A strategy without a roadmap is a vision statement on a conference room wall. A roadmap without a strategy is a list of projects that don't add up to anything.
The 18-month roadmap bridges these two failure modes. It translates your data vision into concrete initiatives with owners, timelines, budgets, and success metrics. And it gives you the tool you need to say "no", the CDO's most important skill.
Your strategy answers *why and what*: Why does data matter to this organization? What capabilities will you build? What will you be able to do that you can't do today?
Your roadmap answers *how and when*: In which order will you build these capabilities? What resources do you need? What does success look like at 6, 12, and 18 months?
The most common mistake: confusing a technology roadmap with a data strategy. "We will migrate to Snowflake and implement dbt by Q3" is not a data strategy. It's infrastructure planning. The strategy should drive technology choices, never the other way around.
Borrow from Google's innovation portfolio framework:
70%, Core programs: Initiatives directly serving your current business model and near-term priorities. Improving Data quality for customer analytics. Building the unified customer . Rationalizing your stack. These are your bread and butter, they keep the lights on and build credibility.
20%, Adjacent bets: Initiatives extending your capabilities into new areas. Building a Data productData productA data asset managed like a product, with an owner, defined users, guaranteed quality, and measurable business value.Voir la définition complète → for a new internal use case. Piloting machine learning in a high-value operational process. Exploring a data partnership.
10%, Strategic bets: Long-horizon, high-uncertainty initiatives. An external data monetization business. A real-time personalization engine at scale. An AI-driven operations center.
This allocation prevents two failure modes: spending all your budget on core infrastructure (safe but not transformational) or spending too much on moonshots before the foundation exists.
Vérification des acquis
1. According to the lesson, what fundamental question does a data STRATEGY answer, as opposed to a roadmap?
2. Why does the lesson consider 'We will migrate to Snowflake and implement dbt by Q3' a flawed example of a data strategy?
3. In the 70/20/10 portfolio approach, what is the primary role of the 70% 'Core programs'?
4. Select ALL statements that correctly describe the purpose of an 18-month roadmap according to the lesson.
Sélectionnez toutes les réponses correctes.
5. Select ALL failure modes that the 70/20/10 allocation is designed to prevent.
Sélectionnez toutes les réponses correctes.
OKRs (Objectives and Key Results) are the best mechanism for translating your 18-month roadmap into actionable plans.
Structure your OKRs at three levels:
Example:
Key discipline: fewer OKRs, not more. A data function tracking 40 KPIs is tracking nothing. Three to five critical outcomes get the attention they deserve.
The biggest roadmap mistake: sequencing technology before talent.
You can buy Snowflake, dbt, and Looker in 90 days. You cannot hire 20 senior data engineers, train 50 business analysts in data literacy, and build a governance culture in 90 days.
Sequence your roadmap accordingly:
1. People: Hire or upskill the team before you start building
2. Technology: Implement tools once you have the people to run them
3. Governance: Build governance around real use cases, not in the abstract
Organizations that fail at data transformation almost always got this sequence wrong. They bought the technology first, hired the people second, and never reached governance at all.
Ces actions sont compilées dans le plan d'action du rôle.