Chief Data Officer - creating a vision

5min read

The article uses illustrative examples from my experience working with Chief Data Officers or people in similar roles (e.g., Head of Data & Analytics). Specific situation might require a different focus than examples in the article.

It is written in the form of a Socratic dialogue (also known in 2024 as GenAI question and answer format). I promise it was not written by GenAI.

What would the conversation be if Chief Data Officer had an Ultimate Strategic Partner at their fingertips?

Q (CDO): I have 30 things in my head. Can I give you a laundry list and we take it from there?

A (CDO_USP): Sure. I have worked with many people in your position. It’s normal and I have the right experience to help you.

Q (CDO): I started my new role as a CDO a few months ago, inheriting many hot potatoes (adoption of data platform is low, no common data definitions, multiple data quality problems, business is not convinced of potential in data, data compliance issues) combined with high expectations from my boss (the COO) who wants me to deliver miracles. I’m also realizing our organization is very political and there are different competing interests. And above all, I only have a small team, limited own budget and authority on the topics. It feels like I need to sell my role and value first before I can really start working. What should I do?

What successful CDOs get right

A (CDO_USP): It’s a lot, but you’re spot on – you cannot do it on your own and this is at the core of the CDO role – namely to enable the wider company to create value from data by improving the data value chain.

Successful CDOs get the following three things right:

1. They have the right vision – to describe a data-augmented future that inspires and excites people, to bring clarity to what is to be done & why, to get buy-in from key stakeholders, to create objectives for their team that drive high performance, and to be able to demonstrate value against transparent KPIs that matter.
2. They have the right roadmap and execute on it to operationalize the vision – this requires focus on the right priorities, having the right resources, partnering with the right vendors and internal teams, and overcoming blockers that inevitably present themselves on the journey.
3. They deliver tangible incremental value for their stakeholders and data users – this is important to demonstrate credibility that keeps stakeholders engaged and accumulates trust. Best evidence of that are happy internal customers (work of CDO helps them deliver more with data) that want to increase collaboration and serve as a champions to other teams.

Why CDOs need a vision

Q (CDO): Ok, I see the logic. What is a good vision?

A (CDO_USP): A good vision helps you, your team, and the broader company. These are the three angles to the vision that should fit together – like nested dolls. And when you stack them together, they reinforce each other and give direction and motivation to the team around you and broader stakeholders.

• What is a vision for yourself, for your career? Where do you want to be in 3 years? Example: CDO is part of the C-Suite or at minimum a key Leader that C-Suite trusts and involves in making key strategic decisions. C-Suite and the rest of organization sees CDO team as an enabler, based on results CDO has delivered in the last 3 years.
• What is a vision for your team? Where do you want your team to be in 3 years? Example: Business teams are delivering results and taking advantage of common solutions enabled by the CDO. Key strategic initiatives (e.g., AI-powered Sales Advisor for Sales Representatives, conversational data portal for all employees) are having a clear top- and bottom-line impact. CDO operationalizes common data governance standards that help everyone to develop and share data products, provides targeted support in cross-functional squads that develop data products, and delivers a set of data products used by the Business.
• What is the data vision for the company? What future should be achieved in 3 years? Example: Data is democratized – everyone in the company can ask questions of data in natural language and get the answer, people across different business functions (Marketing, Sales, Supply Chain, Finance etc.) make data-augmented decisions every day (e.g., where to spend marketing spend, which customers to target with what message, what products to produce in what quantities). Share of data services has grown from 5% to 25% of our revenue – this has increased our overall operating profit margin from 15% to 20%.

Q (CDO): Hold on. This sounds like rocket science. We will never be able to achieve this. I need to focus on pragmatic things ahead of me – classification of sensitive data, access management mechanism that adheres to access rules and is efficient, and data quality issues. Can you help me?

Q (CDO_USP): Yes, I understand there are operational pressing matters, and we can certainly help. They need to be handled and they fit into the tangible incremental value mentioned before. But the right way to address them is to link them to the vision, to the big picture that excites people. In practice this means that the vision must be put in motion in parallel to tackling the operational topics that deliver quick wins. Over time operational topics converge to execution “to-do list” on the roadmap that delivers the vision.

Q (CDO): Ok, to paraphrase what you’re saying is that “just doing the topics that I know need to be done is not enough to have a staying power and grow as the CDO”. I need to link them to the vision that delivers business value. How do I start shaping the vision, who do I need to involve?

A (CDO_USP): Yes, precisely. The pure data governance and management topics are dry and won’t get sustained buy-in. There is a strong need to link them to ultimate value, which is benefiting the data users and is ultimately reflected in higher revenue or reduced cost.

How to create a compelling data vision

CDO will pull the vision together, communicate it, and lead (parts of) the execution but they will need input and support from others. It’s like being a football coach that needs to interact with players (his team + other teams), the club management (C Suite), and the fans (data users).

Concretely, here are some steps CDO can take to craft the vision:

• Understand strategic priorities and challenges of important stakeholders (CTO, CRO, Heads of Business Units) – what do they try to accomplish?
• Get to know initiatives already running or in planning, which CDO can enable (e.g., transition to new data platform, large data product investments – e.g., Next Best Action for Sales representative) – the work that CDO does needs to be useful to these initiatives.
• Do the first mapping of potential data topics that should work in a consistent, common way across the company because this would help with delivery of data products across the company, map out the value chain to demonstrate what areas to focus on for improvements.
• Get to know the current physical reality of these potential data topics – e.g., what are the activities that bring data from sources to data platform and transform it to common data products (how is access granted, how is DQ ensured, is there a need for common definitions)
• Draft a first version of the vision and bring stakeholders on board and use their knowledge to improve the vision.

Ultimately, the vision needs to have …

• WHAT the future is that we want to drive toward
• WHAT the benefit will be / WHY it is meaningful to deliver this future
• WHAT the proposed topics are that will in combination deliver this change
• HOW we will achieve this change (roadmap - activities, effort, resources)

… and should be agreed by stakeholders that provide resources to make it happen.

Example of a data vision - Everyone in the company can ask questions of data in natural language

Q (CDO): Ok, can you give me a concrete example of the vision and the benefits it unlocks?

A (CDO_USP): Sure, this of course is an example of the vision. The right vision for your situation may very well be different.

WHAT is the future that we want to drive toward

Example: Data is democratized – everyone in the company can ask questions of data (in line with access rules) in natural language and get an answer. This has significant impact on the revenue side (use cases and decisions that can lead to increase sales can be done much faster and in greater depth and breadth) and on the cost side (delivering any business improvements can be done faster and costs less because data is available in natural language.

WHAT the benefit will be / WHY it is meaningful to deliver this future

Benefits need to be looked top-down and validated bottom up through the lens of specific use cases.

Let me give you a few concrete examples how the above vision when realized impacts business outcome and will then circle back to a more general reasoning.

Example – making customer service and product improvements based on customer and market feedback:

• Imagine you sell a product that has many customer reviews – some are online product reviews from customers, some are expert 3rd party assessments and reports, others are from direct interactions that customers had with your customer support team.
• If you can interact with all this data in natural language, you can much faster sense what customers think of your product, what they expect, what are deficiencies, and how that sentiment changes over time. This can feed into decisions to improve the processes for customer interaction (see this recent case study) or your product roadmap.
• Compare this scenario to often typical situation where this information is separate, there is specific analysis on it done once every year (because it’s too costly to do it more often), and where the feedback only reaches the people who can make improvements after months (or never).

Example – assessing shortlisted targets for acquisition and making the right decision:

• Imagine you’re in the process to buy a company in an adjacent market where you want to grow. You have shortlisted a few target companies and they have accepted your interest. At this point you’re collecting different data in so called “virtual data room”, which includes excels, PDF documents, perhaps even code repositories or audio/video recordings of board meetings or interviews. This data covers all aspects of the business (financial, operational, legal, strategic etc.).
• If you can interact with all this data in natural language, you can understand the connections between data better, spot areas where data does not provide sufficient answer to your questions or areas where there is discrepancy to be investigated. Ultimately, you create a more thorough decision process that leads to a better decision (what target to acquire and at what price).
• Compare this scenario to often typical situation where time is tight and multiple teams analyse the information but still miss on important topics, and analysis is consumed by decision makers who have follow-up requests, which requires new analysis for which there is not enough time to complete.

General top-down reasoning as to why this vision is meaningful:

• This vision is bridging that gap between data/tech people who can analyse large data and functional experts who understand the business but may not be able to code.
• This vision removes the time to decision because the data does not need to go through multiple teams in a classical processing chain.
• This vision allows for interaction and quick pivots, just like a real conversation does.

Q (CDO): Hold on, this sound very business related.? What will need to be done to get this vision into reality? I want to know concrete building blocks that need to be put in place. I will need to go now, but I want you to think about what it would take to realize this vision and we can discuss next time.

A (CDO_USP): It is business related because any improvement you aim to deliver needs to follow from business benefits. Therefore, the work that you will do on data & tech side needs to be linked to business benefits from the start. Looking forward to discussing concrete steps to progress towards the “Data is democratized” vision in our next session.