C E R T I F I C A T I O N G U I D E Qlik • QSDA • Platform-neutral QSDA Certification Guide: Qlik Sense Data Architect Syllabus, sample questions and a study plan for the Qlik Sense Data Architect exam Inside: the exam fact sheet checked against Qlik’s own page, all five domains broken out to every published objective, why two of them decide two thirds of your result, what a data architect is actually paid and where the credential moves the needle, a five-step route from load script to exam day, and ten sample questions with answers. T H E B L U E P R I N T Identify requirements 20% Data connectivity 8% Data model design 28% Data transformations 38% Validation 6% E X A M A T A G L A N C E 50 QUESTIONS 90 MINUTES 62% PASS SCORE 5 DOMAINS 19 OBJECTIVES Prepared by AnalyticsExam • www.analyticsexam.com www.analyticsexam.com Qlik • QSDA Qlik Sense Data Architect Certification 1 CERTIFICATION GUIDE Contents Ctrl+click any line to jump straight to that section. The same seven entries appear in your PDF reader’s bookmark pane. SECTION 01 Exam Overview 2 SECTION 02 Two Domains Decide Two Thirds 3 SECTION 03 The Five Domains, Objective by Objective 3 SECTION 04 What the Credential Is Worth 5 SECTION 05 Getting Certified, Step by Step 6 SECTION 06 QSDA Sample Questions 7 SECTION 07 Where to Go Next 10 www.analyticsexam.com Qlik • QSDA Qlik Sense Data Architect Certification 2 SECTION 01 Exam Overview The Qlik Sense Data Architect (QSDA) certification exam is Qlik’s back-end credential. Where the Business Analyst exam asks what a chart should show, this one asks how the data underneath it was loaded, keyed, transformed and proved correct. It is deliberately platform-neutral, so the same paper serves a client-managed Qlik Sense estate and a Qlik Cloud tenant. Every figure below is taken from the official Qlik exam details page and cross-checked against the QSDA exam syllabus. Certification Qlik Sense Data Architect Exam code QSDA Exam name Qlik Sense Data Architect Certification Exam Scope Platform-neutral — applies to both client-managed Qlik Sense and Qlik Cloud Number of questions 50 Duration 90 minutes Passing score 62% Exam fee Not stated on Qlik’s exam page; a 2021 Qlik announcement priced it at USD 250 — confirm before booking Delivery Local test centre or online proctoring; registration through Pearson VUE Badge Qlik Sense Data Architect certification badge, issued through Credly Domains Five, with published percentage weightings Two numbers there are worth sitting with. Fifty questions in 90 minutes is a shade under two minutes each, and QSDA items are long: candidates comparing notes in the Qlik Community certification threads describe stems that run to several sentences with load-script lines as the answer options. Reading speed is a real constraint on this paper, not a footnote. And 62% is a comparatively forgiving pass mark — you can lose nineteen questions and still pass — but Qlik states that "exam content is updated periodically" and that "the passing score is adjusted to maintain a consistent standard", so treat it as the current figure rather than a permanent one. What Qlik recommends before you book Production-quality Qlik Sense app development, not just tutorial work. Writing load scripts and validating the data they produce. A basic grasp of ETL, plus creating and using connectors. The QVD layer, Qlik platform architecture, SQL and relational databases. www.analyticsexam.com Qlik • QSDA Qlik Sense Data Architect Certification 3 Qlik labels those as recommended experience rather than prerequisites, and it publishes no formal entry requirement, no version code and no stated validity period for this exam. Nothing stops you booking it cold. The recommendation is still the honest signal: the seven lines above describe a working data architect, and the paper is written for someone who has been one. SECTION 02 Two Domains Decide Two Thirds Most blueprints spread their weight fairly evenly and leave you to guess where to start. This one does not. Add the two script-and-model domains together and the arithmetic is blunt: The weighting, read as a study plan Data Transformations 38% + Data Model Design 28% = 66% of the paper. That is roughly 33 of the 50 questions, against a pass mark of 31. Data Connectivity and Validation together are 14% — about seven questions. Identify Requirements is 20%, and almost none of it is technical. The consequence is uncomfortable but useful: you could answer every requirements, connectivity and validation question correctly and still fail , because those three domains come to 34% and the pass mark is 62%. The script and the data model are not the bulk of this exam, they are the exam. Everything else is the margin. Read the other way, the small domains are the cheapest marks on the paper. Validation is 6% — about three questions — and it has just two published objectives: validate and test scripts, validate and test data. Connectivity is 8% across two objectives. An evening spent on those four objectives covers a seventh of the exam, and they are the kind of thing a working architect does by instinct without ever having named it. Sample question 5 in this guide is exactly that shape: check the key field for duplicates and nulls before you trust it. Identify Requirements deserves a separate warning. At 20% it is the second-heaviest domain, and it is the one technical candidates skip, because it reads as soft. It is not soft — it is where slowly changing dimensions, granularity, aggregation level and security scoping are tested, and those are design decisions with a single defensible answer. Two of the ten sample questions later in this guide sit in this domain, and neither can be reasoned out from load-script knowledge alone. SECTION 03 The Five Domains, Objective by Objective Qlik publishes an exact percentage for each domain rather than a range, so the chart below is the real blueprint and not a proxy for it. The objective lists beneath are reproduced as Qlik words them. www.analyticsexam.com Qlik • QSDA Qlik Sense Data Architect Certification 4 Identify requirements 20% Data connectivity 8% Data model design 28% Data transformations 38% Validation 6% Published weightings from the Qlik Sense Data Architect exam details page. Notice the shape rather than the individual bars. Data Transformations is more than six times the size of Validation, and the two smallest domains together are smaller than Identify Requirements on its own. There is also an asymmetry worth planning around: Data Model Design carries 28% across four objectives, while Identify Requirements carries 20% across five. Model design therefore goes deeper per objective — expect several questions on the same idea approached from different angles, which is why synthetic keys, AutoNumber and NoConcatenate all appear in the ten sample questions below. 1 Identify Requirements — 20% of the exam, 5 objectives • Determine primary requirements for business users. • Identify stakeholders using a given scenario. • Determine metrics and levels of granularity and aggregation. • Determine dimensionality and the need for slowly changing dimensionality support. • Determine the appropriate level of security. 2 Data Connectivity — 8% of the exam, 2 objectives • Determine the data sources and connectors needed. • Determine the appropriate method to create connections for data sources. 3 Data Model Design — 28% of the exam, 4 objectives • Determine the measures and attributes from each data source. • Identify the appropriate type of data model. • Determine the correct method to optimize the data model for Qlik Sense. • Determine the correct method to implement data structures efficiently. 4 Data Transformations — 38% of the exam, 6 objectives • Determine the correct method to build data content based on requirements. • Analyze and evaluate null and blank data handling required to support filtering. • Determine the correct method to document Data Load scripts. • Determine the correct method for date handling techniques. • Determine the correct method to perform script organization and cleansing. www.analyticsexam.com Qlik • QSDA Qlik Sense Data Architect Certification 5 4 Data Transformations — 38% of the exam, 6 objectives • Analyze relevant variables to build scripts for incremental loading for the extract layer. 5 Validation — 6% of the exam, 2 objectives • Determine the appropriate method to validate and test scripts. • Determine the appropriate method to validate and test data. Two features of that list shape how you revise it. First, almost every objective begins "Determine the correct method to..." or "Determine the appropriate...", which is blueprint language for you will be given a scenario and asked to choose . That is why the sample questions are all named companies with a stated problem rather than definitions — the paper tests judgement about which technique fits, not whether you can recite what the technique is. Second, the objectives are written in product-neutral analytics vocabulary — granularity, dimensionality, incremental loading — but the answers are Qlik-specific. Knowing what an incremental load is will not tell you that WHERE NOT EXISTS against a QVD is how Qlik does it. It is worth reading the objective list next to the AnalyticsExam QSDA syllabus, because AnalyticsExam expands the same five domains with the sub-topics candidates report seeing, and the two lists agree domain for domain and percentage for percentage. SECTION 04 What the Credential Is Worth The role this exam certifies is well paid, and unusually well documented. Coursera’s data architect salary round-up, updated 28 February 2026, gathers five independent sources and they disagree in an informative way: the US Bureau of Labor Statistics puts the median data architect salary at USD 135,980 , Payscale at USD 135,121 and Salary.com at USD 127,032 , while Glassdoor’s median total pay is USD 178,000 and Zippia’s average is the outlier low at USD 113,700 Read that spread honestly. The gap between Zippia’s 113,700 and Glassdoor’s 178,000 is not measurement noise — it is the difference between base salary and total compensation, and between a generalist and someone owning the data layer of a funded platform. Glassdoor’s own experience curve makes the same point from inside a single dataset: USD 132,000 at 0–1 year rising to USD 209,000 at 15 or more . What moves you along that curve is scope, not certificates. What a certificate does is make you legible for the scope in the first place. Be equally honest about growth. The BLS projects data architect roles growing 4% from 2024 to 2034 , an increase of about 7,800 positions — in line with the national average, not a boom. This is a mature, stable, well-paid specialism rather than a rising one, and that is an argument for depth over breadth: the return here comes from being demonstrably good at a specific platform, which is precisely what a platform certification evidences. www.analyticsexam.com Qlik • QSDA Qlik Sense Data Architect Certification 6 Where QSDA actually helps It certifies the back end — integration, modelling, optimisation, not dashboards. It is platform-neutral, so it survives a move from client-managed Qlik Sense to Qlik Cloud. It maps to data architect, ETL developer and data integration specialist roles. The badge is issued through Credly, so it is verifiable by a recruiter without your CV. If you are choosing between Qlik’s two Sense credentials, AnalyticsExam’s side-by-side comparison draws the line where it actually falls: the Business Analyst exam centres on "data visualization and creating interactive dashboards for end users" and leads toward business analyst, BI developer and dashboard designer roles, while this one "concentrates on back-end processes like data integration and model optimization" and leads toward "Data Architect, ETL Developer, and Data Integration Specialist". Pick by which half of the work you want more of, not by which exam looks harder. SECTION 05 Getting Certified, Step by Step 1 Get a Qlik Sense environment you can break A trial tenant or employer estate. Every heavy objective is something you do, not read. 2 Build a QVD layer end to end, once Extract, store, reload from QVD. It is the spine of the 38% transformations domain. 3 Break your own data model deliberately Force a synthetic key, then fix it three ways. Model design is 28% and rewards scars. 4 Sweep connectivity and validation in one evening Four objectives, 14% of the paper, and the cheapest marks available to you. 5 Book it, and read the stems slowly 50 questions, 90 minutes, 62% to pass. Time pressure here is real. Step three is the one candidates skip and regret. A synthetic key you have read about is a definition; a synthetic key you caused at 6pm on a Friday is a reflex. Force one, then resolve it by composite key with AutoNumber, by renaming a field, and by commenting the field out — that single exercise covers most of what sample question 7 below is testing. When you want to check the recognition has landed under time pressure, AnalyticsExam’s QSDA practice exam is built to this same syllabus, and candidates comparing preparation notes and exam-day experience gather in the Qlik Community. On booking: in its own exam announcement Qlik states that the exam "can be taken in a local test center or via online proctoring", with registration handled through Pearson VUE, and prices it at USD 250. Check the current fee and scheduling options on the official page before you pay — Qlik links out to its own purchase system, and the details page carries the standing caveat that exam content is updated periodically. www.analyticsexam.com Qlik • QSDA Qlik Sense Data Architect Certification 7 SECTION 06 QSDA Sample Questions These come from AnalyticsExam’s QSDA sample questions page, reproduced verbatim and shown here in shuffled order. Eight are single-answer and two ask for two responses, and both of the multiple-select items say so in the stem — "which two actions", "which two statements" — which is worth noticing, because on the real paper that phrasing is your only warning. Notice how many are a named company with a specific failure rather than a definition; that scenario framing is characteristic of this exam. Answers follow the set. Q1. Finance wants a sales report that shows each sales rep's commission percentage as it was at the time each historical sale occurred, even though a rep's commission rate has changed several times since. What kind of requirement is this? a) A granularity requirement, since it only concerns how finely the fact table's rows are broken down. b) A slowly changing dimension requirement, since a historical attribute value must be preserved as of the transaction date. c) A simple overwrite requirement, since only the sales rep's single current commission rate needs to be stored going forward. d) A connectivity requirement, since it depends on which connector is used to load the sales rep table itself. Q2. Compared with only reviewing an existing static report to infer requirements, why is interviewing the actual business users generally more reliable when scoping a new data model? a) Interviews are generally faster to complete than reading through a single existing report line by line. b) Interviews remove the need to profile the source data before the model is designed. c) Interviews let the data architect skip validating the finished data model with users at a later stage. d) Interviews surface the real decisions users need answered, not just an existing report's layout. Q3. Emberwood Foods loads a RegionSales table, then later in the script loads a StoreSales table that happens to have identical field names and count to RegionSales. The data architect wants StoreSales to remain a separate table rather than being appended onto RegionSales. What should be added to the StoreSales load to prevent the automatic append? a) The QUALIFY statement, applied before the StoreSales load b) The NOCONCATENATE prefix c) A CROSSTABLE prefix on the StoreSales load d) A RESIDENT clause referencing RegionSales www.analyticsexam.com Qlik • QSDA Qlik Sense Data Architect Certification 8 Q4. A data architect builds a composite key by concatenating three long text fields together, producing a key value that is dozens of characters wide. Reload times and memory usage both increase noticeably once this key is added. Which technique addresses this without changing what the key represents? a) Split the composite key back into its three original text fields and load each one separately again, undoing the earlier consolidation into a single key. b) Wrap the composite key in ApplyMap() so its value is looked up from a mapping table rather than stored directly on every row. c) Wrap the key in AutoNumber(), replacing its long text values with compact sequential integers. d) Apply NoConcatenate to the load that builds the composite key. Q5. A data architect validating a newly built Customers table wants to confirm that the field intended as the primary key is actually unique and complete before relying on it for associations. Which validation step directly addresses this? a) Check that the Customers table's tab name in the script matches the table name shown in the data model viewer. b) Check that the Customers table was the first table loaded in the script sequence. c) Check the key field for duplicate values and for nulls or blanks that would leave rows without a usable key. d) Check that the Customers table sorts correctly when a user clicks the column header in a table chart. Q6. Meridian Sports has a Products table where one field stores a comma-separated list of tags for each product, such as 'outdoor,footwear,clearance'. The data architect wants one output row per product-tag pair. Which function should be used on the tags field, without specifying a field-index argument? a) ApplyMap b) IntervalMatch c) Crosstable d) Subfield www.analyticsexam.com Qlik • QSDA Qlik Sense Data Architect Certification 9 Q7. A data architect at Bluepeak Retail reloads a data model and notices an unintended synthetic key has appeared between the Orders and Returns tables, formed from two shared fields that were not meant to be a joint key. Which two actions would resolve this without discarding needed data? CHOOSE TWO a) Build a deliberate composite key by concatenating the shared fields, often wrapped in AutoNumber(), and drop the original fields. b) Apply a WHERE NOT EXISTS clause when concatenating the Returns load against the existing QVD so only new or changed rows are appended. c) Delete the Returns table from the data model entirely so no association can form. d) Reload the tables in reverse order so the Orders table associates first. e) Rename or comment out the accidental shared field in one of the two tables so only the intended key remains. Q8. Basalt Ridge Mining's data architect applies a WHERE clause to a source load that discards roughly a third of the incoming rows, then examines both RecNo() and RowNo() on the rows that remain. Which two statements correctly describe the result? CHOOSE TWO a) RowNo() values on the surviving rows form an unbroken sequence starting at 1, because it only counts rows that actually make it into the load. b) Applying a WHERE clause resets RecNo() back to 1 for the first surviving row, the same way it does for RowNo(), since any row-counting function restarts fresh once filtering removes rows. c) RecNo() and RowNo() always return identical values, because both simply count rows in the order they are loaded. d) Neither RecNo() nor RowNo() can be used together with a WHERE clause in the same LOAD statement. e) RecNo() values on the surviving rows can have gaps, because it reflects each row's position in the original, unfiltered source. Q9. After reloading a data model, a data architect notices that selecting a specific product on one sheet leaves several customers greyed out who should logically still have purchased that product. What should the data architect check to validate whether the associations are behaving correctly? a) Whether the key fields joining the affected tables are spelled, cased, and typed identically. b) Whether the connection used to load the affected tables has since been renamed by an administrator. c) Whether the app's dashboard sheets have been published to the correct stream or space. d) Whether the fact table has been pre-aggregated to a coarser grain than the original source data. www.analyticsexam.com Qlik • QSDA Qlik Sense Data Architect Certification 10 Q10. The merchandising director at Harrow Retail Group asks a data architect to build a new app so the team can “see how products are doing,” without further detail. What should the data architect do first to determine the primary requirement for the business users? a) Begin building a star schema immediately, using every field found in the product and sales source tables without further discussion. b) Meet with the director and the report's actual users to confirm which decisions the app must support. c) Reuse the data model from an unrelated finance app, since it is already built, tested, and running in production. d) Ask the IT department to decide which specific metrics the merchandising team should track for the new app. Answer key Q1 Q2 Q3 Q4 Q5 Q6 Q7 Q8 Q9 Q10 b d b c c d a, e a, e a b SECTION 07 Where to Go Next One caution about study material. A great deal of data-modelling content online is written for dimensional warehousing in SQL, and the concepts transfer while the answers do not. Qlik’s associative engine, the QVD layer, synthetic keys, AutoNumber, NoConcatenate and Subfield are Qlik constructs, and fluency in star schemas will not produce the right answer about any of them. Work from the Qlik exam details page for the blueprint and the fee, keep the wider Qlik catalogue in view on the AnalyticsExam Qlik certification list, and read the exam-day threads on Qlik’s own forum before you book — they are the only place candidates describe what the questions actually feel like. Quick reference Official exam details, objectives and purchase - learning.qlik.com Syllabus, sample questions and practice test - analyticsexam.com/qlik-certification Candidate discussion and exam-day reports - community.qlik.com Booking and proctoring - Pearson VUE www.analyticsexam.com Qlik • QSDA Qlik Sense Data Architect Certification 11 Before you book — the guide in one card Questions 50 Duration 90 minutes Passing score 62% Exam fee Not published by Qlik Delivery Test centre or online proctoring Booking Pearson VUE Scope Client-managed and Qlik Cloud Badge Issued through Credly Heaviest domain Data transformations (38%) Lightest Validation (6%) Objectives 19 across five domains Where the exam is won Transformations + model design = 66% Good luck. The useful thing about this paper is that Qlik tells you exactly where the weight sits, and the weight sits on the load script and the data model. If you can build a QVD layer, spot a synthetic key before it spots you, and say why you chose one granularity over another, you are most of the way there already.