Skip to content
Tech Interview Prep

Top 100 Product Manager Practice MCQs

Every question on one page: open one, choose an answer, and see why at once. Signed in, your latest answer to each question counts, and you can answer again after 24 hours.

Curated: · Written: · Reviewed:

Reviewed 100Review pending 0

Answers to reviewed questions count toward this role and the overall leaderboard; your latest answer to each one is the one that counts.

100 questions

MCQ-52Two teams publish a weekly active accounts metric for the same product from the same events table. For the week of March 3, Team A reports 412,000 and Team B reports 334,000. Their definitions differ in three clauses. Population: Team A counts every account with at least one authenticated request; Team B excludes accounts whose only requests in that period went to the public API. Event: Team A counts any request; Team B requires a session start. Window: Team A uses a rolling 7 days ending Sunday 23:59; Team B uses the calendar week Monday 00:00 to Sunday 23:59. Team A's query returns 61,000 accounts whose only requests in that period went to the public API. Which conclusion about the 78,000 gap is correct?
MCQ-33A B2B SaaS team set a weekly target on demos booked as its input metric. Baseline: 200 demos, 20% close rate, $20,000 average deal, $800k revenue. Latest quarter: 320 demos, $830k revenue. The 120 added demos came from a new self-serve segment that closes at 5% on a $5,000 average deal; the original 200 enterprise demos closed at the same 20% and $20,000 as before. Which conclusion should the team act on?
MCQ-53A B2B SaaS growth team runs a Monday review whose job is to catch a metric moving before it compounds. The page carries six tiles. Last quarter the weekly active rate fell from 41% to 35% and nobody raised it, while the cumulative signups tile, which rose every week, sat at the top. The team's rule is that a tile stays only if some value of it would change what they do that week; the growth team's levers are onboarding and support. Trial-to-paid rate is owned by the pricing team and acted on in their Friday meeting. Which set of tiles should the Monday review carry?
MCQ-55A B2B analytics product defines activation as 'completed profile setup'. Over two quarters the team raised it from 51% to 78% of new accounts with in-app prompts; month-3 retention stayed flat at 31%. They measure first-week behaviours against month-3 retention: | First-week behaviour | Retained if done | Retained if not | Share of new accounts doing it | | --- | --- | --- | --- | | completed profile setup | 34% | 31% | 78% | | invited a teammate | 62% | 24% | 41% | | ran two imports | 58% | 22% | 12% | | connected a data source | 55% | 20% | 6% | The team can only run onboarding prompts on behaviours that at least 30% of new accounts already perform in week one; below that, there is no audience to test against. Among the behaviours that clear that bar, they will pick the one with the largest retention gap. Which behaviour should the team adopt as its activation metric?
MCQ-70A subscription product's six-month renewal is the outcome the team steers on, and renewal itself is already tracked on the exec dashboard. A mobile redesign shipped three quarters ago. Correlations between each candidate first-week indicator and six-month renewal, measured on accounts that have aged to six months: | Indicator | Pre-redesign cohorts (n = 4,200) | Post-redesign cohorts (n = 2,600) | | --- | --- | --- | | first-week logins | 0.61 | 0.14 | | second-week import | 0.40 | 0.58 | | teammate invited | 0.49 | 0.52 | The PM must add one leading indicator to the dashboard for the next two quarters. Which indicator should it be?
MCQ-56Signup conversion (visit to completed account) fell from 3.2 percent to 2.0 percent quarter over quarter, and a rebuilt signup form is ready to ship. The same two quarters broken out by acquisition source: | Source | Share last quarter | Share this quarter | Conversion last quarter | Conversion this quarter | | --- | --- | --- | --- | --- | | organic | 62% | 34% | 3.1% | 3.1% | | paid social | 8% | 48% | 0.4% | 0.4% | | referral | 30% | 18% | 4.0% | 4.0% | Which conclusion is justified?
MCQ-54A subscription app ships a checkout redesign in Q2. The dashboard reports blended month-2 retention of 38% for Q1, Q2, and Q3, so the redesign is called neutral. The team splits the last six monthly acquisition cohorts by channel: | Cohort | Month 2 | Month 6 | | --- | --- | --- | | Paid search, pre-redesign | 44% | 39% | | Paid search, post-redesign | 33% | 11% | | Organic, pre-redesign | 52% | 48% | | Organic, post-redesign | 50% | 46% | | Blended | 38% | 25% | Which conclusion does the cohort split support?
MCQ-97Your team ships a new onboarding flow. Activation rises 8% but 30-day retention is flat. What should the PM do first?
MCQ-24A payments team's quarterly dashboard reports a 23% checkout completion rate. The dashboard counts every account active in the quarter in the denominator and treats a missing `checkout_completed` event as a non-completion. Ten moderated sessions with long-tenured merchants all completed checkout without trouble. An engineer finds the event is emitted only by the client build shipped 5 months ago; accounts created before that release make up 69% of the accounts in the dashboard. What does the evidence support?
MCQ-63A workspace product's growth dashboard defines a monthly active user as any device that opened the app in the last 30 days. For March it reports 2,100,000 MAU, 41,000 paying workspaces, and a trial-to-paid conversion of 1.95% (41,000 ÷ 2,100,000). The board target is 4%, so the VP proposes cutting the free tier. A data engineer re-runs March counting distinct account IDs instead of devices and gets 780,000 MAU; 1,320,000 of the devices belong to accounts that had another device active that month. Which conclusion does the corrected figure support?
MCQ-32A B2B SaaS team steers on accounts created. Over two quarters the metric rose 34%, but the growth came entirely from accounts that never started a trial; the share of new accounts completing a first real task fell from 44% to 19%, and support tickets about unusable trials tripled. The team has one quarter to fix its steering metric. Which conclusion is justified?
MCQ-73A SaaS company's quarterly satisfaction score fell from 64 to 52. The survey goes to every active account; the Q2 send added an in-product prompt and the number of responses nearly tripled. Churn and downgrades were flat quarter over quarter. The board meets Friday to lock next quarter's headcount and wants to know whether the customer experience actually changed. Which approach should the team take?
MCQ-75A payments team's service dashboard has 42 panels. Over one quarter the access log shows 1,240 views: 1,180 from the on-call rotation during incidents, 60 from everyone else. Panel-level counts show 6 panels (error rate, p99 latency, saturation, queue depth, deploy markers, dependency health) account for 91% of panel opens, and 19 panels were never opened. A separate weekly business review dashboard has 8 panels, 340 views, 12 distinct viewers, and is opened every Monday by the same 12 people. Which change does the data justify?
MCQ-91Your B2B SaaS product sends one analytics event per page view to a third-party vendor. Each event carries 38 properties, including `page_template`, `referrer_host`, `full_url` (with query string), `user_email`, and `account_id`. The only consumer of this pipeline is the growth team's funnel dashboard, which reads `page_template` and `referrer_host`. The vendor's retention default is 24 months and the team never changed it. A security review finds `full_url` has been capturing password-reset tokens. Which change removes the most exposure while keeping the funnel dashboard working?
MCQ-72A checkout redesign ships in 3 days. The spec lists four funnel steps (cart_view, checkout_start, payment_submit, order_confirm) but defines no events for them. Staging emits none of the four. The previous version's dashboard has cart_view and checkout_start counts only. The PM wants to ship on time and add events next sprint. What is the correct call?
MCQ-76Before launch, a team pre-registered a bet: ship feature X to all eligible accounts, and success is 60% of eligible accounts using it weekly by the end of Q3. The quarterly reads were 19% (Q1), 17% (Q2), and 18% (Q3). The enterprise cohort alone reached 55% while the SMB cohort reached 12%, feature page views rose 40% quarter over quarter, and the business case had forecast 15,000 weekly-active accounts (the Q3 read is 18,000). At the Q3 review, the team disagrees about what to do. Which action follows from the pre-registered threshold and the three reads?
MCQ-59A team pre-registers a two-week A/B test: 50,000 users per arm, one analysis at the end, α = 0.05. They watch the dashboard daily, and on day 3 (8,000 users per arm) the treatment shows +4.2% conversion at p = 0.03. The PM wants to stop and ship. Which move is defensible?
MCQ-61An A/B test of a new onboarding flow ran on 60,000 accounts. The pooled lift in 30-day retention was +0.4% (p = 0.62), and the team is about to ship. The test plan pre-registered one cut, seat count, and the results by that cut are: >5 seats, 21% of accounts, +22%; 2–5 seats, 16%, +9%; single seat, 63%, −9%. Which conclusion does the data support?
MCQ-67A checkout experiment runs four weeks with 5 million sessions per arm. Conversion is 3.40% in control and 3.427% in treatment: a relative lift of +0.8%, p = 0.018, 95% CI [+0.14%, +1.46%] relative. Before the test the team pre-registered +2.0% relative as the smallest lift worth shipping, because the change adds a permanent code path and a config flag costing about 2 engineer-weeks a quarter to maintain. What should the team do?
MCQ-65A B2B SaaS team's 90-day retention analysis shows accounts where a teammate was invited in the first week retained at 62%, against 24% for accounts with no invite. The PM proposes prompting every new account to invite a teammate. A randomized test on 20,000 new accounts lifts first-week invitations from 20% to 28%, and 90-day retention comes out 26.0% in the prompt arm versus 25.2% in control. Which conclusion does the test support?
MCQ-66A consumer app ships 24 experiments over a year, each with its own A/B test. Summed, the winning tests claim a 38% lift in the primary metric, weekly active days per user. Throughout the year the team also held 2% of its 40 million weekly users on the old experience. At year end that holdout reads 4.10 weekly active days per user against 4.47 for everyone else. Five of the 24 experiments ran on the same home-screen surface, so their gains counted the same users twice. What does the evidence support?
MCQ-68A ride-hailing team tests a driver bonus that pays for accepting short pickups. They randomize 4,000 drivers in one metro 50/50 for two weeks. Treated drivers average 12 rides/day, control drivers 10 rides/day, and the team reports a 20% lift. Before the test the metro averaged 44,000 completed rides/day; during the test it also averaged 44,000. Rider demand, not driver supply, sets how many trips happen. What should the team conclude?
MCQ-69A hypothetical checkout experiment that removed a shipping-cost surprise lifted conversion from 3.1% to 3.35% (n=40,000 per arm, p=0.01). The team wants to apply the same change to the signup flow, where shipping cost is not shown. What should they do next?
MCQ-60A mobile game studio ships a redesigned level-complete screen to 100% of players and measures sessions per player against the four weeks before launch. New installs (first 7 days) were about 55% of weekly active players during the test. | Week | Pooled | New installs | Returning players | | --- | --- | --- | --- | | 1 | +12% | +22% | -1% | | 2 | +6% | +14% | -3% | | 3 | +2% | +7% | -4% | | 4 | 0% | +4% | -5% | The studio must decide whether to keep the redesign. What does the data support?
MCQ-57A team pre-registers two guardrails before launching a new paywall: 30-day refund rate with a stop threshold of 1.5x control, and support contacts per 1,000 orders with a stop threshold of 1.25x control. Control refund rate is 4.0%; control support contacts are 40 per 1,000 orders. After two weeks: trial starts (the primary metric) rose 11%, treatment refund rate is 5.8%, treatment support contacts are 52 per 1,000 orders, and 90-day churn for converted users is 2.4x control. Which call does the stopping rule require?
MCQ-58A team wants to know whether a redesigned checkout raises conversion. Checkout gets 400 visitors a week, the smallest lift worth shipping for is 2 points on a 4% baseline, and the call has to be made before the holiday freeze in three weeks. Which plan holds up?
MCQ-71A team ships an optional bulk-edit feature. A randomised availability test assigns 2,500 users to each arm. In the available arm, 34% enable the feature; task completion is 71% in the available arm and 70% in the withheld arm. Among enablers, completion is 88%; among non-enablers in the available arm, 62%. The feature costs 2 engineer-weeks a quarter to maintain. What does the test support?
MCQ-34A team ships two changes to the signup flow: an import fix and guided setup. The import fix pilot ran with guided setup already enabled for the whole cohort, and that cohort activated 6 points above the 31 percent baseline measured over the prior eight weeks. Guided setup alone, in the closest analogue, moved activation 4 points. Which target should the team commit to for the quarter?
MCQ-16A team is deciding whether to build an 'auto-reconcile' feature. In a concept test with 12 target users, a facilitator walked through a static mock: 11 of 12 said they would use it (mean intent 4.4 out of 5). Shown the same mock with no narration and asked to describe what it does and when they would use it, 3 of 12 got it right. Offered a two-week pilot that required 30 minutes a week of manual data entry to get early access, 2 of 12 signed up. Which conclusion does this evidence justify?
MCQ-4A PM surveys 40 account admins about a planned bulk-export feature: 34 say they would use it weekly. She then interviews 12 of them, asking each to walk through the last time they exported data. Nine describe exporting once a quarter or less, and seven of those nine have never exported more than 50 rows. The team meets to decide what to build. Which decision follows from the evidence?
MCQ-13A B2B SaaS team surveys all 50,000 active accounts on six candidate features for next quarter's roadmap. The six features were drawn from the support tickets of accounts older than 2 years. 3,500 accounts respond: a 7% response rate. Feature X ranks first overall, picked first by 42% of respondents. The respondent mix: accounts older than 2 years are 20% of the base but 65% of respondents; accounts younger than 90 days are 30% of the base but 6% of respondents. Among over-2-year respondents X is first (55%); among under-90-day respondents X is fifth (4%) and their top pick is Feature Y. What should the team do?
MCQ-62A support org's quarterly review shows first-response time fell from 4.1 hours to 22 minutes after an auto-acknowledgement was added to the intake queue, while resolution time rose from 19 hours to 25 hours and CSAT fell from 74 to 62. The VP of support wants to keep the first-response target and add a bonus for hitting it. Which recommendation should the PM make?
MCQ-2A launch review has one quarter left. The team's problem statement is "We need a notification centre." Discovery found 12 admins miss 4 escalations a week; 78% of missed events are escalations that should auto-assign, and the rest are customer messages. The team has already spent 6 weeks building the notification centre. Which recommendation holds up?
MCQ-3Your workflow product has 4,200 accounts: 61% self-serve, 22% mid-market, 17% enterprise. Last quarter 340 accounts churned, 265 of them self-serve. You have 8 discovery interview slots for a feature that will ship to all tiers, and the PM's draft list is 8 enterprise accounts from the customer advisory board. Which interview list should replace the draft?
MCQ-9Last quarter your team lost 12 competitive deals. The CRM loss reasons: onboarding time 7, price 3, the competitor's new dashboard builder 2. The dashboard builder shipped eight weeks ago and got press coverage; two reps have asked for parity. Engineering says an equivalent is 9 weeks, and the planning cycle has room for one 9-week project. No work is currently scheduled against any of the three reasons. Which project does the loss data support?
MCQ-10A scheduling tool's landing page positions it against two enterprise vendors. An exit survey of 500 trial signups finds 64% arrived from a spreadsheet, 11% from Vendor A, 18% from nothing at all. The spreadsheet cohort activates at 22% and reports spending about four hours a week on scheduling; the other cohorts activate at 9%. The campaign ships in two weeks and the goal is trial-to-paid conversion. Which change should the review approve?
MCQ-1A billing team's invoice API returns date, customer and line items; the invoice ID exists only in the emailed PDF. Three enterprise accounts have asked by name for a CSV export of invoices, each naming a different column set: date, per-customer subtotal, raw line items. All three copy the invoice ID out of the PDF by hand each month. Support logs show 22 of 40 enterprise accounts reconcile by hand, median 3 hours per account per month. Engineering estimates the export at 11 weeks and adding the invoice ID to the invoice API response at 2 weeks. The team has one planning slot. Which action does the evidence support?
MCQ-12A B2B SaaS company reports 40% month-on-month growth in new accounts. Breakdown of last quarter's new accounts: paid acquisition 78% of signups with 9% month-6 retention; organic referral 14% with 44%; partner 8% with 31%. When paid spend was paused for six weeks, total new accounts fell 82% while organic and partner signups held flat. Which reading of product-market fit does this data support?
MCQ-21Your team has 30 customer interviews and one quarter to ship a single fix. The interviews and usage data give you this: | Problem | Named top-3 pain | Median account spend on a workaround | Notes | | --- | --- | --- | --- | | Duplicate invoice matching | 21 of 30 | 0 h/month | A 2023 default auto-merges duplicates; 2 accounts disabled it | | Month-end close reconciliation | 12 of 30 | 9 h/month | 4 accounts pay a contractor about $1,800/month for it | | Access provisioning | 9 of 30 | 0 h/month | One account paid an engineer 6 h to script around it 14 months ago | Which problem do you build first?
MCQ-5A 12-person team sells a reporting tool at $89/month. Of the 40 accounts that churned last quarter, exit calls show 26 pulled one quarter's numbers into a board deck and then stopped logging in, 9 moved to a cheaper spreadsheet template, and 5 named a missing SSO integration. The Q4 roadmap's top item is a 6-engineer-week interactive dashboard builder, justified by the job statement 'when I track my metrics weekly, I want a live dashboard, so I can spot trends.' What should the team do with the Q4 roadmap item?
MCQ-22A product-sense interview has 20 minutes and a 20-point rubric: 5 points for naming a user segment and the goal, 5 for stating two or three problems and choosing one with a reason, 5 for proposing two solutions with a trade-off between them, and 5 for closing with a success metric that measures the user goal. The candidate's transcript: 0:00–0:30 'I'd add dark mode to the maps app.' 0:30–14:00 describes toggle placement, colors, and animation. 14:00–18:00 'The user is everyone. The goal is engagement. Problems: not enough features. I'd pick dark mode. Solutions: dark mode and route sharing. Trade-off: dark mode is better. Metric: daily active users.' 18:00–20:00 answers implementation questions. Each rubric section is all-or-nothing. What score should the interviewer record?
MCQ-6A PM must give the board a market size for a new product that only works with one ERP. The analyst report puts the category at $12B. The CRM lists 2,100 accounts in the target segment running that ERP. In the 40-account pilot, 6 accounts bought, averaging 8 seats at $1,200 per seat per year. Which figure should the PM report?
MCQ-8A B2B SaaS team surveyed 200 accounts; the average stated acceptable price was $180/mo. They then ran two paid pilots: 80 accounts offered $99/mo, of which 4 signed (5%); 80 accounts offered $49/mo, of which 16 signed (20%). The forecast for the $99 pilot was 25%. Which conclusion do the numbers support?
MCQ-23Your team is forecasting first-year accounts for a new compliance-automation product sold to mid-market finance teams; every buyer runs a security review before signing. A paid landing-page test drew 6,000 visitors at $0.80 a click, and 4.0% joined the paid waitlist: 240 signups in 90 days. Three analogues are on the table, each with its 90-day waitlist signups and its first-year accounts: a self-serve developer tool (2,000 signups → 40,000 accounts), a mid-market finance tool that also requires a security review (400 → 3,200), and an enterprise compliance tool that requires a security review (15 → 180). The mid-market segment holds about 150,000 firms. Which first-year figure belongs in the forecast?
MCQ-14A self-serve analytics product has one roadmap slot this quarter. The board set one goal: lift new-account activation, flat at 31% for two quarters. Last quarter's support tickets cluster into three groups: | Cluster | Tickets | Distinct accounts | | --- | --- | --- | | export formatting | 380 | 6 | | first import fails | 94 | 210 | | billing confusion | 120 | 88 | The PM has one roadmap slot. Which cluster should take it?
MCQ-15A product team's opportunity assessment for a Q3 initiative covers 4,000 accounts in the target segment. The behaviour that must change is weekly imports per account, from 1 to 3; the value if it holds is $1.2M annualized; the failure condition is a population average below 1.6 weekly imports per account at day 90. On day 90 the dashboard shows 2,400 accounts at 3 imports/week and 1,600 still at 1. The team lead reports the check passed and asks for the same funding next quarter. Which read of the day-90 data is right?
MCQ-26A team ranks next quarter's work with RICE, (reach × impact × confidence) / effort, where reach is the number of users who encounter the change during the quarter. The product has 12,000 accounts; 900 of them open the bulk-edit screen, 3,000 use import, 8,000 open saved views, and 1,300 sign up. The scored table: | Item | Reach | Impact | Confidence | Effort | | --- | --- | --- | --- | --- | | bulk edit | 12,000 | 1 | 100% | 2 | | import retry | 3,000 | 2 | 80% | 1 | | saved views | 8,000 | 1 | 50% | 1 | | new-signup fix | 1,300 | 3 | 100% | 1 | The review must fund one item. Which recommendation holds up?
MCQ-27A platform team has 18 team-weeks of capacity next quarter. Four candidates are on the list, each with a cost in team-weeks and an annual value: search relevance rewrite, 12 weeks, $840k; checkout latency fix, 6 weeks, $600k; admin audit log, 4 weeks, $200k; mobile push revamp, 8 weeks, $400k. The VP wants to approve all four because each returns more than it costs. What should you do?
MCQ-29Your platform team has 5 engineer-weeks of debt capacity this quarter, separate from feature work. Three debt proposals are on the table, with last year's incident counts and the committed roadmap items each one blocks: | Area | Incidents last year | Committed items blocked | Estimated fix | | --- | --- | --- | --- | | checkout service | 9 | 3 | 4 weeks | | admin tooling | 1 | 1 | 1 week | | reporting service | 2 | 0 | 6 weeks | The reporting service is the oldest module and the team has wanted to rewrite it for two years. Which allocation of the 5 weeks is best supported by this data?
MCQ-30You own a 4-engineer platform team with 6 weeks left in the quarter: 24 engineer-weeks, fully committed. The plan: billing migration, 10 engineer-weeks, contractual date in week 6; audit log, 8 engineer-weeks, the prerequisite for a compliance review in week 5; retry fix, 6 engineer-weeks, no external date. A VP asks for a partner export, which you estimate at 6 engineer-weeks, and wants it this quarter. There is no overtime budget. Which plan delivers the export this quarter without breaking a dated commitment?
MCQ-44A platform team of 5 engineers plans a 10-week quarter, giving 50 engineer-weeks of capacity. In each of the last two quarters it committed all 50 engineer-weeks to planned work. Unplanned work (support escalations, incidents and compliance reviews) consumed 32% of capacity in Q1 and 36% in Q2, and the team met 4 of 5 commitments in Q1 and 3 of 5 in Q2. Hiring is frozen for the next two quarters. What should the team commit to for the coming quarter?
MCQ-47Your team plans each quarter by scoring backlog items and taking the top ones. A data-residency rule takes effect 1 July. The compliance work scores 40; two already-committed features score 900 and 700. The compliance work is estimated at 11 weeks, and over the last eight quarters the team has finished 1–3 weeks late on every item estimated above 8 weeks. The two features together take 9 weeks. Today is 1 March. What should the plan do?
MCQ-48A subscription company spends $1.5M a quarter to acquire 5,000 accounts, a $300 cost each. Month-6 retention has held at 20% across the last three cohorts, and the curve is flat from month 6 on, with about 1.5% of the survivors churning each month. Each surviving account contributes $20 a month in gross margin. Finance proposes doubling acquisition spend next quarter. Which conclusion is justified?
MCQ-49A tutoring marketplace operates in two cities. Last week: City A had 20,000 student requests, 2,000 active tutors, a 15% match rate, and a 2-day median time to match. City B had 1,000 student requests, 1,500 active tutors, a 98% match rate, and a 5-minute median time to match. The growth team has budget for one intervention next month: recruit 1,200 tutors or acquire 3,000 students, in either city. What should it do?
MCQ-40A B2B SaaS product sells three tiers: Starter ($49/mo, 5 seats), Team ($199/mo, 25 seats), and Enterprise (custom). Audit logging sits in Enterprise only. Over the last two quarters, sales granted 19 one-off contract exceptions to Team customers who needed the audit log, and 6 of those deals were lost; a compliance rule that took effect last year makes audit logging mandatory for every customer in Team's segment. The seat limit produced 1 exception and 0 lost deals; the API-volume cap produced 2 exceptions and 1 lost deal. Which conclusion does this data justify?
MCQ-43Your team's strategy doc lists four bets and five things you will not build. Over the last two quarters, 14 requests reached the roadmap: four matched the bets and shipped; five fell inside the 'will not build' categories, and three of those shipped anyway after an exec escalation; the remaining five were in areas the doc never mentions, and all five shipped. The VP asks whether the doc is doing its job. What should you conclude?
MCQ-39Your team sells a $20,000 ACV product at 75% gross margin and closes 300 mid-market deals a year. A competitor cuts list price 25% and your sales lead asks to match it. The last 25 losses were reviewed: price was the deciding factor in 6, and integration fit or implementation risk decided the other 19. What does matching the cut require, and what is the right move?
MCQ-42An API product's quarterly revenue fell from $1.80M to $1.28M over the year, a 29% drop. The fall is uniform (the top 20 accounts, mid-market, and self-serve each dropped 26–31%) and no deal was lost on features; two competitors cut list prices 30% in Q2. Gross margin fell from 71% to 58% while fixed infrastructure costs stayed flat, and fully-loaded cost to serve is now $1.5M a quarter against $1.28M of Q4 revenue. Two engineers maintain the product. Which recommendation holds up?
MCQ-51Your roadmap uses three bands: committed (may be repeated to customers as-is), planned (theme only, with a 'not yet scheduled' caveat), exploring (internal only). Item R was marked committed and the repeat log shows sales repeated it to 5 accounts; two of them have already asked their rep for the ship date. Item R has just slipped out of the quarter. The same slide reached 9 other accounts that never had item R repeated to them, and support has 63 open tickets about roadmap items. What should you do?
MCQ-31A B2B SaaS company publishes a quarterly roadmap. Across the last four quarters, 14 items carried dates, 5 shipped on their original date, and the account team quoted 9 of the dated items in signed contracts. Date reliability splits cleanly: items due within 6 weeks hit their date 90% of the time, items more than two quarters out hit it 20%. Every one of the 31 escalations that followed a slip was on an item sales had quoted. Which change to the roadmap process removes most of the escalations without cutting what sales can promise?
MCQ-46A PM at a B2B SaaS company gets an escalation from a customer. The feature request has no fixed scope: two integrations, each estimated at 3–7 weeks by the team, plus a shared auth change estimated at 2 weeks that must land before either integration can be tested. The two integrations must be built sequentially by one engineer. The customer has said the renewal depends on having both integrations live by week 12. Which response should the PM choose?
MCQ-45Six engineers, a 13-week quarter: 78 engineer-weeks of capacity. The plan commits 30 engineer-weeks to A (a contractual deliverable due in week 12), 24 to B (a platform migration that A's final integration depends on), and 18 to C (internal tooling with no external dependency), leaving 6 engineer-weeks of buffer. In week 3 a regulatory change lands: it must ship by week 10 and its scope is fixed at 20 engineer-weeks. Which plan fits the quarter without breaking a committed date or the regulatory deadline?
MCQ-64A team has 12 engineer-weeks before a committed launch date. The build estimate is 20 engineer-weeks and the plan also carries a 2-week load test. The estimate breaks down as: | Slice | Engineer-weeks | Share of target usage | | --- | --- | --- | | core create/edit flow | 6 | 68% | | spreadsheet import | 4 | 17% | | approval routing | 5 | 9% | | audit log | 3 | 5% | | bulk export | 2 | 3% | Which plan is justified?
MCQ-78A B2B analytics product ships on a contractual date six weeks out. Four capabilities are in flight, each with two engineer-weeks of work remaining: CSV import, a report builder, filter-aware export, and a dashboard builder. Six engineer-weeks of capacity remain. The contract's acceptance test is a customer importing their own file, building a report, and exporting it without a workaround; the dashboard builder is a separate bet with no customer commitment. Which plan should the PM take to the release?
MCQ-41A payments team has one engineer free for the next two-week sprint, and two candidates compete for it. A rounding defect in the invoice PDF affects 1,100 accounts a month; each invoice is off by an average of $12, and finance issues a manual credit for every affected account at a handling cost of $15 per account. A self-serve upgrade flow is modelled at 300 incremental upgrades in the quarter, each worth $200 in first-quarter gross margin. Only one can ship this sprint. Which call does the arithmetic support?
MCQ-18A B2B analytics team has two engineers free next quarter. The report builder is one candidate: five rounds of usability fixes over seven weeks raised task completion from 40% to 88%, but only 4% of accounts opened the builder in any month before or after. The ingest pipeline is the other candidate. Which recommendation holds up?
MCQ-20A platform team's quarterly review shows 42 of 42 services registered in the internal catalog, 38 of 42 teams reporting satisfaction, and 17 teams still running their own Jenkins deploy jobs. The platform's deploy path takes 9 days from request to first production deploy; the Jenkins path takes 40 minutes. Of the 17 teams with Jenkins jobs, 12 also deploy at least one service through the platform. The top support ticket category is 'how do I add a service to the catalog' (31 tickets). The team has one quarter of budget left. Which investment does the evidence support?
MCQ-19A platform team has declined the same request three times in 18 months: bulk CSV export of audit logs. Each round cost about 8 hours of re-analysis (5 PM, 3 engineering), and the requester reopened it because nobody could find the earlier reasoning. The blocker is that a full export scans 400 GB and pins a read replica for about 40 minutes; a migration to a columnar store lands in Q3 and is expected to cut that to under 2 minutes. The requester is a $120k-ARR account. What should the PM do with this fourth request?
MCQ-28A platform team owns an API that a mobile team must integrate against. The API is estimated at 8 weeks; the last three comparable APIs each finished 3 weeks late, because a third-party schema changed mid-build, and the API's critical path is a single design decision owned by one engineer. Mobile needs 5 weeks of integration once the response shape is final. The release is committed at 13 weeks. Freezing a contract stub costs the platform team 2 days; mobile's integration then needs 3 days of rework when the real API lands, against 11 days if mobile builds on the unfrozen draft. What should the two teams do?
MCQ-38The config-layer rewrite would take two of the four engineers for a quarter. The next two quarters' roadmap is committed and the platform team says the rewrite makes several of those items cheaper: the dependents span billing, onboarding and reporting. No customer contract or compliance deadline depends on it, and it is not yet funded. Which response should the product manager choose?
MCQ-36A B2B SaaS company is choosing how to provide full-text search over the documents in its customer portal. A vendor licence is $90k a year, takes six weeks of one engineer to integrate, and about two engineer-months to remove later. Building in-house is 12 engineer-months for a first version, then 1.5 engineers permanently on relevance tuning. A partner wants 8% of the portal's revenue, which is $4M ARR today, and would own the roadmap. Loaded engineer cost is $20k per engineer-month. In win/loss interviews from the last 40 closed deals, search came up in 3; the product's differentiation is its workflow automation. Which path should the company take?
MCQ-35A payments team is migrating 240 enterprise accounts from a legacy ledger to a new service. The full estimate is 24 engineer-months: 8 for automated reconciliation, 6 for support tooling, 4 for cutover automation, and 6 for account migration. The new service serves no traffic until every account has been migrated; it runs in shadow mode until then. Support tooling is required before any cohort larger than 60 accounts can cut over; smaller cohorts can use the existing manual runbook. The legacy vendor bills a flat $24,000/month regardless of how many accounts remain on it, and the contract cannot be reduced before it ends in 18 months. Finance offers 12 engineer-months. Which decision is defensible?
MCQ-37A team is retiring a legacy export endpoint. 120 accounts call it daily; 4 are top-10 by revenue. They announced a 60-day sunset. At day 45, 18 accounts remain, including 1 top-10 account whose nightly batch now calls the endpoint three times more than before. The other 17 accounts each need 0.5 engineer-day to migrate. The top-10 account needs a custom adapter: 5 engineer-days to build and 10 business days to validate. Two engineers are available for the remaining 15 business days. Which plan is sound?
MCQ-90A SaaS company is moving 10,000 existing accounts from $20/month to $29/month. A pilot on 2,000 accounts ran two arms, both with 90 days' notice before renewal. The first 1,000 accounts got no grandfathering and churned at 9% over the first renewal cycle. In the second 1,000, the 400 accounts older than 3 years were grandfathered at $20 for 12 months; that arm churned at 4%, leaving 960 accounts, 400 of them at the grandfathered price. Baseline churn is 3%. The remaining 8,000 accounts are next, and finance will pick the rollout that produces the most revenue over the next 12 months. Which should they choose?
MCQ-74A subscription business builds its Q3 plan as a baseline plus named contributions: existing-cohort baseline 5.4M (no new work), pricing change 1.4M, new self-serve segment 1.4M, annual-plan upsell 0.6M. 8.8M total. Q3 closes at 6.7M, measured the same way: baseline 5.4M, pricing 0.5M, new segment 0.2M, upsell 0.6M. Which Q4 plan does the close support?
MCQ-87A subscription product runs two changes in the same quarter. The paywall team, measured on trial starts, ships a shorter paywall: trial starts rise 15%, but trial→paid conversion falls 20% (relative) because the added traffic is lower-intent. The onboarding team, measured on trial→paid conversion, ships a longer onboarding: conversion rises 10%, but trial starts fall 5% from the added friction. Baseline is 100,000 trial starts and 20% conversion a month, and both changes stay live all quarter. What is the net effect on monthly paid subscribers?
MCQ-50An enterprise contract renews in 3 weeks and requires a WCAG 2.2 AA conformance report for the checkout flow. An audit of that flow returns 40 issues: 5 block task completion outright, 15 are reachable only with a workaround, 12 are minor but each fails a Level AA success criterion, and 8 fail only Level AAA criteria. Two engineers are available full-time and each closes about 6 issues per week. Which commitment should the team make?
MCQ-11You have 8 engineer-weeks before a 12-account pilot of a B2B data-onboarding product. The hypothesis under test is that a workspace admin can connect their own data source without a call. Pilot accounts are already provisioned, so nobody needs SSO to sign in; two of the twelve accounts asked for SSO during sales, and compliance wants audit-log export before the first renewal, not the pilot. The previous cohort saw 8 of 10 admins abandon at the import step. Which scope do you commit to?
MCQ-17A payments team must decide by Friday whether to ship a redesigned four-step refund flow. The open risk is whether customers can find the refund entry point and complete the steps unaided. Design has three working days and eight testers are booked. Which artefact should the team build for this study?
MCQ-7A 4,000-account B2B tool has 90-day retention cut by industry vertical. Reviewer share is the fraction of accounts that run a weekly review meeting. The team can fund one onboarding change next quarter. | Vertical | Accounts | Reviewer share | Retention, all | Retention, reviewers | Retention, others | | --- | --- | --- | --- | --- | --- | | A | 900 | 53% | 47% | 70% | 21% | | B | 700 | 48% | 45% | 69% | 23% | | C | 1,200 | 50% | 46% | 71% | 20% | | D | 600 | 55% | 48% | 70% | 22% | | E | 400 | 50% | 44% | 68% | 19% | | F | 200 | 55% | 49% | 70% | 24% | Which action should the team fund?
MCQ-79Your team estimates a launch at 8 weeks: 3 weeks of CRUD work it has done before, 4 weeks migrating an undocumented legacy schema, and 1 week integrating a vendor API whose production rate limits differ from the sandbox. Leadership wants the date committed at 5 weeks. Last quarter a similar push took an 8-week estimate to 5; the work ran 11 weeks including 2 weeks of rework, and the team has since refused to quote anything below its worst case. What should you do?
MCQ-83You are the delivery lead for a vendor data migration serving 8 enterprise customers. On March 3 you learn the cutover will slip from March 31 to May 12 because the upstream API changed. Two customers have internal training scheduled for April 7 and April 14; one contract has a $25k/week penalty starting April 1. A partial path can move 3 of 5 data domains by March 31. What is the best next action?
MCQ-80On 3 March a usage dashboard ships to all 4,200 accounts. Support macros and the docs page still describe the old usage report, and the pricing FAQ has no entry for the dashboard. By 7 March support has logged 380 tickets: 140 ask whether the dashboard is billable, 90 ask where the old report went, 60 report a chart that reads zero for annual-plan accounts, and 90 are unrelated. The pricing owner confirms the billing decision is still open. At the 10 March review, which conclusion does the ticket data support?
MCQ-81A change to the account-export service is planned for staged rollout. The only known risk is for accounts with more than 10,000 records: in a canary, 5% of those accounts hit a timeout, while accounts at or below 10,000 records show no failures. The rollout plan is: stage 1, 1% of 20,000 accounts for 4 hours; stage 2, 50% for 2 hours; then 100%. Stage 1 drew 200 accounts, none with more than 10,000 records. Stage 2 drew 10,000 accounts, 1,200 with more than 10,000 records. The current stop condition halts the rollout if the overall error rate across the stage exceeds 1%. During stage 2 the overall error rate is 0.6%. What should the rollout owner do before proceeding to 100%?
MCQ-82A checkout change, a 40-line diff to the authorization retry path, went out at 14:05. Before launch the team agreed the rollback rule: revert if the 15-minute failed-authorization rate exceeds 2.0%, or if p99 authorization latency exceeds 900 ms. At 14:20 the dashboard reads 4,800 attempts, 118 failures, p99 640 ms. The last time this path sat at 2.4%, it reached 9% within 40 minutes and 1,100 orders were declined. The on-call engineer has 20 minutes before the next review. What is the right call?
MCQ-84A payments team has logged four deploy incidents in 14 months. Each postmortem named the on-call engineer who skipped step 4 of the deploy runbook and closed with the action 'retrain the on-call'. Step 4 is a manual check, and the deploy tool proceeds whether or not it is done. Incidents 2, 3 and 4 each skipped step 4. A fifth incident is now open. What should this review produce?
MCQ-86Your payments service runs at p99 = 210 ms. A senior stakeholder, the VP of Engineering, wants to migrate to a new provider before the three-week change freeze, citing the vendor's benchmark of 90 ms p99. Your own sandbox run at production traffic shape measured 190 ms. The decision meeting is Thursday. What should you do?
MCQ-88You lead the platform team and send the VP a status every Friday. In week 3 a dependency your team owns blocked the payments ledger migration. You expected to clear it yourself and did not raise it. In week 12 the VP forwards you a message from the payments lead asking about the same blockage. The migration has a regulatory deadline 3 weeks out, and clearing the dependency needs two of your engineers for the full 3 weeks. Those two engineers are allocated to the Q3 roadmap item the VP approved last week. Which action do you take now?
MCQ-89A PM and a staff designer disagree about a 4-step onboarding flow. Completion is 38%, and 61% of drop-off happens at step 3, which asks for company size. Design says users abandon because they cannot tell how many steps remain; the PM says the company-size field is the blocker. Launch is fixed in 5 days by a partner commitment. A 5-user usability test takes 1 day. A power calculation says an A/B test needs 9,400 users per arm to detect the 2-point difference design predicts, and the flow gets 12,000 signups a week. Which action should the PM take?
MCQ-92A team is launching a new permissions model to 2,000 accounts next week. A two-week pilot on 500 accounts logged 60 support contacts at an average handling time of 9 minutes: 33 about a permission-denied error message, 15 about how to invite teammates, and 12 about billing. The permission-denied message is a one-line copy change already merged in the release branch; the other two are answered by help articles. Which action should the team take before launch?
MCQ-93A four-engineer team took over a payments-retry service in January. The package: the repo, a 40-ticket backlog, runbooks, an architecture diagram, the 2023 retry-storm postmortem, and six weeks of overlap covering a full on-call rotation. By the end of the overlap they had closed 22 of the 40 tickets and passed all 14 runbook smoke tests. In March they deleted a 30-second jitter on retries as dead code; during the next ledger outage, retries synchronized and ledger p99 went from 120 ms to 4.1 s, repeating the 2023 signature. Which sign-off check would have caught this before the overlap ended?
MCQ-94You own checkout and need the payments platform team to add an idempotency key to the charge API before your March 14 launch. You have no authority over them. Their published quarterly goal is to cut pages per quarter from 24 to 12; last quarter 14 of their 24 pages were duplicate-charge alerts from retried requests. Their sprint planning closed Friday and the next sprint starts in three weeks. Which move should you make?
MCQ-98You join a payments team as a staff engineer three weeks before its launch review. The scope has five items. The team's decision log for the last two quarters records nine stopped items; three of the five in scope appear there, each stopped for a blocker the log still marks open: two for 'no owner for the ledger migration' and one for 'regulator has not signed off on the new KYC flow'. The two items with no stop record have passed readiness review and have named owners. The team is split 4-2 in favour of shipping all five. Usage data shows the feature is live for 2 of the 40 target accounts. Which action should you take at the review?
MCQ-99You start as PM for a checkout team on day 1: six engineers, two-week sprints, so six sprints before your day-90 review. Two things are fixed: a tax-calculation fix is contractually promised to the EU launch and must ship by sprint 2, and the funnel dashboard was last refreshed five months ago. What is known: 38% of sessions abandon at the address step, and an A/B test the previous PM ran last year showed a shorter address form lifts conversion 2.1%, but the change was never shipped. The VP of Product has said the day-90 review will be judged on one thing: a customer-facing metric that moved because of something you shipped. Which plan is defensible?
MCQ-25A PM is writing the planning document for a 9-engineer, 2-designer team. Last cycle the team worked from a 14-page spec: 3 of 9 engineers read it fully, and two engineers spent 5 weeks building overlapping services because the scope boundary was stated only on page 11. The PM has one week before kickoff. Which action is most likely to prevent the same overlap this cycle?
MCQ-77A payments team is replacing a CSV importer for expense reports. The old importer committed rows 1–3,999 when row 4,000 failed validation; support reconciled 640 accounts over 4 weeks. The new importer must accept 50 MB files, reject duplicate rows, and handle permission-denied uploads. The tech lead has one week to write the spec engineers will build from. Which spec content should the tech lead require?
MCQ-85A platform team has 6 engineer-weeks unallocated this quarter. Its largest account (18% of ARR) escalates six weeks before renewal: their compliance team needs a record of admin actions. 47 other accounts have requested the same record over the past year. A general audit trail is estimated at 8 engineer-weeks; a bespoke export for this account at 5 engineer-weeks, usable only by them. Support can produce a manual export for the account in two days. Which plan is sound?
MCQ-95An interviewer scores a 45-minute behavioural loop with three elements, each 0–5: decision ownership, measured outcome, and change since. A candidate's failure story runs: 'We missed the Q3 launch by three weeks because the platform team slipped on the API. I escalated to my manager and we shipped in October. I learned to communicate earlier.' The debrief records 0/5 on decision ownership, 0/5 on measured outcome, 2/5 on change since. The candidate needs 10/15 total to advance. Which single follow-up question recovers the most scoreable evidence?
MCQ-96A candidate rehearses a conflict story for a staff loop: a teammate wanted to raise the DB connection pool from 20 to 60; the candidate argued the pool was fine and took it to the tech lead, who sided with the candidate; three weeks later p99 went from 40 ms to 900 ms at 2,400 concurrent users and the pool was raised to 60. The draft ends: 'I held the line on capacity, and the lead backed me.' In a mock loop, three of four interviewers wrote 'no evidence of learning.' Which single edit most directly fixes what they flagged?
MCQ-100You have three questions left in the final 10 minutes of the loop. The debrief form scores four lines, each 0–2: a line with no question scores 0, one question scores 1, and two or more score 2. The weights are how decisions get made 50%, what the first year is judged on 30%, ownership scope 15%, technical depth 5%. The loop has already produced strong evidence (2) on ownership scope and technical depth. Which three questions do you ask?