The Dashboard Nobody Opens
A logistics firm built a weekly operations dashboard for its 14 depot managers. It pulls 40 metrics from three systems: dispatch, warehouse, and finance. Six months after launch, usage logs show 9 of 14 managers open it less than once a month; 4 have never logged in. In interviews, managers say the numbers "don't match what we see" and that the finance figures arrive four days late. One manager keeps a personal spreadsheet that reconciles dispatch and warehouse counts by hand each Monday. The operations director wants to "add more drill-downs" to increase engagement. The dashboard project has consumed 300 analyst hours so far.
1. Evaluate the director's proposal and recommend an alternative approach with concrete first steps.
Read the worked answer
Adding drill-downs treats a relevance problem as a feature problem. The evidence points elsewhere:
- Trust failure. Managers say numbers don't match reality. Until the reconciliation logic is explained and validated with them, more views of untrusted numbers will not help.
- Timeliness failure. Finance data lags four days, so the weekly view is stale for the decisions managers make on Monday. A metric that arrives after the decision is made has no value regardless of its presentation.
- Shadow system. The hand-built spreadsheet is the real requirement in disguise. It tells you which metrics matter and how they are calculated.
Recommended approach: run a short requirements session with 3–4 managers, starting from the personal spreadsheet. Ask what decision each number supports and when it is needed. Then rebuild a smaller dashboard (8–12 metrics) with documented definitions, a visible data-refresh timestamp, and a reconciliation note. Defer drill-downs until usage rises. Measure success by weekly active use and by whether the manual spreadsheet is retired.
2. How should the team handle the finance data delay? Give at least two options with trade-offs.
Read the worked answer
The delay is a data-supply constraint, not a visualization issue, so it needs an explicit design decision.
Option A: Split the dashboard by cadence. Show dispatch and warehouse metrics daily or near-real-time, and show finance metrics in a separate monthly or lagged section clearly labeled with the data-as-of date. Trade-off: users see two views and must understand the difference, but each number is timely for its purpose. This is usually the best default.
Option B: Change the finance close process. Ask finance to produce a provisional weekly figure with a defined accuracy tolerance, later replaced by final numbers. Trade-off: requires finance effort and creates a risk of provisional numbers being treated as final; needs clear labeling and a reconciliation step.
Option C: Remove finance metrics from the weekly view and keep them in a monthly report. Trade-off: simplest, but managers lose cross-domain visibility.
Recommend A now and explore B with finance. Whatever the choice, stamp every metric with its refresh time and definition so users can judge fitness for their decision. Document the decision and the tolerance in the dashboard's data dictionary.