But wait a minute. How do we get there? Can we just transform the spreadsheets to a fancy PowerBI dashboard and leave it with that? The journey is longer and luckily it’s also very exciting.
Let me share 5 of the first experiences on the journey to becoming a data-driven company.
# 1. Select the right technical data-driven platform
To become a data-driven company you need a place to host and grow your data - a data platform. We are talking huge and growing amounts of data so you need a flexible and cost-efficient platform built for the future.
This can be a quick and easy choice, you just listen to what the consultants and industry experts recommend for you. And they will of course recommend that you build your data-driven home in the cloud of their choice. Can’t blame them for that, we all have our preferences. But keep in mind that you are the one who will be the owner and maintainer of the data long after the consultants are gone. You need something for the long run.
Since this is a very important decision it is worth taking time, to investigate and compare before the decision is made. Take into account all the requirements. As an example, it can be in the areas of compliance, technology, security and cost:
Do you have sensitive data that need to be stored on-premises? Is the public cloud an option? What kind of in-house competence do you have? Which training options do you have access to? What kind of agreements and discounts can you get with the different vendors? Are there any special requirements for your data? Any new regulative and compliance issues you need to take into account? Does the platform support the business continuity and disaster recovery plans you need for those vital tools? Are you aware of changes in data in the future? Et cetera et cetera. It’s a lot to consider at this important stage of the development.
Select 3 alternatives, run some POCs, talk to your data community and do a structured analysis. And don’t forget to document the process. 2 years from now people will ask you why you made this choice and you can refer to the documented decision.
# 2. Define the data-driven strategy and follow it
When the data platform is up and running you need to have a long-term plan. Where do you want to go with your data platform? A tempting solution is to grow the data lake with as much data as you can find. A rich data platform is gold, but it should bring value from the start so early adoption might be a smarter first goal.
A common strategy is to invest time in building a client on top of the platform to make the data visually available for all the stakeholders. This means your audience doesn’t need technical knowledge to use it for their analysis and needs. Again — spend some time choosing the right client before you commit. License costs can skyrocket when the product gets popular.
Another good thing to invest in from the start is hygienic factors such as data governance. This gets more complex to introduce the longer you wait so it is good to start early.
The data strategy should always align with the overall company strategy. If for example innovation is high on that list you should focus more on laying the ground for it by providing the data the organisation needs for the innovation initiatives. If for example, cost efficiency is the top company strategy you should focus more on providing data for financial analysis and forecasting.
No strategy is set in stone, you can adjust as you go. Just make sure to communicate it and involve the people so that everybody knows which way to go at all times.
# 3. Ensure ownership and adoption
To keep the pace up and the development going you need support, investment and attention from the main stakeholders. They can be both internal and external.
Even if engagement can mean a constant stream of Slack messages, questions and emails it is the best situation you can be in. Take a deep breath and get control of your backlog. Because it means that people care and have a desire to see your product evolve. Ownership and involvement from top management are also a must. Data should be an interest of the whole company, not just the Technology or the Product department. Without support from the whole organisation, the area will not get priority and can stop evolving.
It is up to you to make your product popular. Provide webinars, training and Q&A sessions. Put up the monitors by the coffee machines. Invest in good user interfaces to make them attractive and easy to use.
# 4. Prioritise right
In the start, you will most likely have a very big backlog with tons of small and bigger initiatives, bug fixes and requests for new development. Keep the stakeholders warm by involving them in the priorities, and keep the backlog transparent.
Always keep in mind to do what brings value, not what is easiest to do. For example, should you allocate the next sprints to integrate with the CRM system, or should you work on transaction forecasting? Find out what brings value to the company, and always refer to the strategy. In case of a prioritisation conflict, bring the stakeholders together to align on the decision.
One can also argue that a middle road can work here. You don’t want to get stuck spending months trying to implement something that might not work out as planned due to, for example, bad data quality. Evaluate what is within reach before you start. It should both bring value and be realistic.
# 5. Improvement over time
At the beginning of your data-driven journey, you should not strive for perfection. Ok, so you get a lot of your data from manual spreadsheets. We have all been there. That is fine as long as you have a plan and know where the real data source is, not everything needs to be perfect from day one. Get the dashboards out early based on the spreadsheets and start to gather feedback from the users.
Manual jobs are also ok in the start phase. Maybe not all batch jobs are automated yet and you need to click some buttons to get the imports running, there is probably some bad data quality and missing documentation. Evaluate what you can live with, and what can wait until later. In most cases, it is more valuable to get the MVP out than to have a clean kitchen in the backend.
More advanced methods, like machine learning and the use of AI, can be slowly introduced. Start small and learn from it, become smarter with experience. Seek expert advice on complex matters.
Your data will grow over time. In the beginning, you will probably not have big amounts of historical data. That is ok. Data can also be mocked and simulated. Use what you have first and make a plan for how to grow it. Your data-driven platform will get richer day by day.
Via iterative development and MVPs, you will continuously adapt to the changing needs of the business. A data platform is not a static product. What was relevant and correct yesterday might be something else tomorrow. The data is used in different ways and fetched from different sources. This is also what makes it fun. You learn and get more insight as you go on your way to becoming a data-driven organisation.
These were a few takeaways from the start of our exciting data-driven journey. Stay tuned for more experiences from the development soon. Good luck on your journey and please don't hesitate to reach out if you would to share experiences or discuss with us.