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 → is not a dashboard. It's not a report. It's not a raw dataset.
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 → is a curated, managed, and governed dataset that is treated with the same rigor as a software product: it has an owner, an SLA, documentation, versioning, and quality guarantees.
This distinction matters because it fundamentally changes the economics of data. Raw datasets are internal assets that disappear when the engineer who built them leaves. Data products are organizational assets that persist, scale, and accumulate value.
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 → has five defining characteristics:
1. Clear ownership, A named individual or team is accountable for this data productdata productA data asset managed like a product, with an owner, defined users, guaranteed quality, and measurable business value.Voir la définition complète →'s quality, availability, and evolution. Not "the data team" generically.
2. Defined consumers, The data productdata productA data asset managed like a product, with an owner, defined users, guaranteed quality, and measurable business value.Voir la définition complète → knows who uses it and why. Consumer requirements drive the product roadmap, not internal convenience.
3. Quality SLA, Committed freshness (data updated within X minutes), completeness (Y% non-null for critical fields), and validity (Z% of records pass business rules).
4. Documentation and discoverability, Consumers can find the product, understand its schemaschemaA schema is the formal blueprint that defines how data is structured, named, typed, and related within a database, file, or message.Voir la définition complète →, and trust its definitions without asking the owner.
5. Versioning and stability guarantees, Breaking changes follow a deprecation process. Consumers aren't surprised by schemaschemaA schema is the formal blueprint that defines how data is structured, named, typed, and related within a database, file, or message.Voir la définition complète → changes.
Vérification des acquis
1. According to the lesson, which company pioneered the Data Product Manager model?
2. How many defining characteristics of a data product does the lesson identify?
3. Based on the lesson's logic, why is treating data as a 'product' economically superior to treating it as a raw dataset?
4. Select ALL elements that the lesson defines as part of a Quality SLA for a data product:
Sélectionnez toutes les réponses correctes.
5. Select ALL correct statements about the Data Product Manager (DPM) role according to the lesson:
Sélectionnez toutes les réponses correctes.
Building data products at scale requires a new role: the Data ProductData ProductA data asset managed like a product, with an owner, defined users, guaranteed quality, and measurable business value.Voir la définition complète → Manager (DPM). This role sits at the intersection of domain expertise, technical understanding, and product management discipline.
A DPM for the customer domain would own: the customer 360 data productdata productA data asset managed like a product, with an owner, defined users, guaranteed quality, and measurable business value.Voir la définition complète →, the churn prediction feature storefeature storeA centralised repository managing ML features, ensuring consistency between training and serving environments.Voir la définition complète →, and the customer events stream. They work with data engineers to build and maintain these products, with domain stakeholders to understand requirements, and with governance teams to ensure compliance.
This is distinct from a traditional BIBITechnologies and processes that turn raw data into actionable insights via reporting, dashboards and analysis, so teams can decide based on facts rather than intuition.Voir la définition complète → analyst (who consumes data products) and a data engineer (who builds the infrastructure). The DPM owns the product lifecycle.
Airbnb pioneered this model. Their "data productdata productA data asset managed like a product, with an owner, defined users, guaranteed quality, and measurable business value.Voir la définition complète → managers" own specific data domains and are accountable for the quality of data products those domains produce. The result: clearer accountability, faster delivery, and higher quality.
The Domain Event Stream, A real-time stream of everything that happens in a domain. Example: all checkout events with standard schemaschemaA schema is the formal blueprint that defines how data is structured, named, typed, and related within a database, file, or message.Voir la définition complète →. Consumers build their specific views from this stream.
The Aggregate Entity, A curated view of a core business entity. Example: the "customer 360" that aggregates behavioral, transactional, and demographic data about each customer. High value, high maintenance.
The Feature Dataset, Precomputed ML features served to models in real-time or batch. Example: "customer purchase probability features" computed daily, served to the recommendation model. Managed by the ML platform team.
The Metric Dataset, Standardized, agreed-upon business metrics. Revenue, DAU, conversion rateconversion rateThe percentage of visitors or prospects who complete a desired action (purchase, sign-up, contact form), calculated as conversions divided by total opportunities.Voir la définition complète →, defined once, computed consistently, used everywhere. This eliminates the "why do the finance and marketing dashboards show different revenue numbers?" problem.
Data productData productA data asset managed like a product, with an owner, defined users, guaranteed quality, and measurable business value.Voir la définition complète → quality is not subjective. Define it with measurable SLAs:
Publish these SLAs. Track them. Alert when violated. Report them to data consumers. This is product management applied to data.
1. Quelle est la principale différence entre un data productdata productA data asset managed like a product, with an owner, defined users, guaranteed quality, and measurable business value.Voir la définition complète → et un dataset brut ?
A) Le data productdata productA data asset managed like a product, with an owner, defined users, guaranteed quality, and measurable business value.Voir la définition complète → est stocké dans un système différent
B) Le data productdata productA data asset managed like a product, with an owner, defined users, guaranteed quality, and measurable business value.Voir la définition complète → a un propriétaire, un SLA de qualité, de la documentation et un versioning, traité comme un produit logiciel
C) Le data productdata productA data asset managed like a product, with an owner, defined users, guaranteed quality, and measurable business value.Voir la définition complète → est uniquement pour les données en temps réel
D) Le data productdata productA data asset managed like a product, with an owner, defined users, guaranteed quality, and measurable business value.Voir la définition complète → est plus facile à maintenir
Réponse: B
2. Quel est le rôle du Data ProductData ProductA data asset managed like a product, with an owner, defined users, guaranteed quality, and measurable business value.Voir la définition complète → Manager (DPM) ?
A) Il remplace le data engineer
B) Il consomme les data products pour crcrThe percentage of visitors or prospects who complete a desired action (purchase, sign-up, contact form), calculated as conversions divided by total opportunities.Voir la définition complète →éererThe ratio of interactions (likes, comments, shares) to reach for a given piece of content, used to gauge how well audiences respond relative to how many people saw it.Voir la définition complète → des dashboards
C) Il possède le cycle de vie du data productdata productA data asset managed like a product, with an owner, defined users, guaranteed quality, and measurable business value.Voir la définition complète →, entre expertise domaine, technique et management produit
D) Il gère la sécurité des données
Réponse: C
3. Quelle métrique de qualité mesure si les données sont mises à jour suffisamment rapidement ?
A) Completeness
B) Accuracy
C) Freshness
D) Uptime
Réponse: C