Practical framework / 5 MIN READ
Use a log scale only when the comparison needs ratios
Make multiplicative change readable without hiding absolute size.
AI-assisted research and analysis for the SNS Creative Fieldbook
Sources checked 2026-11-06
A logarithmic scale changes what equal visual distances mean. It can help compare proportional movement across very different magnitudes, but it can also make a large absolute gap look deceptively modest if the reader expects a linear axis.
What you will leave with: Choose and explain a scale using the decision question, valid values and a companion absolute-size view.
Use this when
- A few large campaigns make smaller series unreadable.
- A chart uses a log scale without explaining how to read it.
This lesson does not recommend logarithmic scales for every skewed dataset or authorize removing inconvenient zero values.
Choose between amounts and ratios
Ask what the reader needs to compare. If the decision concerns total cash required, absolute differences matter. If it concerns proportional growth across differently sized markets, ratios may be useful. Write that purpose before choosing the scale.
The ONS advises preferring intuitive linear scales and limiting non-linear scales to cases where they are needed for clarity. First consider small multiples, a focused range or separate views. A sophisticated axis is worthwhile only if it makes the intended comparison more understandable to the actual audience.
Explain equal distance with a small example
On a base-ten logarithmic axis, 10, 100 and 1,000 are equally spaced because each step multiplies by ten. On a linear axis, equal distances represent equal additions. That difference changes how readers interpret slopes and gaps.
Use explicit tick labels in ordinary units and label the scale. If the audience needs help, explain one interval directly on the chart. Avoid combining log axes with bars whose length invites a zero-based magnitude reading. A line or point display with clear context is often easier to interpret for a ratio-focused question.
Handle zeros and negative values openly
An ordinary logarithm is undefined for zero and negative values. Do not silently drop zero-spend campaigns or add an arbitrary constant solely to make the software draw them. Such changes alter the comparison and can hide important business states.
Choose another display or separate those cases with a visible explanation. If a specialized transformation is genuinely appropriate, describe it and obtain statistical review where needed. Keep a table of actual values available so the visual transformation does not become the only record of the evidence.
Preserve the business consequence
Show absolute amounts alongside proportional movement when both matter. A small campaign can double while contributing far fewer additional customers than a large campaign growing slowly. The log view should help identify patterns, not erase the scale of resource commitments.
Test the chart with someone unfamiliar with the transformation. Ask what an equal vertical gap means and which market added the most in absolute terms. If they read it incorrectly, simplify the display or add a companion view. State that similar slopes do not establish similar marketing quality or a causal relationship.
Make the reasoning concrete.
A hypothetical market rises from 100 to 200 orders while another rises from 10,000 to 20,000. Both double and have equal vertical movement on a log axis, but the additions are 100 and 10,000 orders respectively.
Use only the log chart
Opening: Describe the growth as equivalent business impact.
Hypothesis: The proportional change is equal.
Risk: The absolute additions differ by a factor of 100.
Pair ratio and amount views
Opening: Show proportional growth with an explicit scale and an absolute-change table.
Hypothesis: Readers can compare both pace and contribution.
Risk: Different market economics still require separate interpretation.
Use the log display only if proportional comparison is the intended task, and retain the absolute values needed for budget decisions. Do not label equal slopes as equal commercial value.
Take the questions into your next brief.
Use these prompts in your own notes. Nothing is submitted from this page.
Question
Does the decision depend on absolute change or proportional change?
Alternatives
Would small multiples or separate views be clearer?
Domain
Are there zero or negative values?
Labels
Can readers explain equal spacing correctly?
Context
Where are absolute amounts and limitations visible?
Watch for the wrong conclusion.
- Leaving the logarithmic scale unlabeled.
- Dropping zero values without disclosure.
- Adding arbitrary constants to force a plot.
- Equating equal proportional growth with equal business contribution.
Your next action
Redraw one wide-range chart in linear and log forms, then choose the version that readers interpret correctly for the decision.
Sources & scope
Frameworks, worksheets and hypothetical examples are original editorial synthesis. Linked primary sources support only the stated concepts. Source checking is not an independent audit of effectiveness. No named human author or reviewer is claimed.
- ONS: Axes and gridlines, non-linear scales →Primary statistical-agency visualisation guidance · Supports: Advises preferring intuitive linear scales and using non-linear scales only when necessary for clarity. The numerical log-scale examples and marketing decisions are original illustrations.