A platform-growth chart makes a small vendor look almost flat beside a much larger competitor. Another chart using the same data gives both lines the same upward slope. Neither picture is necessarily wrong. The vertical axis determines whether equal visual distances represent equal additions or equal proportional changes.
EcomToolkit recommends stating the comparison question before choosing a chart scale. This guide shows how linear and logarithmic axes change the reading of ecommerce platform statistics. All platform names and counts in the examples are fictional. They illustrate chart interpretation rather than current adoption, market share or a vendor recommendation.
Table of Contents
- Ask whether size or growth is the decision
- Work through the same counts on both scales
- Read slope only with the time axis in view
- Treat zero and missing observations honestly
- Make the chart readable without a tooltip
- Audit a vendor chart before using it
- The EcomToolkit view
Ask whether size or growth is the decision
A buying team might want to understand the size of an ecosystem because it needs local implementation partners and experienced staff. Another team might want to understand how quickly adoption is changing. Absolute counts and proportional growth are relevant to different questions, and neither establishes platform suitability on its own.
On a linear axis, moving from 100 to 200 uses the same vertical distance as moving from 1,000 to 1,100. Both changes add 100. On a logarithmic axis, moving from 100 to 200 uses the same distance as moving from 1,000 to 2,000. Both changes double the value.
Datawrapper’s line-chart guidance describes the choice between linear and logarithmic scales. The charting option should follow the analytical purpose. Changing a scale to make a preferred vendor look more impressive reverses that relationship and makes the report harder to trust.
Keep the underlying measure explicit. Detected websites, active merchants, installed stores and revenue are different quantities. An axis transformation does not make unlike definitions comparable. Establish the unit, observation dates and source methodology before deciding how to display the series.
Work through the same counts on both scales
Suppose Cedar grows from 100 to 200 to 400 observed stores across three equally spaced annual observations. Elm grows from 10,000 to 20,000 to 40,000. Both double at each step, but Elm adds far more stores in absolute terms.
| Fictional platform | First year | Second year | Third year | Total addition |
|---|---|---|---|---|
| Cedar | 100 | 200 | 400 | 300 |
| Elm | 10,000 | 20,000 | 40,000 | 30,000 |
On a common linear axis, Cedar’s line sits close to the bottom because its absolute values are much smaller. On a common logarithmic axis, the lines have equal slopes over equal time intervals. Their vertical separation remains meaningful: Elm is one hundred times Cedar at every observation.
With a base-ten logarithmic axis, the first Cedar point has coordinate 2 and the first Elm point has coordinate 4. The same two-unit gap remains at the later observations. Those are transformed coordinates used to position marks; the chart should normally label the axis with the original counts for readers.
The example supports two simultaneous statements: the platforms have equal proportional growth, and their observed scale is very different. It does not establish that they attract equally valuable merchants, retain customers equally well or offer equivalent support capacity. Those conclusions require different evidence.
Read slope only with the time axis in view
Equal slopes on a logarithmic vertical axis indicate equal proportional change per equal horizontal interval. If one observation covers a month and another covers a year, visually similar line segments do not represent the same growth rate unless time spacing is handled correctly.
A straight rising line on a semi-log chart is consistent with constant proportional growth over equal periods. It is not evidence that growth will continue. Few observations, methodology changes or selective date ranges can create a persuasive-looking line without establishing a stable underlying process.
Use our platform CAGR and small-base guide for the endpoint arithmetic behind growth claims. The chart adds a different layer: how visual spacing affects what readers notice. A correct percentage can still be presented in a way that conceals its small starting base.
If the purpose is to compare change from a common starting point, an indexed chart may also help. Setting both first observations to 100 removes the original size difference, so show original counts alongside it. Indexing and logarithmic scaling are separate transformations; disclose both if they are combined.
Treat zero and missing observations honestly
An ordinary logarithmic scale requires positive values. A recorded zero cannot simply be placed on that axis, and a negative net change is not a valid positive count. First establish whether the value means no observed stores, an unavailable observation or a measure that can legitimately be negative.
Do not silently add one to every observation to make the chart render. That changes the ratios, especially for small values. Moving from zero to one becomes a different mathematical story after adding a constant, and the chosen constant becomes part of the result rather than a harmless display setting.
For a series containing real zeros, consider a linear chart, a separate panel or an explicitly described transformation designed for that purpose. For missing observations, leave a visible gap or annotate the absence. Neither a zero nor an invented interpolated value should stand in for unknown platform coverage.
Keep closed or undetected stores in the source methodology. If a provider changes its detection rules, a sudden jump may reflect measurement rather than adoption. Annotate the break and avoid drawing an uninterrupted growth narrative across incompatible definitions simply because the charting software connects the points.
Make the chart readable without a tooltip
Label the scale directly, using wording such as logarithmic vertical axis alongside the chart. Choose meaningful tick labels and show the unit. Readers should not need to infer the transformation from an unusual sequence of numbers or discover it only by opening a settings menu.
Annotate the start and end counts for important series. Where the comparison is proportional, explain that equal vertical distances represent equal ratios. Where the decision is about ecosystem size, include a linear view or a table of absolute counts so that small and large platforms remain distinguishable.
| Chart choice | Best question | Main interpretation risk |
|---|---|---|
| Common linear axis | How different are absolute counts? | Small-series changes become hard to see |
| Common logarithmic axis | How do proportional changes compare? | Readers assume distances mean additions |
| Indexed series | How much has each changed from its start? | Starting scale disappears |
| Separate panels with different ranges | What is each series doing locally? | Slopes appear comparable when they are not |
| Data table beside the chart | What are the exact observations? | Trend shape is less immediate |
Avoid using a logarithmic axis for bar lengths that imply a meaningful zero baseline. A line or dot chart is usually easier to explain for positive values spanning large ranges. Whatever the format, provide the underlying table so readers can verify the comparison without relying on visual estimation.
Audit a vendor chart before using it
Ask for the source counts, observation dates and population definition. Recreate the chart with a common linear axis and a common logarithmic axis, then compare the statements each supports. If the recommendation changes solely because the picture looks steeper, the decision criteria need clarification.
Check for truncated date ranges, selectively omitted competitors and inconsistent measurement coverage. A visual transformation cannot correct those problems. Also inspect whether the report confuses growth in observed stores with growth in market share; the latter requires a denominator that may itself be changing.
For benchmark summaries involving ratios rather than visual axes, see our geometric-mean speed comparison guide. Taking logarithms in an analytical calculation and displaying a logarithmic chart are related mathematical ideas, but they solve different reporting problems and need separate explanations.
The EcomToolkit view
A useful platform-growth chart makes its comparison rule visible. Show absolute scale when capacity and ecosystem size matter, proportional change when momentum matters, and both when the decision needs both. Keep zeros, missing values and methodology breaks visible rather than smoothing them into a persuasive story.
Bring a vendor growth chart to an EcomToolkit platform review. The practical next step is to recover the counts and connect them to the operating questions your store needs answered before selecting a platform.
