- How the DP-900 Blueprint Is Built
- Domain 1: Describe core data concepts
- Domain 2: Identify considerations for relational data on Azure
- Domain 3: Describe considerations for working with non-relational data on Azure
- Domain 4: Describe an analytics workload on Azure
- Reading the Weight Ranges Correctly
- Exam Mechanics That Shape How You Study
- Sequencing the Domains Over Four Weeks
- Where the Credential Fits Professionally
- Frequently Asked Questions
- DP-900 has four published domains; core data concepts and analytics workloads each carry the highest range, 25-30%.
- Relational data on Azure is published at 20-25%; non-relational data at 15-20%.
- Treat percentages as ranges. Microsoft does not publish exact per-domain item counts for the live exam.
- You need a scaled score of 700 or greater on a 1-1,000 scale, which is not 70% correct.
How the DP-900 Blueprint Is Built
Microsoft Certified: Azure Data Fundamentals (DP-900) is organized around four weighted objective areas published in Microsoft's study guide for Exam DP-900. The guide's skills-measured list was current as of July 21, 2026, and the guide page itself was last updated June 23, 2026. Those four areas are the only weighted structure Microsoft gives candidates, so every hour of preparation should map back to one of them.
Two caveats from the study guide are worth absorbing early. First, the subtopic bullets under each domain are illustrative: the guide states that related topics may also appear. Treat the bullets as a floor, not a ceiling. Second, the percentages are published as ranges. Anyone who tells you that a domain is "exactly 27%" of your exam, or that you will see a fixed number of questions from each area, is inventing precision that Microsoft does not provide.
If you are new to the credential itself, the explainer What Is DP-900 Certification? covers the basics, and the DP-900 Study Guide 2026 turns the blueprint into a full preparation path. This article stays focused on the blueprint: what each domain asks of you and where candidates tend to lose points.
Domain 1: Describe core data concepts (25-30%)
This domain shares the highest published range and is the conceptual foundation for the other three. It is vendor-neutral in flavor: you are expected to understand data as a discipline before you map anything to an Azure service. Candidates who skip it because it "feels easy" often discover that the vocabulary resurfaces inside questions from every other domain.
What candidates must understand
The vocabulary and patterns of data work, independent of any single product.
- Data representation: structured, semi-structured, and unstructured data, and how to recognize each from a short scenario (a table of customer records versus a JSON document versus an image library).
- Data storage options: file formats and stores, including what makes some formats suited to analytics and others to exchange between systems.
- Common data workloads: transactional processing versus analytical processing, and why the two have different design pressures.
- Roles and responsibilities: how database administrators, data engineers, and data analysts differ in what they own and what they deliver.
Where points are lost
The most common error is confusing the workload with the technology. A question may describe many small, concurrent reads and writes that must be consistent, and the tested idea is "transactional workload," not a product name. Practice stating the workload type first, then selecting the service. Another frequent trap is the role boundary: a question about preparing and moving data between systems points to the data engineer, while building reports and visualizations points to the analyst.
Domain 2: Identify considerations for relational data on Azure (20-25%)
This domain moves from concepts to Azure's relational offerings. The verb in the title matters: you are asked to identify considerations, which means choosing appropriately among options rather than configuring them. Expect scenario-style prompts in which a business requirement leads you to a deployment choice.
Core ideas to master
Relational concepts first, Azure relational services second.
- Relational fundamentals: tables, rows, columns, keys, and the purpose of normalization, plus indexes, views, and stored procedures at a conceptual level.
- SQL basics: recognizing the role of data definition, data manipulation, and query statements, and reading a simple query to understand what it returns.
- Azure relational services: how Azure SQL options differ (fully managed database service versus managed instance versus SQL Server on a virtual machine), and when open-source relational engines offered as managed Azure services fit.
- Management trade-offs: how much administration the customer retains under each deployment model.
The deployment-model question pattern
A reliable way to approach this domain is to rank the Azure options by how much control and responsibility the customer keeps. A scenario emphasizing minimal administration points toward the most managed option; a scenario emphasizing compatibility with an existing on-premises environment or operating-system-level access points toward the less managed end. Build a small comparison grid for yourself and quiz from it.
| Decision Factor | What the Question Is Really Testing |
|---|---|
| Minimal administration required | Recognizing the most fully managed relational option |
| Broad compatibility with existing server-based workloads | Recognizing a managed-instance style option |
| Need for operating-system-level control | Recognizing infrastructure-hosted database software |
| Open-source engine already in use | Recognizing managed open-source relational services |
Domain 3: Describe considerations for working with non-relational data on Azure (15-20%)
This is the lowest-weighted domain, but it is also the one that introduces the most unfamiliar names for candidates coming from a relational background. A lower range does not mean skippable: a domain at 15-20% still represents a meaningful slice of the exam, and the concepts here appear in comparison questions that reach back into Domains 1 and 4.
Core ideas to master
Know the storage types and when each data model fits.
- Azure Storage capabilities: blob storage for unstructured objects, file shares for shared access, and table storage for key-value style data, including why a hierarchical namespace matters for analytics scenarios.
- Azure Cosmos DB: its role as a globally distributed, multi-model database service, and the idea that different APIs expose different data models (document, key-value, column-family, graph).
- Non-relational data models: identifying when a document, key-value, column-family, or graph model suits the data described.
- Consistency and scale considerations: understanding at a conceptual level why distributed data stores trade off consistency, latency, and availability.
Matching the model to the scenario
Questions here typically describe the shape of the data and the access pattern, not the product. Highly connected entities where relationships are the point suggest a graph model. Self-contained records with varying fields suggest a document model. Simple lookups by key suggest key-value. Learn to read the scenario for these signals before you scan the answer choices.
Domain 4: Describe an analytics workload on Azure (25-30%)
Tied with Domain 1 for the highest published range, this is the domain where candidates most often underinvest because it sounds like "someone else's job." It is not optional. Note also a small wording point in Microsoft's materials: the certification page and the skills-at-a-glance list name this area "Describe an analytics workload on Azure," while the guide's detailed heading omits "on Azure." Both refer to the same objective area.
Core ideas to master
How raw data becomes insight, and which Azure services play which part.
- Analytical data stores: the idea of a data warehouse versus a data lake, and why analytics systems are shaped differently from transactional ones.
- Ingestion and processing: batch versus streaming data, and the extract-transform-load pattern and its variants.
- Azure analytics services: the roles of the major Azure data and analytics services in ingesting, transforming, storing, and querying data at scale.
- Real-time analytics: recognizing streaming scenarios and the services that process event data as it arrives.
- Data visualization: the role of a business intelligence tool in building reports and dashboards, including common chart choices and what each communicates.
Think in pipelines
The most effective mental model for this domain is a pipeline: sources feed ingestion, ingestion feeds processing and storage, storage feeds analysis, and analysis feeds visualization. For any service name you meet, place it on that pipeline. Questions frequently ask which service fits a particular stage, and a candidate who can locate the stage almost always narrows the choices to one.
Reading the Weight Ranges Correctly
| Domain | Published Range | Character of Questions |
|---|---|---|
| Describe core data concepts | 25-30% | Vocabulary, workload types, roles |
| Identify considerations for relational data on Azure | 20-25% | Service and deployment-model selection |
| Describe considerations for working with non-relational data on Azure | 15-20% | Storage types and data-model matching |
| Describe an analytics workload on Azure | 25-30% | Pipeline stages, services, visualization |
Two practical conclusions follow. Together, Domains 1 and 4 represent the largest share of the exam by published range, so a plan that gives them proportionally more time is defensible. Second, do not convert ranges into midpoints and treat them as exact. Microsoft does not disclose how many items from each area appear on a given form, so a safe plan covers all four domains with real depth and does not gamble on skipping the lightest one.
Key Takeaway
Use the ranges to allocate time, not to predict your exact exam. A candidate who is strong in Domains 1 and 4 but ignores Domain 3 is still exposed, because scenario questions routinely compare a relational choice with a non-relational one.
Exam Mechanics That Shape How You Study
Knowing the domains is half the picture; knowing the conditions under which you will demonstrate them is the other half. These facts come from Microsoft's published documentation:
- Delivery: a proctored, computer-based assessment through Pearson VUE, with online proctoring available in supported regions (ID, system, and room requirements apply). Certiport is also offered to students and educators.
- Time: 45 minutes for the assessment, within a 65-minute Fundamentals seat time.
- Questions: the count varies. Microsoft publishes no fixed total for DP-900, and no scored/unscored split has been verified. Certiport's exam-length listing indexes 43-48 items, but that is a provider-specific listing with a retrieval/version caveat, not a guaranteed count for your attempt.
- Format: the specific question-type mix is undisclosed, and the exam may be interactive.
- Resources: Microsoft Learn is not available during Fundamentals exams, so you cannot look up a service name mid-exam.
- Passing score: 700 or greater on a 1-1,000 scaled score. This is not "70% correct," and some questions may be unscored. See DP-900 Passing Score 2026 for the full explanation.
On cost: the issuer page's indexed rendering lists a U.S. list price of $99, but country, tax, and eligible discounts affect what you actually pay, and Microsoft publishes no general member/non-member tier. The details are in DP-900 Certification Cost 2026. For retakes, Microsoft's policy allows a second attempt after 24 hours following a first failure, then 14 days between later attempts, with at most five attempts in 12 months. Fundamentals credentials do not expire and have no renewal assessment.
Microsoft's pass rate is not publicly disclosed in the official sources reviewed, so treat any specific percentage you see elsewhere with suspicion; DP-900 Pass Rate 2026 discusses what can and cannot be said. And because Microsoft does not require a preceding certification or minimum training hours, DP-900 Requirements 2026 explains the recommended (not mandatory) background. For a realistic sense of effort, read How Hard Is the DP-900 Exam?
What no-lookup means for your preparation
Because you cannot consult documentation, passive recognition is not enough. For each Azure service in the blueprint, you should be able to produce, from memory, its category (relational, non-relational, analytics, or visualization), its primary use case, and one scenario where it would be the wrong choice. A one-page summary such as the DP-900 Cheat Sheet is a good final-week review tool, but build your own version as you study, since writing it forces recall.
Sequencing the Domains Over Four Weeks
One short plan, tied to the blueprint rather than to generic habits. The order below follows dependency: concepts first, because every later domain reuses the vocabulary.
Core data concepts
- Cover structured, semi-structured, and unstructured data; transactional versus analytical workloads; data roles.
- Do not rush: this domain's vocabulary reappears everywhere.
Relational data on Azure
- Learn relational fundamentals and SQL basics, then the Azure SQL deployment models.
- Build your control-versus-management comparison grid.
Non-relational data on Azure
- Study Azure Storage types and Cosmos DB, with the four data models.
- Practice naming the data model from a scenario description.
Analytics workloads, then mixed review
- Walk the ingestion-to-visualization pipeline and place each service on it.
- Finish with mixed-domain practice from the main practice test site, focusing on cross-domain comparison questions.
If you have only a short runway, compress by giving Domains 1 and 4 the larger share, but touch every domain. The DP-900 training overview describes the one-day Microsoft course and other preparation routes, and booking logistics are covered in DP-900 Exam Dates 2026.
Where the Credential Fits Professionally
DP-900 is a fundamentals-level credential. It signals data literacy on Azure rather than hands-on engineering depth, which makes it useful for people entering data roles, for sales and project staff who talk to technical teams, and for career changers building a first Microsoft credential. It is also a common stepping-stone toward role-based Azure data certifications. For how employers actually use it, see DP-900 Jobs; for compensation context, DP-900 Salary Guide 2026; and for a value judgment, Is the DP-900 Certification Worth It? The domains themselves are the best map of what a hiring manager will assume you can discuss: workload types, relational versus non-relational choices, and how analytics pipelines work.
Frequently Asked Questions
Microsoft publishes four: Describe core data concepts (25-30%), Identify considerations for relational data on Azure (20-25%), Describe considerations for working with non-relational data on Azure (15-20%), and Describe an analytics workload on Azure (25-30%).
Core data concepts and analytics workloads share the highest published range at 25-30% each. Microsoft does not disclose exact item counts per domain, so the ranges guide time allocation but do not predict your specific exam.
It is risky. Non-relational data is published at 15-20%, which is still a meaningful share, and its concepts appear in comparison questions that touch other domains. Study every domain, even if you give some more time than others.
No. The passing mark is a scaled score of 700 or greater on a 1-1,000 scale, which is not the same as 70% of questions correct. Some questions may be unscored, so you cannot compute a raw-percentage target.
No. Microsoft Learn access is unavailable during Fundamentals exams, so you must recall service names, categories, and use cases from memory. Practice retrieval rather than re-reading.