The Zero Baseline Rule
manipulated axis = manipulated perception (users aren't stupid)
Most people start y-axis at random numbers. Looks dramatic. Feels dishonest.
A 2% change looks like 200%.
Users spot it. Trust dies.
A tiny story
A CEO shared his growth chart in a meeting. Y-axis started at 95.
A 3% increase looked like a cliff jump.
One analyst said: "This feels misleading."
Meeting derailed.
The real issue
Non-zero start. Zoomed scale. Exaggerated peaks. Fake urgency.
Makes tiny movements look massive.
Data integrity questioned.
What you need now
Start at zero for bar charts. Always. No exceptions.
Start at zero for area charts. Shows true proportion.
Line charts get flexibility. But label it clearly if truncated.
Use annotations when context matters. "Y-axis starts at 50."
Show full range first. Zoom in second. Give both views.
Honesty = lasting credibility.
Two spicy takes
Hot take 1: Truncated y-axis is the fastest way to look like you're hiding something.
Hot take 2: If your metric needs a zoomed axis to look good, your metric sucks.
3 actions this week
Check every bar chart. Reset y-axis to zero.
Add a note if you must truncate for valid reasons.
Show the full picture before zooming into details.
Instant integrity.
Closing
When you show truth, you earn trust.
If you want help, reply to this mail or fill this form.
