The CDMP is DAMA International's certification for data management professionals, and it has been around long enough that a lot of governance job ads still name it. The CDMP site says around 20,800 people hold Associate, about 1,560 hold Practitioner and roughly 430 hold Master, as of July 2026. Search interest for "cdmp certification" sits around 1,300 a month in the US, down about 38% on last year, while "dama cdmp" is up sharply from a small base. My read is that interest is shifting from people browsing to people who already know what they want. The structure is simple once you get past the website.
Everyone starts with the Data Management Fundamentals exam. Score 60% or more and you are an Associate. To reach Practitioner you need 70% or more on Fundamentals plus two specialist exams, and Master needs 80% or more on all three, plus an experience submission for Master. The specialist exams are Data Modelling and Design, Metadata, Data Quality, Data Governance, Data Warehousing and BI, Reference and Master Data Management, and Data Integration and Interoperability.
Each exam is 90 minutes, or 110 for non-native English speakers, and proctored online through Honorlock. DAMA lists each exam at USD $311, and a practice exam comes free with every purchase. The study material is the DAMA-DMBOK2 Revised Edition, and DAMA tells you to use that version. This is both the strength and the weakness. The DMBOK is a genuinely useful book if you have never seen data management laid out as a whole: governance, architecture, modelling, storage, security, integration, documents, reference and master data, warehousing, metadata and quality, all linked together.
People who come from engineering often have never been taught most of that. But the exam leans heavily on the book's own definitions and frameworks, so a lot of preparation is learning to answer the way DAMA thinks rather than the way your organisation works. The AI angle is the reason this is on an AI course review site at all. Almost every enterprise AI project I have seen stall did so on data problems, not model problems: nobody owns the data, nobody knows what a field means, quality is unmeasured, and lineage is a guess. That is CDMP territory.
The problem is that the exam does not say so. The current body of knowledge predates generative AI, and DAMA's DMBOK 3.0 project, which kicked off in June 2025 and explicitly targets AI governance and cloud platforms, has no confirmed publication date. Trade press points to 2027. DAMA also runs a GenAI for data management affinity group, but none of that is in the exam yet. On cost, Associate is affordable at one exam.
Practitioner or Master is three exams, so over $900 before renewal, and renewal is yearly with CPD and a fee. That is a lot for a credential that hiring managers outside governance-heavy sectors often do not recognise. Who should take it? Data governance leads, data stewards, data architects in large organisations, and consultants who sell governance work. If that is you, Associate is a sensible first step and Data Governance plus Metadata are the specialist exams I would pick, because they map most directly to the work AI governance teams are being asked to do. If you build pipelines or models for a living, read the governance and metadata chapters of the DMBOK and keep your money for a hands-on exam.