Where MI most often goes wrong

Leadership · ~6 minute read

Data, not improvement

Most MI packs in contact centres produce data and not improvement. The metrics are technically correct, the layout is reasonable, the cadence is regular — and the operation doesn’t change in response. Senior leaders skim the pack, planners produce it dutifully, and the same trends repeat themselves quarter after quarter. The function exists. The function doesn’t do anything.

The diagnosis is recognisable across operations. MI fails for six predictable reasons. This series walks through each, with one article per failure mode and a focus on the discipline that catches it rather than the technology that produces it.

Six places MI can go wrong 1. Designed for producer Not for the audience 2. Downstream metric Tracking the outcome, not the cause 3. All lagging Confirms, doesn’t predict 4. Nobody acts on it Fails the “so what?” test 5. Pack aged out Reports an operation that no longer exists 6. Composite hides truth Headline conceals the components None of the six requires a new tool. All six require discipline the producer rarely has time for.
The diagnostic landscape. Most operations have at least three of the six; many have all six simultaneously.

The six failure modes

1. Designed for the producer, not the audience. The pack the analyst enjoyed building and the operations director can’t use. Audience-led design, not analyst-led. See designed for the producer, not the audience.

2. The metric you track is downstream of the metric that matters. SL, AHT, CSAT — each is a real outcome and each is downstream of behaviours, decisions, and disciplines that the MI rarely captures. See the metric you track is downstream of the metric that matters.

3. All lagging, no leading. The pack that confirms last week’s outcomes and predicts none of next week’s. See all lagging, no leading.

4. The MI nobody acts on. Metrics that have been reported for years and never produced a decision. See the MI nobody acts on.

5. The MI pack that aged out. The pack that was right for last year’s operation and isn’t for this year’s. See the MI pack that aged out.

6. Composite metrics that hide the truth. The 87% composite quality score that hides accuracy collapsing while compliance compensates. Treated extensively in the existing composite metrics article; the diagnostic version is the natural endnote of this series.

Why this matters now

MI used to be expensive to produce. Spreadsheets needed building, data needed extracting, charts needed formatting. The constraint forced discipline because the producer had to be selective. Modern MI tooling has removed that constraint — dashboards build themselves, metrics multiply because they’re cheap, and packs balloon to 40 pages of slides the audience doesn’t read.

The new constraint is attention. The operations director has ten minutes for the pack. The CFO has five. The MI function that doesn’t respect this attention budget produces packs that are technically impressive and operationally invisible. The disciplines in this series are mostly about reclaiming that attention.

The series ahead

Each piece treats one failure mode. As with the forecasting, scheduling, and real-time series, the fixes are unglamorous — audience-led design, leading indicators, the “so what?” audit, the quarterly review. None requires a new platform. All require sustained editorial discipline that most MI functions don’t protect time for. The operations that do find their MI pack becomes one of the function’s most useful outputs; the operations that don’t produce packs that everyone receives and nobody reads.

Next in the series: Designed for the producer, not the audience.

Pair this with designing meaningful MI, the one-page MI pack, and composite metrics hide truth.