The QSDA exam (Qlik Sense Data Architect) now has 45 questions in 120 minutes, and you need 60% to pass. When we read the same Qlik page on 13 October 2026, it listed 50 questions, 90 minutes and a 62% pass mark, with a percentage weight for each topic. Those weights are gone, and the topic list has a new first section.
If you planned your QSDA study from an older guide, a forum post or a video, some of your numbers are now out of date. This entry sets the old and new format side by side, works out what the change means for your pacing, and shows how to split your study time now that Qlik no longer says how much each topic counts. It ends with a four-week plan, the mistakes that cost the most points, and five short answers. Every exam figure below was read on Qlik's own pages on 28 October 2026.
What changed on the QSDA exam?
Four things changed on the official exam page: the number of questions, the time, the pass mark and the way topics are shown. You can check every one of them on Qlik's Qlik Sense Data Architect exam details page, which is public and needs no sign-in.
Here is the old format, as we recorded it on 13 October 2026, against what the page shows on 28 October 2026:
- Questions: 50 before, 45 now.
- Time: 90 minutes before, 120 minutes now.
- Passing score: 62% before, 60% now.
- Topic weights: five percentages before (Identify Requirements 20%, Data Connectivity 8%, Data Model Design 28%, Data Transformations 38%, Validation 6%), none now.
- First topic: "Identify Requirements" has been replaced by "Introductory Concepts in Data Modeling".
Some things did not change. The page still describes an exam that measures your ability to identify requirements for data models, design and build data models, and validate the data. It is still platform-neutral, so the content applies to both client-managed Qlik Sense and Qlik Cloud. And it still carries the same warning: exam content is updated from time to time, the number and difficulty of questions may change, and the pass mark is adjusted to keep the standard the same.
One detail is easy to miss. The page still shows "Published: Sep 2, 2024" at the top, so the date does not tell you the content was updated. The only safe habit is to read the numbers themselves on the day you book.
What do 45 questions in 120 minutes mean for your pacing?
You now have about 2 minutes and 40 seconds per question, up from 1.8 minutes. That is roughly 48% more time for each question, which is our own arithmetic from Qlik's figures (120 divided by 45, against 90 divided by 50).
The pass mark works out the same way. If every question counts the same, 60% of 45 is 27 correct answers, so you can miss up to 18. Under the old format you needed 31 of 50. Qlik does not say that every question carries the same weight, so treat 27 as a planning number, not a promise.
More time per question usually means longer questions. A data architect exam lends itself to scenarios: a short description of source tables, a few lines of load script, or a data model viewer problem, followed by "what is the best fix?" Plan for reading time. A good rhythm is one careful pass through all 45 questions at about two minutes each, which leaves around half an hour to come back to the ones you flagged.
Which topics are on the QSDA exam now?
The exam page lists five topics and 15 distinct objectives. Here they are in Qlik's order, with the number of objectives under each:
- Introductory Concepts in Data Modeling (1 objective). Describe data modeling.
- Data Connectivity (2). Decide which data sources and connectors you need, and the right way to create the connections.
- Data Model Design (4). Pick the measures and attributes from each source, choose the right type of data model, optimize the model for Qlik Sense, and build data structures efficiently.
- Data Transformations (6). Build data content from requirements, handle null and blank data so filtering still works, document load scripts, handle dates, organize and clean scripts, and use variables to build incremental loads for the extract layer.
- Validation (2). Validate and test scripts, and validate and test data.
The page prints the "script organization and cleansing" objective twice under Data Transformations. We count it once, which gives six objectives there.
How should you split study time now that Qlik shows no weights?
Give Data Transformations the most time, Data Model Design the next most, and do not skip the small topics. With no published percentages, the best guides left are the objective list and the badge criteria.
Data Transformations holds 6 of the 15 objectives, which is 40% of the list. That is our count of objectives, not an official weight. It lines up with the old page, where Data Transformations was the largest topic at 38%.
The second guide is the Qlik Sense Data Architect badge on Credly. Its criteria spell out the techniques Qlik expects you to know: join, keep, resident load, where clauses, incremental load, purge strategy and variables. They also cover picking a model that favors reuse, scalability and efficiency, such as a tiered architecture. You should be able to check a model against its source data with visualizations, and fix synthetic keys, circular references and script errors. The badge is listed as an Advanced, paid certification.
A practical split for most candidates:
- About 40%: Data Transformations. Write scripts, break them and fix them.
- About 30%: Data Model Design. Star schemas, link tables, key fields and performance.
- About 30% together: Data Connectivity, Validation and the introductory concepts. These have fewer objectives, but each one is quick to learn, so they are cheap points.
What experience does Qlik expect before you sit the QSDA?
Qlik expects hands-on experience building several production-quality apps in Qlik Sense, not just course hours. The recommended experience on the exam page is:
- Practical experience developing multiple production-quality applications in Qlik Sense.
- Writing Qlik Sense load scripts and validating data.
- A basic understanding of extract, transform and load (ETL).
- Creating and using connectors to different data sources.
- Understanding the QVD layer and the architecture of the Qlik platform.
- Architecting data for good performance.
- Familiarity with SQL and relational databases.
For preparation, the page now points to two Qlik Learning paths: "Data Modeling with Qlik Cloud Analytics / Qlik Sense" and "Advanced Data Modeling with Qlik Cloud Analytics / Qlik Sense". Both open only after you sign in with a Qlik account, so search for them by name once you are logged in.
What does a QSDA-style question look like?
Expect a short scenario with a script or a data model, and four possible fixes. Here is one we wrote ourselves in that style. It is not a real exam question.
Scenario. A load script brings in two tables. Orders has OrderID, CustomerID and OrderDate. Shipments has OrderID, CustomerID and ShipDate. After the reload, the data model viewer shows a synthetic key, $Syn 1, between the two tables. Each shipment belongs to exactly one order. What is the best fix?
- A. Leave it, because synthetic keys are always the correct way to link tables.
- B. Load CustomerID only in Orders, so that OrderID alone links the two tables.
- C. Add DISTINCT to the Shipments load.
- D. Store both tables as QVD files and reload them.
Answer: B. Qlik's help page on how Qlik Sense creates synthetic keys says they appear when two or more tables share two or more fields. Most of the time they point to an error in the script. The advice is to link tables only on the fields that logically connect them and to avoid redundant fields. Since a shipment's customer is already known through its order, CustomerID in Shipments is redundant. If two fields truly form the key, you build your own single key with string concatenation inside AutoNumber.
The short below shows one more sample question from AnalyticsExam. Its YouTube description was written before the change and still lists the older 50-question, 90-minute, 62% format, so use the clip for the question, not for the numbers.
What is a good four-week QSDA study plan for the new format?
Four weeks of about six to eight hours a week works for someone who already builds Qlik Sense apps. If you have never written a load script, add two weeks at the start for basic scripting.
- Week 1: Connectivity and model design. Connect to at least three kinds of source (a database, a flat file and a QVD). Build one star schema and one model with a link table, and note which you would choose for which business question. Read about circular references, where the load finishes but the data model viewer shows red dotted lines, and fix one by renaming a field.
- Week 2: Transformations, part one. Practise joins, keep, resident loads, mapping loads and date handling. Then work through NULL handling. Real NULLs come from ODBC sources, joins and forced concatenation, and they cannot be used for associations or selections unless you use NullAsValue. Text files cannot contain NULLs at all.
- Week 3: Transformations, part two. Build all three cases from Qlik's guide to incremental load with QVD files: insert only, insert and update, and insert, update and delete. Note that WHERE NOT Exists(PrimaryKey) forces the QVD read into standard mode, and that the delete case needs an inner join on the source keys. Comment and organize your scripts as you go, because documenting scripts is an objective too.
- Week 4: Validation and timing. Check every model against its source with simple tables and counts. Then sit one timed set of 45 questions in 120 minutes and review every miss by objective.
Which mistakes cost the most points on the QSDA?
Most lost points come from out-of-date numbers, untidy keys and skipped edge cases, not from hard theory.
- Studying from old figures. Many guides, videos and practice sets still say 50 questions, 90 minutes and 62%. Plan your timing on 45, 120 and 60%.
- Treating every shared field as a key. Fields like Comment or Description, or a Date that means order date in one table and ship date in another, should not link tables. Qlik's help uses exactly these examples.
- Forgetting how NULL moves through expressions. A plus NULL returns NULL, but A and NULL joined with the & operator returns A. Small rules like this decide filter and measure questions.
- Learning one platform only. The exam is platform-neutral. If you only use Qlik Cloud, spend an evening in client-managed Qlik Sense, and the other way round.
- Skipping Validation because it is small. It has only two objectives, but each one is a near-certain question type, and the badge criteria name validation against source data directly.
How does the QSDA compare with Qlik's other Sense exams?
The QSDA is now the odd one out: the business analyst and system administrator exams still show 50 questions, 90 minutes and 62% on their pages today. If you are choosing an order, the hands-on QSBA topics we mapped earlier are a natural first step, because they cover apps and visualizations before you go deep into scripting.
All three exams warn that their content can change. Check the page for your exam on the day you book, and again a week before your exam date.
If your job is running servers rather than building models, the seven admin skills behind the QSSA are the better next read.
Frequently asked questions
1. How many questions are on the QSDA exam now?
The Qlik Sense Data Architect exam has 45 questions, according to Qlik's exam page on 28 October 2026. On 13 October 2026 the same page listed 50 questions.
2. How long is the QSDA exam?
You get 120 minutes, up from 90. That is about 2 minutes and 40 seconds per question, by our own arithmetic.
3. What is the QSDA passing score?
The passing score is 60%, down from 62%. If every question counts the same, that is 27 correct answers out of 45.
4. Does Qlik still publish topic weights for the QSDA?
No. The page lists five topics and their objectives with no percentages. Data Transformations has the most objectives, 6 of 15, so give it the most study time.
5. Is the QSDA exam about Qlik Cloud or client-managed Qlik Sense?
Both. Qlik describes the exam as platform-neutral, so the content applies to client-managed Qlik Sense and to Qlik Cloud.