Skip to content
Tech Interview Prep
Technical interview guide

Product Roadmapping

Communicating direction and sequencing over time, without over-committing to dates that will be wrong.

Read
30 min
Practice MCQs
25
Interview Q&A
25
Edition
v4
Editorial status
Reviewed

Scope: GOV.UK Service Manual, Scrum Guide 2020, Atlassian product roadmapping guidance, DORA, and NIST RMF current 2026-08-31.

Curated: · Written: · Reviewed:

25 questions

MCQ-8A platform team of 6 engineers plans a 13-week quarter. Its roadmap lists five feature launches committing 76 of the 78 available engineer-weeks. Reliability work, dependency upgrades and an in-flight service migration are tracked in the platform backlog but appear nowhere on the roadmap. Over the last four quarters, that work consumed 22, 27, 31 and 24 engineer-weeks. What is the most defensible read of this roadmap?
MCQ-17A platform team of six engineers plans a 12-week quarter, giving it 72 engineer-weeks of capacity. The roadmap lists five features at 12 engineer-weeks each and leaves the remaining 12 engineer-weeks unassigned. In each of the previous three quarters the team shipped three of the five planned features; the rest of its capacity went to two incident-driven migrations, a runtime upgrade, accessibility remediation, and decommissioning a legacy service: about 30 engineer-weeks per quarter. What does the arithmetic say about this quarter's plan?
MCQ-1A platform team's annual roadmap lists 40 items spread evenly across four quarters: 10 per quarter. Over the last four quarters the team shipped 6, 7, 5 and 6 items. Sales and support treat every roadmap item as a commitment, and the PM has to make the plan defensible for the year ahead. Which change best fits what the numbers show?
MCQ-23A six-engineer platform team has shipped about 4 roadmap items a quarter for four quarters running, and its quarter-end work-in-progress over that stretch ranged from 4 to 16 items without changing that rate. Twelve items are open now, and product wants the average time from an item being picked up to it shipping cut roughly in half over the next two quarters, with no new hires. Which change to how the roadmap is run does that?
MCQ-3A team keeps one board. Four quarterly outcomes sit at the top; below them are 180 estimated, assigned delivery items. The product owner reorders the whole board every sprint, and sales quotes ship dates from it. Over six weeks, sales has given three different dates for the same feature. What causes the dates to keep moving?
MCQ-7A product team keeps `roadmap.yaml` as the only record of its 24 initiatives. Each entry has `priority` (P1, P2, or P3), `stage` (`planned`, `committed`, or `cancelled`), and `audience` (`internal` or `external`). The sales deck lists initiatives where `audience` is `external` and `stage` is not `cancelled`; the exec summary lists initiatives where `priority` is P1 or P2 and `stage` is not `cancelled`; the engineering plan lists every initiative. The file holds 6 P1, 8 P2, and 10 P3 initiatives; 11 are `external`; 4 are `cancelled`: one P1, one P2, and two P3. The two `external` initiatives that are `cancelled` were cancelled last week; the P1 and P2 `cancelled` initiatives were cancelled a year ago. After a refactor, the generated views show 11 initiatives in the sales deck, 14 in the exec summary, and 24 in the engineering plan. The sales deck still shows the Payments revamp, which is `external`. Which defect explains the counts?
MCQ-2A team owns the checkout service. Its p95 checkout latency is 1.8 s, and 1.2 s of that is a third-party payment provider's API call the team cannot change. The team is writing the Q3 roadmap item for checkout latency. Which anchor lets the team declare pass or fail at quarter end on work it controls?
MCQ-4A B2B SaaS team keeps a Now/Next/Later roadmap. SAML SSO sits in Later. Six of its thirty enterprise accounts have asked for it, sales has begun mentioning it on calls, and there is no design doc and no estimate. All three backend engineers are committed to a database migration through the end of Q3. In the roadmap review the PM moves SSO to Now and tells sales to quote a Q3 delivery date. What is the correct read of that move?
MCQ-20A PM presents a 12-month roadmap with committed ship dates for each quarter. The core workflow has been tested with 4 of 40 target accounts, and there is no retention or pricing data yet. Two committed dates have already slipped, and sales has been quoting them to prospects. Which practice is the anti-pattern that produced this?
MCQ-6A payments team keeps a rolling 12-month roadmap. Items in the next quarter carry story-level acceptance criteria and estimates with a median error of ±20%. Items six to twelve months out are one-line themes. Over the last eight quarters, 55% of items planned more than two quarters ahead changed scope or were dropped, and their estimates missed by a median of 3x. A VP asks the team to write full acceptance criteria and a committed delivery date for a Q4 item, three quarters out, so it can go into a customer contract. What should the team do?
MCQ-10A payments team has 12 sprints of throughput history: a median of 5 stories per sprint and a 15th-percentile throughput of 3 stories per sprint. The next release backlog is 45 stories. The team reports '9 to 15 sprints: 9 at the median, 15 at the 85% confidence bound.' A director signs a customer contract naming 15 sprints as a guaranteed date with a penalty for slippage, and adds no capacity and no change-control clause. Which statement about this arrangement is correct?
MCQ-5You own the roadmap for a 4-engineer platform team running 2-week sprints. Across the last six sprints, 25% of each sprint's capacity went to interrupts and on-call. The next item, migrating 40 services to the new auth library, is estimated at 12 engineer-weeks. A partner team's contract requires their client to cut over by the end of Q3, six weeks from today. What should the roadmap show for this item?
MCQ-16A payments team of five engineers works in two-week sprints. A regulator's new data-retention control takes effect in eight weeks; missing it halts card processing. The team's committed roadmap holds 18 engineer-weeks of feature work. The compliance change is estimated at 26 engineer-weeks, and 25% of every sprint is consumed by on-call and support. Hiring is frozen and no contractor can be onboarded before the deadline. What must the team do to hit the deadline?
MCQ-19A payments team has five engineers and six weeks left in the quarter. The roadmap has three committed items: Feature A (regulatory, 2 engineers for 4 weeks), Feature B (customer commitment, 2 engineers for 6 weeks), Feature C (internal cost reduction, 1 engineer for 6 weeks). A major incident caused 3 hours 20 minutes of downtime; the postmortem requires 3 engineer-weeks to add cross-AZ failover. What should the team do with the roadmap?
MCQ-11You own the roadmap item "Move checkout session state to Redis to cut p99 from 900 ms to 300 ms", committed for Q3. Production peak is 20k session ops/s. A load test at exactly that peak shows the shared Redis cluster saturating at 12k ops/s, with p99 at 1.4 s once saturated. The platform team's upgrade, which raises the ceiling to 60k ops/s, is scheduled for Q4 and cannot be pulled forward: it depends on a vendor release. The commitment review is tomorrow.
MCQ-9A staff engineer reviews a 12-week quarter plan. Team A's Checkout v2 is recorded as "blocked on Ledger API: needed by week 6". Team B owns the Ledger change: 3 weeks of work, with 4 weeks of committed work already queued ahead of it. Checkout v2 needs 3 weeks after the Ledger change lands. Team A's release train departs at the end of week 9. Team A also has a 2-week local ledger stub it could build immediately, letting Checkout v2 start in week 3, but the stub appears nowhere in the plan. Which assessment of the plan is correct?
MCQ-15A team has 5 engineer-weeks of capacity left in Q1, and the work is not parallelizable. Two roadmap items compete: SSO for a $200k renewal with a hard deadline in 3 weeks (3 engineer-weeks), and a reliability fix for a bug causing one sev-2 incident per month, each costing about $80k in credits and on-call (2 engineer-weeks). The product manager is on leave; the interim product lead has authority. The CEO says 'don't lose the renewal.' The VP Engineering says 'reliability first.' What should the interim product lead do?
MCQ-14A 9-engineer platform team's quarterly plan has a Now list and a Not now list. The log-retention entry reads: "blocked on the auth migration (lands week 8); revisit if it lands early or a compliance deadline appears." In week 5 the migration lands and a compliance deadline is set for week 12. Log retention is 5 engineer-weeks; the team has 2 engineer-weeks free before its next committed item. What should the team do with the entry?
MCQ-21You own the roadmap for a platform team with 4 engineer-weeks of unallocated capacity over the next 10 weeks; the rest is committed to SOC 2 remediation that the audit date fixes. A customer that is 15% of ARR ($1.8M/yr) renews in 10 weeks and has told sales it will leave unless it can export its data to its own Snowflake account. Engineering estimates that export at 4 engineer-weeks built against that customer's schema, reusing the existing export pipeline, or 10 weeks built as a configurable connector any customer can point at a destination. What do you commit to?
MCQ-22A team owns its quarterly roadmap and the activation-rate goal on it; the platform org's Q4 migration is committed and cannot move. Q4 is 5 engineers × 12 weeks = 60 engineer-weeks, and the migration is scoped at 36 of them. Q3 shipped four of six roadmap items; the two that slipped were the only two that could move activation, and each needs its full 15 engineer-weeks to produce a readable result. Activation ended Q3 flat at 12% against a 20% target. What should the Q4 roadmap review conclude?
MCQ-25A payments team's Q3 roadmap committed to 12 items, each tied to one of two outcome targets: raise checkout conversion from 61% to 65%, and cut p95 authorization latency from 480 ms to 300 ms. At quarter end, 11 of the 12 items had shipped by their planned dates. Conversion was 61.4% and p95 was 470 ms. Three unplanned items (a regulator-mandated field, a partner API change, and a fraud-rule fix) had consumed about 45% of the quarter's engineering capacity. Which conclusion about the roadmap is best supported?
MCQ-13At the quarterly roadmap review, Squad A reports 41 of 44 committed items closed, 3 of them rolled back within a week; the activation metric for the flow those items targeted reads 12.1% against a 12.0% baseline, with ±0.4 pp week-to-week noise. Squad B reports 18 of 30 committed items closed; the same activation metric moved from 12.0% to 19.2% (p < 0.01), and the 12 items it dropped were re-scoped after interviews showed the flow they targeted was used by 2% of accounts. Which reading of the two squads' progress is correct?
MCQ-18A team pre-registers a decision rule for a new onboarding flow: ship only if the primary metric (7-day activation) shows a lift of at least +3.0pp over control AND the guardrail (support-ticket rate) does not exceed its +0.5pp non-inferiority margin. After 21 days with 3,000 users per arm: treatment activation 40.1% vs control 38.2%; treatment support-ticket rate 2.4% vs control 2.1%; p = 0.04 for the activation difference. What does the pre-registered rule support?
MCQ-24A team ships a redesigned onboarding flow to 5% of new signups, pre-registered for a 4-week read on activation rate. At the 2-week mark: activation is 41% (1,200 users) versus 38% (1,180 control), 95% CI on the difference −0.9 to +6.9 points; support tickets mentioning the flow rose from 5 to 14, and 11 of the 14 say they cannot find the skip button. What is the right next action?
MCQ-12A payments team is deciding whether to replace its nightly batch reconciliation job with a streaming pipeline. The discovery plan pre-registers a 3-engineer-week timebox, the hypothesis "reconciliation lag over 4 hours drives more than 2% of merchant support tickets", an evidence threshold of "at least 5 of 12 sampled merchants show lag-driven tickets", and the decision it enables: fund the migration or keep batch. At the end of week 2, 8 of 12 merchants have been sampled and 3 show lag-driven tickets. What should the team do with the remaining engineer-week?