With technological advancements, like AI, good data is more important than ever. To get good data, you need a good data strategy.
A good data strategy is an important part of becoming more data-driven and ready for AI. To stop guessing and start knowing.
A good data strategy should focus on your data foundation. Data foundation encompasses all the data in your organization that you collect for reporting, analytics and insights. This is what helps you get the overview of what’s going on, allowing for better decision making and strategic growth.
Creating a good data foundation is a very big topic – and it always depends on your organization, your industry, your strategic aims, and the context you operate in.
In the following, I’ll outline six steps to go through to create your data foundation. These are the things you need to consider and good questions to ask in order to set up the right procedures and select the right technology to ensure you’ll always have the data you need.
One of the main considerations is data gathering – simply identifying where you need to gather data from. Which systems are relevant? Which parts of the organization?
I always recommend gathering data before you actually need it. Take the example of starting a new project. When starting a new project, you should gather information before as well as after it begins. This sets you up to compare the before situation with the after situation, so you can assess whether the project has led to the desired change.
The way you collect data from your source systems is an important part of your data foundation. It should be flexible enough to keep track of any customizations or modifications. For example, if you add a new property field, your data capturing method should adapt to this and include the new field in the collection of data, e.g. you decide to track a new KPI on all your projects. A flexible, adaptable data capturing method means you get reports that always include the latest information.
A flexible data gathering method ensures that you always have as much information as possible and the right information. That way you avoid a situation down the line where you might be missing important historical data from a project or initiative.
I hear it a lot: end-users have it as a requirement that reports are updated real-time. However, I always try to challenge that. Is there a legitimate need for real-time reporting or is it just a short-hand for frequently updated?
In my experience, it’s much more valuable to consider how often do we need our reports updated. In practical terms, hourly refreshes of reports are often more than enough. In fact, it can provide a faster and smoother user experience.
However, it’s always a good idea to keep your data foundation real-time updated. Keep that inflow of data incoming because that’s what keeps you flexible. You can tweak how often your reports should be updated and you have the option to make real-time reports if necessary – but again, it might come at a cost.
Having a centralized database for collecting the data makes it simpler to use the data. There is one place you go to retrieve data for use and you have more options for creating logic and combining data from different sources. I find it’s often in the combination of data from different sources that real value can be found.
A centralized database is also a matter of security. It’s easier to control permissions and access to data when it’s in a central place instead of distributed across databases.
Your data should be available to wherever it can create value – and to who it can create value for.
For example, your BI specialist who wants to create a new dataset for a report or analytically minded project manager who wants to connect excel to the dataset to get a step deeper into insights than the current reports available to her.
One thing is the technical aspect. Of course, you need your data foundation to be available to the required technical platforms and tools, in a secure way. But you also need the data to be available in the right formats. You need to be able to work with the data in the relevant contexts – which data is needed for which tasks and in which format? Minimizing the friction of format change ensures better use of your data and more accurate information.
Ensuring good data quality is important. It’s the number one thing that ensures that you can use your data to its full potential.
Good data quality starts with the end user – the people who are using the systems and providing the data. They should understand the value that great data brings to reporting for more accurate decision-making.
More importantly, the systems and processes should be set up to keep data quality high, helping end-users provide the right information in the right way.
It almost goes without saying, but you should have a stable system that you can rely on, so you’re never in doubt as to whether data has been updated or not in your reports. Trust and data quality go hand-in-hand.
A data foundation comes in many shapes and sizes. It could be a data warehouse, a datamart, a reporting database. Whatever technical form it takes, a data strategy goes beyond that. It starts with your strategy and organizational. With the six considerations for a building a good data foundation you’re on your way to create a data strategy that fulfills all your needs – and keeps you future proof.
