We were delighted to host the third quarter IDC Digital Leadership Think Talk of 2022 on September 29. Around 40 digital leaders from across Europe joined the call to share their challenges, successes and experiences of data and how it is managed within their organisation.
Marc Dowd and Tracy Keeling from the IDC Executive Advisory team led the discussions.
All About Data
At the start of the session, Marc Dowd set the scene by asking the audience to think about how they manage data access, secure data as it is transferred, manage data cleansing, data integrity, duplication, building data culture, and how to align the business and IT with data.
Our first CIO contributor explained that you need a data glossary as well as a data catalogue as everyone will have a different definition of what they believe data is and what it involves. Secondly, it is important to have a clear view of the data types, including human and machine created data, and how to manage it via stewardship.
Another contributor described how they had also used data journey maps to show where data comes into or is generated by the organisation and how it is transformed and stored as live or archived and allocated ownership to the people closest to the data mainly within the business. This helps them to manage the difference with dealing with your own data versus external data, which may have to be given back to the owner in some cases or returned to you by the vendor at the end of a contract if moving to another product.
Ownership and Management
This sparked a lot of comments from other CIOs on the call — one spoke about democratization to give bounds to data by starting with regulatory rules and GDPR, then you can decide who can have access, how they can access it, and how long it should be kept.
Our next contributor, from the government industry, responded to the question on whether data stewardship is an IT function or business function.
They compared IT to a car leasing company that provides the asset with a set of rules. It is up to business owner to drive the car (or use the data) in the way they want. This was reiterated by another member who mentioned the exception — that machine-generated IoT data, by far their biggest source, is managed by technical teams.
A further comment was around data management in government. One CIO felt it depends on the size of the organisation and that it should sit in IT, but only if the organisation is not too large. If it is a private sector organisation, they felt it was best to push ownership towards the business side, with business ownership of the quality of the data. IT could then manage the system side of where data sits and flows.
Another member from the pharmaceutical industry explained the segregated model they use. One area was where they have a local person who is responsible for data from local trials, while a central team provides governance and IT provides the infrastructure for it to demonstrate and manage its integrity.
Practical Examples
The discussion then moved towards practical examples. We discussed low code/no code and how data was managed within this. Although they worked differently to standard applications, the participants agreed they still need system architects to create a reliable system.
An example of this was an AI low code/no code chatbot solution for doctors. Another contributor highlighted the complexity by explaining that they can consume data from 25 different sources. The healthcare business stakeholders were responsible for the quality of data and set up content checkers to ensure that what the doctors created was understandable.
Another CIO said they asked an important question before they created the data — do they really need the data they intend to collect or create? Once you have established what you need and why it is important, you have to check the quality of the data. The example discussed was around automated systems that are taught about relationships between entities and how to eliminate bias in algorithms for better quality data. Another person mentioned the need to have independence in the process and diversity in the people designing/reviewing algorithms for automatically generated data.
The conversation continued, demonstrating the wide range of data management strategies and use cases. The exchange of information demonstrated the value of these meetings and the value of peer conversations and experience.
Moving to Best Practices
One contributor talked about how they started their data journey with a small step. For example, if you are exploring data mesh or data virtualization, it is best to start with PoC in one region. If successful, you can productize across the whole company with a business change champion to face off to other areas of the business.
To achieve better data governance, you could either “blame” the risk compliance teams, etc., to get business to develop and stick to data governance rules (stick), or you could also explain to business the value they will get from smart use of data (carrot).
Another idea discussed involved getting line of business leaders to “sell” the data initiative to the rest of the organisation, then go back to IT to provide the tools. Sometimes, the result is that too many business-focused colleagues will ask for access — but that is a good problem to have.
Marc Dowd asked about approaches to governing the demand pipeline for data work, prompting several responses. One was if the business becomes super excited about new data streams, etc., you need a steering committee to prioritise data investments based on business return. Some contributors had set up monthly reviews with a data governance board and an executive committee making the final decision.
We discussed the “carrot or stick” approach and which one the contributors used in their businesses. The biggest “stick “is regulatory, often requiring a separate analytics platform gathering data from many systems, including people and machine generated data for compliance.
Just as important, even for the quality of data and even if there is a big “stick”, is that business users need to understand the value of the exercise, with one comment with a quote from Simon Sinek’s work to “repurpose the why”.
It was felt that leadership needs to understand that data-based decisions are better than pure intuition and be informed enough to know all measures are in place to trust the data. And they need to spread that belief throughout the organisation to promote data as an integral part of business operations.
As we ended the session, it was clear that the challenge of data was difficult as it involved not just the creation and journey through the organisation, management, and storage, but also the challenges around ownership and corporate culture to make data interesting and engaging to optimise its management and use to create insight.
The IDC CIO Advisory team would like to thank everyone who came to the call for their input. It is always inspiring to hear from those working with data challenges across the business. We hope this session was valuable and provided many takeaways for you.
Our next session will look in more detail at ICT governance — Does the new environment mean you need to make changes? We will be looking at new trends such as Agile and Pervasive governance.
If you already receive invitations to our sessions, I hope to see you there. If you would like to join this community, please email us at mdowd@idc.com.