The four domains and what each weighs
Lesson 1 of 3 in The DEA-C01 Exam: Blueprint, Logistics, Strategy.
DEA-C01 validates one claim: that you can build data pipelines, and then keep them running, cheap and correct. Not that you can design a warehouse schema on a whiteboard in the abstract, and explicitly not that you can train a model or interpret a business result — both of those are named as out of scope for the person this exam is written for.
The shape. 65 questions, of which 50 affect your score; the other 15 are unscored pilot items AWS uses to calibrate future exams, and they are not identified while you sit. 130 minutes. Results come back as a scaled score of 100–1,000 with a minimum passing score of 720, and the model is compensatory — you pass the exam as a whole, never section by section. There is one version: DEA-C01. There is no C02.
Two consequences fall out before you study anything. Because you cannot tell which fifteen items are pilots, every question gets full effort — the one that feels unfairly obscure may simply never have counted. And because scoring is compensatory, a weak domain is survivable, so study time belongs where weight × weakness is largest rather than where your pride hurts most.
A horizontal bar chart of the four DEA-C01 content domains by percentage of scored content. Data Ingestion and Transformation is 34 percent, about 17 of the 50 scored questions, and is by far the largest. Data Store Management is 26 percent, about 13 questions. Data Operations and Support is 22 percent, about 11 questions. Data Security and Governance is 18 percent, about 9 questions. The four weights total 100 percent and the derived counts total 50. The first two domains together are 60 percent of the score.
Read that chart as a budget, not a table of contents. Domain 1 at 34% is more than a third of the paper, and the heaviest domain here by eight points — about seventeen scored questions on getting data in and shaped. Domain 2 adds 26%, so the first two domains together are 60% of your score: more than half this exam is "where does the data come from, and where does it live". Domain 3's 22% is the half of the job nobody budgets for — automation, querying, monitoring and data quality — and Domain 4's 18% is breadth at shallow depth, because it covers more separate ideas than any other domain in the smallest slice of the paper.
The best way to hold the blueprint is not as four lists but as one pipeline read left to right. Data arrives and gets shaped (Domain 1). It lands somewhere with a schema and a lifecycle (Domain 2). Somebody automates the line, queries it, watches it and checks it is not lying (Domain 3). And the whole thing has to be locked down and provable afterwards (Domain 4). Every question you will see is a station or a stretch of track on that one board.
| Domain | Weight | What you must be able to do | Your stations |
|---|---|---|---|
1 · Data Ingestion and Transformation | 34% · ≈17 questions | Read from streaming and batch sources and handle their throttling. Reshape and join data, convert formats, and diagnose a transformation that is slow rather than broken. Turn jobs into an orchestrated line with retries, alerts and idempotent steps. Read the SQL, Spark, template and deployment code that appears on screen. | Ingestion patterns — five modules |
2 · Data Store Management | 26% · ≈13 questions | Pick a store from the access pattern, not from the brand. Catalog schemas and partitions so engines and people can find them. Move data through storage tiers and expire it on time. Design a model — dimensional, key-value or open-table — and evolve its schema without breaking readers. | Choosing a data store — four modules |
3 · Data Operations and Support | 22% · ≈11 questions | Automate processing on schedules and events. Query and visualise data, and know what makes a query cheap. Monitor a pipeline, read its logs and find the cause rather than the symptom. Define, measure and enforce data quality, and know what to do once bad data has landed. | Automating data processing — four modules |
4 · Data Security and Governance | 18% · ≈9 questions | Authenticate pipelines and people. Authorize down to table, column and row. Encrypt at rest and in transit, and mask what must not be read. Produce logs an auditor accepts. Hold the line on privacy, residency, retention and data sharing. | Authentication and authorization — three modules |
Interactive sorting exercise: Place each scenario in the domain that owns it. This is the drill that makes the blueprint useful: in the exam you do it silently, before you read the options.
Key terms: DEA-C01, Target candidate (DEA-C01), Scaled score, Compensatory scoring, Unscored questions, Exam blueprint
Interactive checkpoint quiz (2 questions) — open this page in a browser to take it.