage your infrastructure on a strategic and organizational level.

What we noticed at Business & Decision is that most companies make the choice to move to the Cloud because of the required costs and resources to manage the back-end infrastructure of the platform. Therefore, we believe in the PaaS (Platform as a Service) approach which offers at the same time the infrastructure from an external provider and a software provider platform to manage data.

In case you would opt for a PaaS solution you would still need governance. What should be considered is the system of “credits” that most of the modern Cloud platforms put in place for companies to pay depending on the amount of work that has been done. You will need to ensure your Data teams are not spending your credits too fast or for unnecessary actions.

What you should also have in mind is the internal resources for the management of the Cloud Data Platform. The profiles for managing these platforms are changing and a lot of Database Administrator’s tasks are already pre-packaged in most of the solutions. We see a shift towards the requests monitoring being executed and how to fine-tune those requests in terms of consumption. For instance, before, Database Administrators needed to ensure the databases could handle the workload with limited resources, but now in an “unlimited capacity world” and thanks to the autoscaling, they need to monitor how the resources are consumed and advise on best practices in terms of querying and data access.

Define your Cloud solutions

Your company should prepare various benchmarks to analyze what solutions would fit it best, in terms of Cloud Solutions as well as on-premises solutions.

Component’s compilation

In comparison to before, where we talked about stack-based offering, a set of tools, solutions and programming languages were combined in one packaged offer, with Cloud Data Platforms, you can create your own customized compilation of services, tailored to your needs and use cases. You can define the components that are important to achieve your objectives, even if they do not come from the same vendor. In this way you have the capability to create a “best of breed” and fit your company’s specific needs as best as possible.

Get your deployment ready 

Some choices need to be done before deploying your Cloud Data Platform to your company depending on your organization’s strategy: are you opting for a cloud or multi-cloud option? What service vendor are you selecting? Once you are set up for this, you will need to have your subscription and credits to get started.

This selection will depend on your company’s needs and challenges, each Cloud provider has got its own specificities. Thanks to our domain expertise, vast experience and best practices, we offer quality services in guiding our customers in that journey.

Go Live!

Once your Cloud Data platform is ready to implement, you just need now to get started using the tool.
At Business & Decision, our experts can guide you at different stages of your digital journey from implementation deployment or maintenance.

This article was written by Luka Riester, Innovation Tribe Lead at Business & Decision in collaboration with Valérie Dézérable, Faction Lead Data Engineering & Architecture at Business & Decision, Benjamin Protais, Tribe Lead Data Management Europe at Business & Decision and Sébastien Trébois, Senior Data Management Consultant at Business & Decision.

Luka Riester Data Innovation Professional
Business & Decision - Belgium

With more than 12 years in the data & digital consulting industry Luka is now responsible for a team of pre-sales experts and is leading consulting and pre-sales engagements for a prominent set of customers in the domain of data innovation. With his team, Luka…

Learn more

Leave a comment

Your email address will not be published. Required fields are marked *

Your email address is only used by Business & Decision, the controller, to process your request and to send any Business & Decision communication related to your request only. Learn more about managing your data and your rights.