A KPI dashboard people actually use

The useful question is not "which KPIs exist" but "which decision rests on this number". Metrics no action hangs on create reporting work and nothing else.

It helps to separate outcome metrics from leading indicators. Outcome metrics — revenue, volume, market share, attrition — report what happened. Leading indicators — awareness, likelihood to recommend, satisfaction, purchase intent — show what is building. A dashboard of outcome metrics alone is always too late to steer by.

• Sales: revenue and volume by region and product, pipeline, win rate
• Marketing: aided and unaided awareness, ad recall, purchase intent, campaign effect
• Market research: market share, NPS, top-box scores, brand image across waves
• Product and service: customer satisfaction, complaint rate, repurchase rate

The least glamorous and most effective step: a binding formula for every core metric, written down and held in one place. Without it, every meeting discusses the number instead of the decision.

In DataLion metrics are defined as calculated KPIs via a SQL expression in the codebook. They live at the data level rather than inside an individual chart — new waves compute with the same formula, and nobody has to update a calculation in twelve views.

• Indices, scores and NPS defined once, identical everywhere
• Top-box, top-2 and top-3 are computed automatically for scale questions
• Formulas stay consistent across waves and data deliveries

A KPI dashboard is only as current as its supply. The connection, not the visualisation, decides whether it still holds up in six months.

DataLion takes Excel and CSV, databases, a REST API, and SPSS files including value labels and weighting variables. Several sources sit side by side in one project, joined on shared keys — time, region, product.

• Excel and CSV import, repeatable per wave
• Database connection and REST API for automatic refresh
• SPSS import with labels, weights and missing values
• Multiple data sources combined in one dashboard

This is where a market-research KPI dashboard parts ways with a BI dashboard. ERP figures are counted; survey figures are estimated — and estimates have a base, a weighting and an uncertainty.

Ten percent of brand-aware respondents and ten percent of the population are two entirely different statements. When a filter changes the base, weighting, significance and minimum base size have to travel with it. Classic BI tools have no place for these concepts: they can filter rows, but they do not know a weighting variable exists or that a percentage has a reference base.

• Weighting that matches the filtered subset
• Significance tests that compute on the visible base size
• Minimum base sizes that trigger before a cell gets too thin
• Wave logic: questions come and go while the base stays visible

A dashboard is an argument in a fixed order, for one audience. A board wants a handful of headline figures; researchers want filters and detail. Two views almost always beat one crowded one.

Building is drag-and-drop: pick a chart type, assign variables, set filters. Over 50 chart types are available, and anyone needing more detail gets a drill-down rather than a second page of charts.

• Separate views per audience, from the same data
• Interactive filters on time, region, product or segment
• Corporate design with your own colours, fonts and logo
• Roles, permissions and access profiles control who sees which rows

Dashboards are not meant to replace the deck — they are meant to produce it. In DataLion reports go out as native PowerPoint with editable charts in your own template, plus Excel report books and PDF.

Editable means the charts are real PowerPoint objects, not images. Anyone who needs to change a colour or a label mid-meeting can, without going back into the tool.