The Marketing Cloud Next Consultant exam has been bookable since 21 August 2026, and the prep market has barely caught up. Most of what exists is either the official Trailhead trail — good, but not shaped like the exam — or generic Marketing Cloud content written for the old Email Specialist and Consultant tracks, which is a different product.
So here are eight real practice questions. These are the same eight that make up the free diagnostic sample on CertSprint, so you can work through them here and then go see how you'd hold up under a timer.
What the exam actually looks like
| Scored questions | 60 (plus up to 5 unscored) |
| Time | 105 minutes |
| Passing score | 72% — 43 of 60 |
| Prerequisites | None |
Two format notes that matter more than they sound.
Three options, not four. Salesforce's exam guide uses the boilerplate "multiple-choice/multiple-select" wording, but candidates who have sat the exam since registration opened consistently report three answer choices per question and no multi-select items. That changes the maths: a blind guess is worth 33%, not 25%, and the gap between guessing and the 72% bar is what you actually have to close. The questions below are written in the three-option format for that reason.
The pass mark is high for a consultant exam. 72% leaves you seventeen wrong answers across the whole paper. There is no section you can write off.
Where the questions come from
The exam has six sections, and they are not evenly weighted:
| Section | Weight |
|---|---|
| Campaign Design & Flow | 30% |
| Data & Identity | 25% |
| Setup & Governance | 13% |
| Consent | 13% |
| Agentforce & AI | 11% |
| Analytics | 8% |
The eight questions below are drawn from the three heaviest-scoring areas — four from Campaign Design & Flow, three from Data & Identity, one from Consent. That is deliberate: it is the 68% of the exam where most candidates either pass or fail. It also means this sample tells you nothing about how you'd do on Setup & Governance, Agentforce & AI, or Analytics. Treat a good score here as partial information, not a verdict.
Question 1 — Campaign Design & Flow
A weekly product digest should go to everyone in the "Newsletter Subscribers" segment each Friday. Which trigger type fits this requirement?
A. A segment-triggered flow firing on membership changes B. A scheduled campaign sending to the segment on its Friday schedule C. A segment-triggered flow with a weekly re-entry rule that resends each Friday to members who still qualify
Answer: B. Recurring sends to the full current audience are exactly what scheduled campaigns are for — segment-triggered entry serves the opposite pattern (reacting to individuals newly qualifying). Option C is the trap: it describes a build that would technically produce Friday sends, which is why it feels defensible, but it re-implements a scheduled campaign inside an orchestration that then has to carry re-entry rules forever.
Question 2 — Campaign Design & Flow
A consultant is asked whether to build a promotion as a single-send campaign or a multi-step flow. Which requirement most clearly calls for the multi-step flow?
A. The promotion must reach several thousand people across more than one market B. Behavior over time should change what each person receives next C. The audience is defined by a segment
Answer: B. Templates, segment audiences, and tracking apply to single sends too — branching on subsequent behavior (opens, clicks, purchases) is what requires orchestration over time. Volume and multi-market reach are not architectural reasons for a flow.
Question 3 — Campaign Design & Flow
Which requirement indicates a transactional/API-triggered send rather than a marketing campaign flow?
A. A weekly curated newsletter B. A password reset email requested by the user C. A quarterly win-back offer for lapsed customers
Answer: B. Transactional sends are user-initiated, functional, and expected — password resets, order confirmations — typically fired via API at the moment of the triggering action, and delivered regardless of marketing consent. Newsletters and win-back offers are marketing communications that belong in campaign flows under consent rules. This distinction shows up on the exam in consent questions too, so it is worth being precise about.
Question 4 — Campaign Design & Flow
A multi-step campaign flow should wait until a contact opens the first email before sending a follow-up, but must not wait forever. What is the recommended design?
A. A fixed seven-day wait followed by the send, so everyone receives the follow-up at the same interval B. A wait step with an engagement condition and a maximum duration, followed by a decision split on the outcome C. A decision split immediately after send, with no wait
Answer: B. A bounded wait with an engagement condition captures openers as they engage, while the timeout path ensures non-openers continue through the flow instead of stalling. Option C fails because engagement data has not arrived yet at the moment of the split — a mistake that is easy to make in a build and easy to miss on a screen.
Question 5 — Data & Identity
The client complains a segment count is far lower than the raw record count in the source system. Which explanation should the consultant investigate FIRST?
A. The platform randomly samples segments B. Identity resolution has unified duplicate records into fewer distinct profiles, and the segment counts profiles rather than records C. Activation targets only profiles with a mapped contact point, so the segment count excludes the rest of the audience
Answer: B. Segments count unified individuals, not raw source rows — successful unification of duplicates is the most common and healthy reason counts drop versus source records. Option C describes something real that also reduces reachable audience, which is what makes it a good distractor; it just is not the first thing to check, and it explains activation counts rather than segment counts.
Question 6 — Data & Identity
Which problem will identity resolution NOT fix on its own?
A. Two records for the same person sharing the same email address B. Incorrect attribute values entered in a source system C. The same person appearing in two source systems with matching identifiers
Answer: B. Identity resolution links records that share identifiers — the same email across records or systems is exactly what it unifies. What it cannot do is repair bad data: a wrong phone number or misspelled name entered at the source stays wrong on the unified profile. Reconciliation rules only choose which value wins, not whether it is true. Data quality is fixed in the pipeline, not by matching.
Question 7 — Data & Identity
What is the practical difference between a standard segment on a publish schedule and a real-time segment?
A. There is no difference except naming B. Real-time segments update membership within moments of the qualifying behavior, while scheduled segments only refresh at their configured publish interval C. Real-time segments evaluate only streaming events, while scheduled segments are the only ones that can read profile attributes and calculated insights
Answer: B. The choice is a latency decision: cart-abandonment style triggers need near-immediate membership changes; weekly newsletters tolerate scheduled refreshes. Option C invents a capability boundary that does not exist — plausible if you half-remember the streaming-versus-batch distinction, wrong if you know it.
Question 8 — Consent
A customer withdraws marketing consent today but an already-scheduled campaign is due to send tomorrow. What should a correctly designed setup do?
A. Send anyway because the schedule predates the withdrawal B. Exclude the person at send time because consent is evaluated when the message goes out C. Send the message and record the withdrawal as effective from the following campaign onward
Answer: B. Consent is enforced at send evaluation, not at scheduling time — the withdrawn individual is suppressed while the campaign proceeds for everyone else. Consent carries 13% of the exam, and most of its questions test this same instinct: consent state is checked at the latest possible moment, and suppression is surgical rather than campaign-wide.
How to read your score
Eight questions is a sample, not an assessment. Two honest calibrations:
- Got 7 or 8 right? You have solid instincts in the two heaviest sections. You still know nothing about your Setup & Governance, Agentforce & AI, or Analytics performance — together 32% of the exam, which is more than enough to fail on.
- Got 5 or fewer? The pattern to look at is which wrong answers attracted you. In three of these questions the distractor describes something that genuinely exists in the platform and would even work; it is just not the right design. That is the exam's favourite trick, and it is not something extra reading fixes — it is fixed by doing enough scenario questions that the second-best answer starts to feel wrong.
The real exam gives you 105 minutes for 60 questions — about 105 seconds each, and the first candidates through have reported using nearly all of it. Reading speed under pressure is part of what is being tested.
What these eight left out
The timer, the 72% pass line, and the three sections this sample deliberately left locked: Setup & Governance, Agentforce & AI, and Analytics. A timed run on CertSprint covers all six sections at the real exam length and scores each one separately, so you find out where you stand on the 32% you have not seen yet.
Where to go next
For the full picture of the exam these eight are drawn from — the final section weights, the three-option format, booking details, and how to weight your study — read our complete 2026 guide to the Marketing Cloud Next Consultant certification.