Plain-language analytics literacy
Dashboards should make decisions easier, not harder to reach.
This is a reading room, not a sales floor. We explain what a KPI actually is, how a simple dashboard gets built in tools you already have, and how to ask the right question before a number changes your mind.
What a KPI Really Is
Not every number on a screen deserves the title.
Build One Dashboard
Google Sheets or Looker Studio, kept to one screen.
Read a Chart Properly
Spot a truncated axis before it fools you.
Questions Worth Asking
Before you repeat a number in a meeting.
Ready-Made Structures
Layouts for weekly and monthly reporting.
The Three-KPI Problem
Thirty metrics tell you everything and nothing at once.
A KPI, a key performance indicator, is a number tied directly to a goal the business has actually agreed on. Not a number that happens to be easy to pull from a report. Not a number that looks impressive in a slide. If nobody can say which decision the number changes, it is not a KPI. It is just data with good posture.
Most dashboards fail managers not because they contain too little information, but because they contain too much of it, arranged with no sense of priority. When thirty tiles compete for attention, none of them win. The eye settles on whichever one is red that week, regardless of whether it matters.
A rough filter for choosing three
Ask whether the number would change a decision within the next month. Ask whether someone in the room owns the outcome behind it. Ask whether it moves for reasons you can explain without guessing. Ask whether tracking it weekly would actually change how the team works, or just how anxious it feels.
Why non-technical doesn't mean not capable.
Nobody hands a manager a course in data modelling before promoting them into a role that requires reading one. That gap gets filled, usually badly, by whoever built the dashboard deciding what matters on your behalf. Reading a chart critically is a learnable skill, not a personality trait some people are born with and others aren't.
The manager who asks "what's the sample size behind this average" in a Tuesday meeting isn't being difficult. They're doing the one part of the job a spreadsheet cannot do for them: deciding whether a number deserves trust before it gets acted on.
A Practical Path
Three steps, in the order they actually happen.
Choose
Pick the three numbers tied to a decision someone will actually make this quarter. Write down why each one was chosen, in one sentence. If you can't, it's not ready yet.
Build
Lay those three numbers out on a single screen, in Google Sheets or Looker Studio. No scrolling required to see all three. One clear trend line each, labelled in plain words.
Question
Before repeating any number, check the axis, check the date range, check who calculated it and how. A dashboard is a starting point for a conversation, not the end of one.
Choosing a Format
Three common ways managers keep an eye on their numbers.
None of these require buying anything. This is a comparison of approach, not a product recommendation.
| Approach | Setup Time | Ongoing Effort | Best Suited For | Learning Curve |
|---|---|---|---|---|
| Manual Check-InsEmail or verbal updates | None | High, repeated weekly | Very small teams, early stage | None |
| Shared Google SheetCommon starting point Often the easiest first step | A few hours | Low, mostly automatic | Teams under 30 people | Gentle, basic formulas |
| Looker Studio DashboardConnected data sources | A day or two | Low once connected | Multiple data sources, wider audience | Moderate, one-time setup |
Either spreadsheet or Looker Studio can hold a genuinely useful dashboard. The difference is rarely the tool. It's whether someone limited themselves to the three numbers that matter, or gave in to the temptation to add "just one more tile" until the page needed scrolling.
Reading Charts Honestly
The same data can look calm or alarming, depending on the axis.
A line that climbs steeply on one chart and barely moves on another might be plotting identical figures. The difference is often where the y-axis starts. Truncating it at 80 instead of 0 turns a two percent shift into what looks like a dramatic surge.
Watch for four habits before accepting a chart's story at face value. Check whether the vertical axis starts at zero or somewhere flattering. Check whether two lines share one axis or quietly use two different scales. Check whether the date range was chosen to include a good month and exclude a bad one. Check whether a percentage change is being shown without the underlying total, since 50% growth on four customers is not the same claim as 50% growth on four thousand.
Before You Trust a Number
Five questions worth asking in any meeting where a report gets waved around.
What period does this cover?
A week, a quarter and a rolling twelve months tell very different stories from the same dataset.
Where did it come from?
A number pulled automatically from a system behaves differently to one typed manually into a slide.
Compared to what?
Growth only means something next to a baseline. Ask what the comparison point actually is.
Who is included, and who isn't?
An average can hide a small group doing very well and a large group doing poorly.
Would it survive being recalculated?
If someone else pulled the same figure from scratch, would they land on the same result?
Been handed a report you're not sure you should trust yet?
This site won't sell you a platform or a workshop. It exists to make the next report you read a little harder to misinterpret, and a little easier to question with confidence.
Send a Question