ACCA APM September 2026: The 25-Mark Data Analytics Question Is Now Guaranteed — Here's How to Answer It
From September 2026, one of your two APM Section B questions is guaranteed to come from syllabus section D: Data science and technology. That is 25 marks on data analytics, AI and information systems, every sitting, no exceptions.
What actually changed in APM for September 2026
The new APM Syllabus and Study Guide (September 2026 to June 2027) rewrites the Section B rules. Previously the examiner could pick Section B topics from anywhere. Now the guide states: "One of the Section B questions will come from syllabus section C, and one will come from syllabus section D." Section C is performance reporting. Section D is the new "Data science and technology for performance and insights."
Section D is no longer a side topic. It has grown to 14 learning outcomes and five of them are brand new: advising on the data science process (goal setting, selecting, cleaning, transforming and storing data), analysing data using regression analysis and identifying biases, patterns and distributions, assessing machine learning and AI for insights, assessing the output of data models and advising on refinements, and advising management on a model's output to form recommendations.
Some old favourites have gone to make room. BCG, Porter's Five Forces and McKinsey 7S have been removed. If your notes still have a Five Forces page, bin it. Total Shareholder Return and Net Promoter Score have been added to the financial and non-financial measure lists.
Why this is where marks get lost
The examining team's updated Examiner's Approach article is blunt about it: candidates "will be expected to analyse not merely calculate numerical data" and lists of rote-learned points "will not produce a complete answer." The same warning has appeared in every big data examiner's report for years. Candidates recite the 4Vs (volume, velocity, variety, veracity) and score two or three marks out of twenty. The examiner wants the data applied to the organisation's specific goals.
That matters more now because the question is compulsory. You cannot dodge it by hoping the technology question doesn't appear. It will.
Worked example: the regression output question
The scenario gives you a model built by the marketing team: Monthly sales = 200 + 4.2 × marketing spend ($000), with R² of 0.31, built on 18 months of data. The requirement asks you to assess the model and advise the board.
Wrong answer: "There is a positive relationship between marketing and sales. Each $1,000 of spend generates $4,200 of sales. The board should increase marketing spend to boost revenue."
Correct answer: "R² of 0.31 means marketing spend explains only 31% of sales movement; 69% is driven by something else, so this model is a weak basis for a spending decision. 18 monthly data points is a small sample and cannot show seasonality reliably. Correlation is not causation: the sales rises may reflect the two product launches mentioned in the scenario, which coincided with higher spend. The model should not be used to extrapolate beyond the range of spend actually tested. Recommendation: refine the model by adding product launch and seasonal variables before committing budget, and pilot any increase in one region first."
The wrong answer states what the numbers say. The correct answer challenges them (scepticism marks), quantifies the limitation (analysis marks) and advises action that is commercially safe (commercial acumen marks). Both took the same time to write.
What to do
Learn the four analytics types as questions, not definitions. Descriptive: what happened? Diagnostic: why? Predictive: what will happen? Prescriptive: what should we do? In the exam, name which type the organisation is currently using and which type its objectives actually need. That gap is usually the answer.
For any data model in a scenario, run the same five checks. Sample size, R² or fit, data quality and cleaning, correlation versus causation, and ethical or bias issues (D2i). Five checks, five paragraphs, each applied to the scenario facts.
Practise the new specimen exam, not old Section B questions. ACCA released a new specimen alongside the syllabus. Old past papers still work for Section A, but their Section B questions do not reflect the guaranteed C and D split.
The numbers
APM's pass rate has sat at 40 to 42% across the last five sittings (June 2025 to June 2026). It is one of the two hardest papers at Strategic Professional. Section B is 50 marks of your 100, and from this session half of that is a data question you can prepare for with certainty.
The examiner has told you exactly what 25 of your marks are about. Take them.