Your First 100 Days as a New Chief Data Officer

In the context of an opening or a leave, you have just been appointed as the Chief Data Officer, Head of Data, or another title that places you in the role of organizing data teams and tools for an organization. Congratulations!

This role is complex due to the wide range of expectations and stakeholders. The CDO must simultaneously:

• Provide business value while orchestrating a technical organization,

• Deliver quick and visible impacts while guiding teams through a gradual transformation of their practices

• and the list goes on !

My experience in this role and in supporting CDOs leads me to suggest four main areas to develop in your first few months:

1. Build a roadmap

2. Position data as a Business Partner

3. Leverage a modern, scalable, reliable, and fitting data platform

4. Communicate / Be a preferred interlocutor for all functions

Build a Roadmap

The data roadmap is a shared vision between the Data Leader and all business units on how data will feed and structure the evolution of work processes.

A data roadmap is a progressive program that balances value production and technological maturity. It is not just a technological trajectory.

The data roadmap focuses on building trust in data and the value provided by data teams.

The data roadmap is fueled by:

• A comprehensive understanding of the business model and its value chain

• A clear assessment of available resources, including team maturity in data management and analytical solutions adoption, technical ecosystem, and existing skills…

• The expectations of your key sponsors, ideally starting with the CEO

Therefore, before diving into roadmap creation, you need to map your environment and possibilities, and to do this, you must empathize with business teams.

⚠️ Warning!

Do not systematically seek to reform the existing organization! The advice above mostly pertains to Data leaders entering an organization where everything needs to be established. Various organizations and architectures can emerge from corporate cultures and multiple historical constraints, of which you may not be aware but must quickly comprehend.

It may be necessary to start fresh, but this should be reserved for the most severe cases, involving serious dysfunction or posing an immediate danger, such as GDPR issues. However, changes in data direction are sometimes observed that bring about shifts not necessarily grounded in rational reasons, but rather tied to personal preferences or internal political considerations. This can significantly delay progress and be poorly received by the teams.

In practice

1. Map the Environment

   1. Meet with all stakeholders and create an initial stakeholder analysis for yourself.

   2. Map the organization's value chain starting from its mission/principal clients and moving through the main and secondary processes. Position the level of maturity of the underlying technology at each step.

   3. Map expressed needs based on their perceived value, technical complexity (prerequisites, development time), and level of sponsorship.

2. Propose a Roadmap

   1. Propose short-term improvements that will have a significant impact in terms of perceived value (reliability and time saved) but require minimal transformation.

   2. Plan to associate the development of reusable components with business use cases as soon as possible: shared data across multiple business units, data source quality monitoring, in-house data analysis and post-processing libraries, web tool structures, or machine learning flows.

   3. Above all, allocate bandwidth for ad hoc requests. They will help you adapt your roadmap on the fly and structure your offerings over time.¹

If you are taking over an existing data organization, prefer a change management approach that aims to begin by maximizing the potential of what already exists and concurrently initiate prototypes of modernization to iteratively validate their adoption.

Position Data as a Business Partner

To bring real progress to an organization, data teams should not just design and deliver "data products" but ensure they meet operational needs and are easy for users to adopt.

The majority of value is and will be be created by business units. It's about enriching and facilitating their daily practices.

Several actions need to be taken in parallel:

• Business/Data Collaboration

Business units often fail to engage data teams or make simple and non-contextualized requests due to a lack of understanding of data operations and capabilities. The goal is to establish collaborative practices between business and data. Gradually, mutual understanding and collective habits are built, making this collaboration beneficial for all.²

• Formalization and sharing of knowledge

Collaboration relies on shared knowledge, just as effective communication relies on a common vocabulary. Establishing a shared space for consensus on processes, metrics, and documentation, beyond saving time and increasing efficiency, is a necessary condition for governance actions.

• Cultivate a culture of measurement

Knowing why actions are taken, how to evaluate them afterward, and how to compare them makes the difference between data-mature organizations and those that are not. Data teams typically have a scientific background, making them well-placed to bring best practices to business units, starting by applying them to their own actions.

In practice

• Always ask your teams to place their tasks in a business context and consider possible technical solutions from the perspective of business users.

• Establish a centralized documentation of business processes and metric definitions as soon as possible. Initially, use it as an internal tool for your teams ("eat your own dog food") and refer your teams and stakeholders to it. Referring to it should become a habit.³

• Utilize ad hoc requests as opportunities to propose evaluation metrics for business actions. You can also suggest multiple ways to evaluate an action and their implications in terms of decision biases to instill doubt and emphasize the importance of definitions!

Leverage a Modern, Scalable, Reliable, and Suitable Data Platform

As a Data Leader, it's your responsibility to choose the technical stack your organization will invest in to advance its transformation and maturity.

This prospecting exercise is challenging in a dynamic environment with constant innovations and integration or specialization processes in specialized tools.

In this context, a technological trajectory should enable flexibility in decision-making, avoiding vendor lock-in, and keeping a watchful eye on innovations.

To achieve this, consistently focus on process formalization, documentation of business rules, and modular architecture design. This allows you to remain as technology-agnostic as possible and to be prepared to reduce costs and risks associated with upcoming technological shifts.

In practice

Whether you are in a medium-term legacy/on-premises environment or already in the cloud, establish or complete the fundamental components of a data platform:

• Data acquisition

• Data storage, data modeling, definition of pipelines/DAGs

• Development, testing, production deployment, orchestration of corresponding source code

• Building and maintaining dashboards

• Conducting in-depth analyses, developing machine learning models, and executing them in production

• Building and running dedicated applications embedding decision-support algorithms

• Managing knowledge and metadata: internal documentation, data catalogs

Communicate / Be a Preferred Interlocutor for All Functions

You have 100 days to make yourself as widely accepted as possible, as it will facilitate the adoption of all data initiatives you plan to launch. If there's one person in the entire organization's data who needs to be visible, it's you.

• Communicate with your CEO

It's not about connecting with the CEO only when you join and when you leave. The exercise starts even before your recruitment – ensure as much sponsorship and alignment (or at least willingness) at the CEO or intermediate level to which you report as possible. While in the role, make sure not to overlook strategically visible issues. This is not personal strategy: top-level support is a necessary condition for the success of the entire transformation that the organization aims to achieve through… you.

• Communicate with business units

Identify individuals in each business unit who show the most analytical interest. They will naturally become your data champions through ongoing requests. Organize regular meetups with these future data champions.

• Communicate with IT

Let's be clear: data, with its close ties to business and its technical underpinnings, can be a challenging client for IT, serving as a bridge between two different dynamics and operating modes. Spend as much time empathizing with IT as with business units. Learn to explain the challenges and constraints of each group to find a middle ground. Yes, you're a kind of diplomat.

• Communicate publicly

There are as many roles for CDOs as there are organizations that host them. In such a dynamic field, you need to draw from what your peers share about their experiences. But you should also share your unique, important, and inspiring experiences after these 100 days. Get out there!

---

That’s all for today!

As a CDO, what would be your astonishment report? Would you spend your first 100 days differently?

Your feedback helps me choose which topics to cover among the many that need attention, so feel free to contact me on LinkedIn or via email: gansanay AT gmail DOT com.

1. On this topic, you can read the previous letter: Handling Ad Hoc Requests: From Bottleneck to Business Partner

2. On this topic, you can read the two parts “Please, no ad hoc data requests any more” and Anatomy of an Ad Hoc Request

3. Use the Diataxis documentation framework, which distinguishes tutorials, how-tos, explanations, and references, as inspiration.