Business Analytics Quiz: Dashboard And Kpi Design
10 questions · exam conditions
0:00
Dashboard And Kpi DesignQuestion 1 of 10

A logistics dashboard reports late-delivery performance for two regions. The North region had 800800 deliveries, of which 6464 were late. The South region had 200200 deliveries, of which 3636 were late.

Which companywide KPI and dashboard treatment are most appropriate?

Report a 10%10\% late-delivery rate based on all deliveries, and provide regional rates as diagnostic drilldowns.
Report a 13%13\% late-delivery rate by averaging the two regional percentages, and display both regions equally.
Report an 8%8\% late-delivery rate because the larger region is most representative, and flag only major deviations.
Report an 18%18\% late-delivery rate because the worst-performing region determines companywide service risk.
← Back to quizzes

Business Analytics Quiz

Business Analytics Quiz: Dashboard And Kpi Design

Practice Dashboard And Kpi Design in Business Analytics with focused quiz questions that help you check what you know, review explanations, and build confidence with test-style prompts.

What this quiz covers

This quiz focuses on Dashboard And Kpi Design, giving you a quick way to practice the rules, question types, and explanations that matter most for Business Analytics.

How to use this quiz

Try each quiz question before looking at the correct answer. Use the explanations to review missed ideas, then come back to similar questions until the pattern feels familiar.

All questions

Question 1

A logistics dashboard reports late-delivery performance for two regions. The North region had 800800 deliveries, of which 6464 were late. The South region had 200200 deliveries, of which 3636 were late.

Which companywide KPI and dashboard treatment are most appropriate?

  1. Report a 10%10\% late-delivery rate based on all deliveries, and provide regional rates as diagnostic drilldowns. (correct answer)
  2. Report a 13%13\% late-delivery rate by averaging the two regional percentages, and display both regions equally.
  3. Report an 8%8\% late-delivery rate because the larger region is most representative, and flag only major deviations.
  4. Report an 18%18\% late-delivery rate because the worst-performing region determines companywide service risk.
Explanation: When building a companywide KPI from regional data, the critical decision is how to aggregate: should you average percentages, or calculate from raw totals? This question tests exactly that, along with sound dashboard design principles. The right approach is to calculate the overall rate from all deliveries combined. Total late deliveries: 64+36=10064 + 36 = 100. Total deliveries: 800+200=1,000800 + 200 = 1{,}000. That gives 1001,000=10%\frac{100}{1{,}000} = 10\% — a weighted aggregate that respects each region's actual volume. Pairing this headline KPI with regional drilldowns lets executives see the company's true performance while enabling analysts to diagnose where problems live. That's exactly what answer A describes, making it correct. Answer B is the classic unweighted average trap: averaging 8%8\% (North) and 18%18\% (South) gives 13%13\%, but this treats a 200-delivery region as equally important as an 800-delivery region — a statistically distorted picture. Answer C cherry-picks only the larger region's rate (64800=8%\frac{64}{800} = 8\%) and discards real data from South entirely; size alone doesn't determine representativeness when you have complete data available. Answer D takes the worst region's rate (36200=18%\frac{36}{200} = 18\%) as the companywide figure, conflating a risk-management concern with a factual measurement — these are different things. Study tip: Whenever a question involves combining rates or percentages across groups of different sizes, always go back to raw counts. Averaging percentages without weighting by group size is one of the most common analytical errors tested on business-analytics exams.

Question 2

A seasonal retailer records December revenue of 12.012.0 million. November revenue was 10.010.0 million, December revenue last year was 13.013.0 million, and this December's approved target was 12.512.5 million.

Which dashboard status statement provides the most decision-useful assessment?

  1. Revenue is improving because it increased 20.0%20.0\% from November, so the December KPI should be marked on track.
  2. Revenue is nearly on target because it reached 96.0%96.0\% of plan, so historical comparisons can be omitted.
  3. Revenue is underperforming because it declined 7.7%7.7\% year over year, so the monthly increase should be omitted.
  4. Revenue rose 20.0%20.0\% month over month but is 4.0%4.0\% below target and 7.7%7.7\% below last December. (correct answer)
Explanation: When evaluating dashboard KPIs, your goal is decision-usefulness — giving stakeholders every relevant benchmark so they can diagnose performance accurately. A single comparison can mislead; good business analytics combines multiple reference points: prior period, prior year, and approved target. Answer D is the strongest choice because it presents all three perspectives without suppressing any signal. The month-over-month gain is 12.0−10.010.0=20.0%\frac{12.0 - 10.0}{10.0} = 20.0\%, which looks positive. But the year-over-year change is 12.0−13.013.0≈−7.7%\frac{12.0 - 13.0}{13.0} \approx -7.7\%, signaling potential concern. And the variance to target is 12.0−12.512.5=−4.0%\frac{12.0 - 12.5}{12.5} = -4.0\%, confirming underperformance against plan. D reports all three honestly, letting decision-makers see the full picture. Answer A cherry-picks only the month-over-month comparison and incorrectly declares the KPI "on track" — but a seasonal MoM increase (October → November → December is naturally an up-ramp) doesn't mean target achievement. Answer B makes the opposite error of omission, focusing on the plan variance while dismissing year-over-year context. Declaring historical comparisons unnecessary because you're "close to target" removes a key diagnostic layer. Answer C cherry-picks in the other direction — leading with the YoY decline and omitting the MoM trend entirely. Selectively omitting data to frame a narrative, whether positive or negative, reduces analytical integrity. The study tip here: watch for answer choices that omit data while justifying the omission. On business analytics questions, any option that says a comparison "can be omitted" or "should be excluded" is almost always a trap — complete, multi-dimensional context is the standard for decision-useful reporting.

Question 3

A customer-support director is responsible for reducing subscription churn caused by poor service. Churn is measured quarterly, while first-response time, unresolved-ticket backlog, and ticket-reopen rate are available daily.

Which KPI hierarchy would best support both accountability and timely intervention?

  1. Use unresolved-ticket backlog as the outcome KPI, and treat quarterly churn as contextual information outside the main dashboard.
  2. Use first-response time as the outcome KPI, and combine churn and ticket-reopen rate into one equally weighted service score.
  3. Use quarterly churn as the outcome KPI, supported by response time, backlog, and reopen rate as leading diagnostic indicators. (correct answer)
  4. Use quarterly churn as the only headline KPI, because adding operational drivers may distract managers from the business objective.
Explanation: When building a KPI hierarchy, your goal is to connect a lagging outcome metric — one that measures ultimate business impact — with leading operational indicators that you can act on before the outcome deteriorates. This question tests whether you understand that distinction and can apply it to a real management scenario. Quarterly churn is the true outcome the director is accountable for: it directly measures the business harm she's trying to prevent. But because it's only visible every 90 days, acting on churn data alone means you're always reacting too late. The right solution, captured in C, is to keep quarterly churn as the headline outcome KPI — preserving clear accountability to the business objective — while treating first-response time, unresolved-ticket backlog, and ticket-reopen rate as daily leading indicators. These operational metrics let managers spot problems early and intervene before churn spikes appear. A is backwards: elevating backlog to outcome KPI demotes churn to "contextual," which breaks accountability to the actual business goal. Reducing backlog is a means, not an end. B creates a misleading composite by blending churn (a lagging outcome) and reopen rate (a leading driver) into a single score — this obscures what's actually causing what and makes accountability murky. D commits the opposite error of C: restricting the dashboard to churn alone removes the early-warning signals that make timely intervention possible. A manager watching only quarterly churn has no levers to pull. A useful rule of thumb: outcome KPIs tell you the score; leading indicators tell you how to play the game. You need both, in a clear hierarchy.

Question 4

A churn model scores 10,00010{,}000 customers each week, but the retention team can contact only 500500. Some high-risk customers are unlikely to respond to an offer, while some moderately high-risk customers have substantial retained margin and strong expected treatment response.

Which KPI view would best convert the predictive model into an actionable retention dashboard?

  1. Rank customers only by predicted churn probability, display the average score of the top 500500, and track contact completion.
  2. Rank customers by expected incremental retained margin net of contact cost, and track realized lift against a holdout group. (correct answer)
  3. Count all customers above a fixed churn threshold, display model accuracy, and allow managers to select contacts subjectively.
  4. Rank customers by current account margin, display total margin contacted, and use churn scores only as explanatory annotations.
Explanation: When a predictive model feeds a business decision with a resource constraint, the right KPI isn't just who is most at risk — it's who delivers the most value if we act. This question tests whether you understand the difference between model output (churn probability) and business outcome (retained margin lift). The winning framework here is expected incremental value: for each customer, multiply the probability they would churn by the margin you'd retain if the intervention works, then subtract the cost of contacting them. Ranking by this metric — exactly what B describes — ensures your 500 contacts go to customers where action actually moves the needle. Tracking realized lift against a holdout group closes the feedback loop, telling you whether the model is generating real business value over time. Choice A fails because raw churn probability ignores two critical dimensions: response likelihood and margin value. A high-probability churner who won't respond to any offer and has low margin is a wasted contact. Choice C compounds this by adding a fixed threshold (which discards rank-ordering entirely) and relying on model accuracy — a technical metric that doesn't translate to business decisions. Subjective manager selection also reintroduces the human bias the model was meant to remove. Choice D flips the error in the opposite direction: sorting by current account margin prioritizes your best customers, not your most actionable ones. A high-margin customer with near-zero churn risk doesn't need a retention call. Study tip: On business analytics questions, whenever you see a constrained resource (limited budget, limited contacts, limited time), look for the answer that maximizes expected incremental value, not raw scores or totals — that's the operationally correct framing.

Question 5

A prescriptive inventory system recommends expedited shipments to reduce stockouts while respecting a weekly premium-freight budget. The current dashboard highlights a 92%92\% recommendation-acceptance rate, but executives cannot tell whether accepted recommendations create value.

Which dashboard redesign would best evaluate and govern the prescriptive system?

  1. Retain acceptance rate as the primary KPI, adding the number of recommendations and average approval time as supporting measures.
  2. Use stockout rate as the sole primary KPI, because freight spending and recommendation overrides are implementation details.
  3. Use predicted savings as the primary KPI, and exclude rejected recommendations so the view reflects actions actually taken.
  4. Use realized net cost reduction versus a credible baseline, with budget adherence, overrides, and constraint breaches as diagnostics. (correct answer)
Explanation: When evaluating a prescriptive analytics system, you need to ask two distinct questions: Is the system producing accurate recommendations? and Is it actually creating business value? A 92% acceptance rate tells you people are clicking "approve" — it says nothing about whether those approvals reduced costs or improved outcomes. This question tests whether you can design governance metrics that close that accountability gap. The strongest dashboard redesign is D because it measures what ultimately matters: realized net cost reduction against a credible baseline. That baseline is critical — without it, you can't separate the system's contribution from market fluctuations or seasonal effects. The diagnostic layer (budget adherence, overrides, constraint breaches) then helps you understand why performance is high or low, creating a feedback loop that genuinely governs the system rather than just monitoring activity. A fails because it doubles down on process metrics — acceptance count and approval time — rather than outcomes. High acceptance rates with fast approvals could still mean millions in wasted freight spend. B is tempting because stockout rate is a real outcome, but it's incomplete. A system could eliminate stockouts by blowing through the premium-freight budget every week, which would appear successful under this single KPI while causing financial harm. C introduces a dangerous selection bias: excluding rejected recommendations hides cases where overrides may have outperformed the system's suggestions. You can't evaluate a model you're only showing in its best light. For exam questions involving analytics governance, watch for answer choices that confuse activity metrics (acceptance rate, volume) with outcome metrics (net cost reduction, stockout impact). Governance requires both outcomes and diagnostics — never just one.

Question 6

A software company defines February-end retention for the January activation cohort as the percentage of January activations that are active at the end of February. The cohort contains 1,0001{,}000 customers. Of these, 100100 canceled by the end of January, another 180180 canceled during February, and 5050 of the customers who had canceled in January reactivated before February ended.

Which KPI value and definition should appear on the cohort dashboard?

  1. Report 72%72\% retention because all cancellations through February remain losses for the original activation cohort.
  2. Report 82%82\% retention because February retention should exclude cancellations that occurred during January.
  3. Report 77%77\% retention because the denominator stays fixed and reactivated cohort members count as active at period end. (correct answer)
  4. Report 85.6%85.6\% retention because only customers active at the start of February belong in the retention denominator.
Explanation: Cohort retention questions test your ability to apply a consistent definition across a measurement window. The key principle: in cohort analysis, the denominator is fixed at the original cohort size, and you measure who remains active at the end of the period — regardless of when churns or reactivations occurred along the way. Here's how the math works for answer C. Start with 1,0001{,}000 customers. By period end, 100100 canceled in January, but 5050 of those reactivated before February ended, so net January losses are 5050. Another 180180 canceled during February. Total inactive at February end: 50+180=23050 + 180 = 230. Active customers: 1,000−230=7701{,}000 - 230 = 770. Retention: 770/1,000=77%770 / 1{,}000 = 77\%. This correctly keeps the denominator fixed and credits reactivations as active at period end. Answer A reports 72%72\% by ignoring the reactivations entirely — treating the 5050 returnees as permanent losses. This violates the rule that retention measures status at period end, not cumulative churn events. Answer B reports 82%82\% by excluding January cancellations from the loss count altogether, as if only February behavior matters. This misunderstands cohort logic: all 1,0001{,}000 original activations remain in the denominator, and all end-of-period departures count. Answer D reports 85.6%85.6\% by shrinking the denominator to only those active at February's start (1,000−100=9001{,}000 - 100 = 900), then dividing 770/900770/900. Adjusting the denominator mid-cohort is the classic trap — it converts cohort retention into a rolling retention metric, which answers a different question entirely. When you see cohort retention problems, lock the denominator at original cohort size and evaluate net active status at the period's close — reactivations restore customers to "active," and mid-period events only matter insofar as they affect that final snapshot.

Question 7

A retailer's desktop conversion rate increased from 8%8\% to 9%9\%, and its mobile conversion rate increased from 2%2\% to 3%3\%. However, desktop's share of traffic fell from 80%80\% to 20%20\% between the two periods, causing the overall conversion rate to decline.

Which dashboard design would best explain the apparent conflict without obscuring overall business performance?

  1. Display the overall conversion rate together with channel conversion rates and traffic mix, using a decomposition for the total change. (correct answer)
  2. Display only channel conversion rates because the overall decline is entirely an aggregation artifact and is not decision-relevant.
  3. Display only the overall conversion rate because segment views can encourage managers to disregard the company's actual result.
  4. Replace conversion rate with total conversions because count metrics are unaffected by changes in channel composition.
Explanation: Whenever you see a scenario where segment-level metrics move in one direction while the aggregate moves in another, you're dealing with Simpson's Paradox — a statistical phenomenon where changes in group composition distort the overall trend. The dashboard design question is really asking: how do you present data so decision-makers understand why the overall rate fell without losing sight of the big picture? Here's the core logic. Desktop converted better (8%→9%8\% \rightarrow 9\%) and mobile converted better (2%→3%2\% \rightarrow 3\%), yet the blended rate fell because desktop traffic collapsed from 80%80\% to 20%20\% of the mix. The overall rate shifted from (0.80×8%)+(0.20×2%)=6.8%(0.80 \times 8\%) + (0.20 \times 2\%) = 6.8\% to (0.20×9%)+(0.80×3%)=4.2%(0.20 \times 9\%) + (0.80 \times 3\%) = 4.2\%. A dashboard showing channel rates, traffic mix, and the overall rate — along with a decomposition of what drove the total change — gives managers both the diagnosis and the business reality. That's exactly what A provides. B is wrong because dismissing the overall rate as "just an artifact" ignores real business consequences — total performance still matters to leadership and stakeholders. C goes the opposite direction, hiding the segment detail that explains why the decline happened, leaving managers unable to act correctly. D is tempting but flawed: switching to raw conversion counts doesn't resolve the interpretive confusion and can introduce its own distortions if total traffic volume also changed. When you spot conflicting aggregate vs. segment trends on the exam, immediately think Simpson's Paradox and ask: does the dashboard show composition alongside rates? That pairing is almost always the right design choice.

Question 8

An online retailer receives order data every 1515 minutes. Refund data are loaded once each night at 22 a.m. and cover transactions only through the prior midnight. At noon, the dashboard currently calculates net sales by subtracting the latest available refunds from live order revenue.

Which redesign would most improve the interpretability of the net-sales view?

  1. Continue showing the mixed-latency net-sales value, but add a warning whenever the current day's order volume is unusually high.
  2. Show finalized net sales through the prior midnight and display today's gross order revenue separately as a provisional metric. (correct answer)
  3. Delay every dashboard metric until the next refund load so that all operational and financial measures share one refresh time.
  4. Show live gross order revenue as net sales and revise historical values after refunds arrive, without distinguishing provisional values.
Explanation: When a dashboard blends data of different freshness levels into a single number, you face a data latency mismatch — and the core challenge is making that mismatch visible rather than hiding it. Ask yourself: does the displayed metric accurately represent what it claims to represent, and does the user know its limitations? Here, "net sales" implies orders minus refunds, but the refund data is always at least 1212 hours stale by noon. Subtracting yesterday's refunds from today's live orders produces a figure that is neither truly current nor truly finalized — it's a misleading hybrid. The cleanest fix is to separate the two streams by their natural boundaries: show finalized net sales through prior midnight (where both orders and refunds are complete) alongside today's gross revenue as a clearly labeled provisional figure. This is exactly what B does — it preserves analytical integrity while still surfacing timely operational data, and it tells users precisely what each number means. A fails because adding a volume-based warning doesn't address the fundamental latency problem — the metric remains a mixed-signal number regardless of order volume. C sounds rigorous but sacrifices operational value entirely; delaying live order data until 2 a.m. makes the dashboard useless for intraday decisions. D is the most dangerous choice — silently labeling incomplete data as "net sales" and revising it later without flagging provisional values actively misleads users until the revision occurs. The broader strategy: when a question involves dashboard or reporting design, always ask whether users can tell how fresh each number is. Transparency about data currency is a core principle of interpretable analytics.

Question 9

In an A/B test, the control experience has a 10%10\% purchase conversion rate and an average order value of 8080 dollars. The treatment has a 9%9\% conversion rate and an average order value of 9292 dollars. The return rate is 5%5\% for control and 8%8\% for treatment. Statistical uncertainty has not yet been assessed.

Which experiment dashboard design would best support a launch decision?

  1. Use average order value as the primary KPI because it increased by 15%15\%, and show conversion rate as supplementary context.
  2. Use conversion rate as the primary KPI because it is closest to purchase behavior, and omit order value until the test ends.
  3. Use revenue per visitor with a confidence interval as the primary KPI, and monitor return rate as a prespecified guardrail. (correct answer)
  4. Use a composite score that equally weights conversion, order value, and return rate, and launch when the score becomes positive.
Explanation: When an A/B test produces mixed signals across multiple metrics, your job is to find a single, unified metric that captures the business outcome holistically — then protect it with guardrails for unintended side effects. Here, conversion rate fell (10%→9%10\% \rightarrow 9\%) but average order value rose ($80→$92\$80 \rightarrow \$92). Neither metric alone tells the full story. The right synthesis is revenue per visitor (RPV), calculated as conversion rate × average order value. For control: 0.10×$80=$8.000.10 \times \$80 = \$8.00. For treatment: 0.09×$92=$8.280.09 \times \$92 = \$8.28. RPV actually favors the treatment — but only if that lift is statistically reliable and not eroded by returns. That's exactly what C provides: RPV as the primary KPI with a confidence interval (addressing statistical uncertainty, which the passage flags as unassessed), plus return rate as a prespecified guardrail since it worsened from 5%5\% to 8%8\%. A is tempting because AOV increased by 15%15\%, but promoting a metric that ignores the conversion rate drop gives you an incomplete and misleading picture of revenue impact. B makes the opposite mistake — conversion rate alone misses the compensating AOV gain, and omitting order value until the test ends unnecessarily delays relevant information. D sounds rigorous but is actually arbitrary: equally weighting three metrics with different units and business significance lacks statistical grounding, and "launch when the score becomes positive" is not a valid significance threshold. The key pattern to memorize: when metrics move in opposite directions, combine them into one revenue-based metric and treat the risk metric as a guardrail, not an input to a composite score.

Question 10

A subscription company wants to increase profitable customer acquisition without reducing customer quality. Marketing can increase sign-ups by offering deeper discounts, but heavily discounted customers historically cancel more often and request more refunds.

Which KPI design would best align an executive acquisition dashboard with the company's objective?

  1. Use sign-up conversion rate as the headline KPI, with website visits and advertising impressions as guardrails.
  2. Use contribution margin per visitor as the headline KPI, with 9090-day churn and refund rates as guardrails. (correct answer)
  3. Use average initial order value as the headline KPI, with total acquisition spending and campaign reach as guardrails.
  4. Use total subscription revenue as the headline KPI, with click-through rate and cost per impression as guardrails.
Explanation: When designing an executive KPI dashboard, your goal is to ensure the headline metric directly captures the company's core objective — in this case, profitable acquisition, not just acquisition volume. A well-built dashboard pairs a headline KPI that reflects the true goal with guardrail metrics that catch unintended side effects. Contribution margin per visitor does exactly this. It measures how much profit each potential customer generates after accounting for discounts, costs, and behavior — so if aggressive discounting attracts low-quality customers, this number falls immediately. The guardrails of 90-day churn and refund rates then flag the specific quality problems the passage warns about. Together, they create a complete picture: are we acquiring customers profitably, and are those customers actually staying? That's why B is correct. A fails because sign-up conversion rate rewards volume, not value. An executive watching conversion rate could celebrate a discount-fueled spike while the business quietly loses money on every new subscriber. C uses average initial order value, which sounds financial but misses the point — a customer who pays a high first-month price and immediately churns creates no long-term profit. Neither acquisition spending nor campaign reach helps detect that problem. D is the weakest choice: total subscription revenue is a lagging, high-level metric, and click-through rate plus cost per impression are pure marketing-efficiency metrics that tell you nothing about customer quality or profitability. A useful pattern to remember: whenever a question involves a tension between growth and quality, the correct KPI design will measure profitability or value at the unit level — not top-line volume.