A platform announces 100% growth while a larger competitor reports 10%. The first figure sounds decisive until you discover that one added 500 observed stores and the other added 10,000. Both percentages can be correct. They describe movement relative to different starting points.
EcomToolkit recommends reading a platform growth claim as a small reconciliation exercise. Identify the measured population, recover the endpoints, calculate the absolute change and inspect the denominator. This guide uses invented datasets to explain the arithmetic; it does not present current vendor market shares or investment forecasts.
Table of Contents
- Recover the starting and ending counts
- Keep market share separate from platform growth
- Use CAGR for endpoint comparisons across years
- Do not average annual percentages blindly
- Check whether the series changed its meaning
- Turn a growth claim into a useful buying question
- The EcomToolkit view
Recover the starting and ending counts
The basic percentage change is the ending value minus the starting value, divided by the starting value, multiplied by 100. Always show the absolute change beside it. A large percentage can reflect a small starting base rather than a large addition to the market.
In an illustrative one-year comparison, Platform A grows from 500 to 1,000 observed stores. Platform B grows from 100,000 to 110,000. A doubles while B grows 10%, but B adds twenty times as many observed stores. Neither calculation alone tells a merchant which platform is suitable.
| Hypothetical platform | Starting observed stores | Ending observed stores | Absolute change | Relative growth |
|---|---|---|---|---|
| A | 500 | 1,000 | 500 | 100% |
| B | 100,000 | 110,000 | 10,000 | 10% |
| C | 20,000 | 19,000 | Minus 1,000 | Minus 5% |
| D | 0 | 100 | 100 | Undefined from a zero base |
Do not replace the undefined percentage for a zero starting value with 100%. There was no positive baseline against which to calculate that ratio. Report the new count and explain that the series began at zero or that the earlier observation is unavailable.
Check the time span as carefully as the arithmetic. Growth over a quarter, a year and three years cannot be ranked directly without addressing duration. A headline that omits dates can make an ordinary cumulative change look like exceptional annual momentum.
Keep market share separate from platform growth
A platform’s observed count can rise while its share falls. Suppose its count increases from 1,000 to 2,000 while the measured market expands from 10,000 to 25,000. The platform doubles, but its share falls from 10% to 8% because the denominator grows faster.
The share change is minus two percentage points. Relative to the original share, it is a 20% decline. Those are different units, so write them explicitly. Calling the movement simply minus two percent would understate the relative change and leave readers unsure which calculation was intended.
Growth in the denominator also needs interpretation. It may represent more active commerce sites, a broader crawl, a revised eligibility rule or improved detection. Before treating the share movement as a market event, confirm that both observations refer to the same kind of population.
W3Techs’ methodology explains how its technology surveys define websites and detect technology usage. Other providers can use different populations and units. The lesson is to read the provider’s method, not to transplant one provider’s denominator into another provider’s growth calculation.
Use CAGR for endpoint comparisons across years
Compound annual growth rate expresses the constant annual rate that would connect a positive starting value to a positive ending value over a stated number of years. Calculate it as the ending value divided by the starting value, raised to one divided by the number of years, minus one.
For an invented series rising from 10,000 to 13,310 over three years, the ratio is 1.331. Its cube root is 1.10, giving a CAGR of 10%. Count elapsed intervals: four annual observations covering the beginning and end of three years do not represent four years of growth.
The BEA explanation of average annual growth uses the compound-interest approach and makes the period frequency explicit. When applying this arithmetic to platform statistics, preserve the observation dates and measurement definition instead of treating every series as equally comparable.
CAGR is a summary of endpoints, not a description of each intervening year. A platform could grow steadily or rise sharply and then retreat to the same endpoint. Inspect the intermediate observations before describing the trend as stable, accelerating or resilient.
Do not average annual percentages blindly
Imagine a count rising from 100 to 150 and then returning to 100. The annual changes are plus 50% and minus 33.33%. Their arithmetic average is about plus 8.33%, yet the final count equals the starting count and the two-year CAGR is zero.
The second percentage applies to a different base. That is why averaging the rates does not reconstruct the overall compounded movement. For a known chain of annual growth factors, multiply the factors to recover cumulative change rather than adding their percentages.
| Hypothetical path | First-year change | Second-year change | Two-year endpoint change | Two-year CAGR |
|---|---|---|---|---|
| 100 to 150 to 100 | Plus 50% | Minus 33.33% | 0% | 0% |
| 100 to 110 to 121 | Plus 10% | Plus 10% | Plus 21% | Plus 10% |
| 100 to 80 to 100 | Minus 20% | Plus 25% | 0% | 0% |
| 100 to 120 to 144 | Plus 20% | Plus 20% | Plus 44% | Plus 20% |
Use unrounded values for calculations and round only the presentation. Reconstructing a long chain from rounded published percentages can create small discrepancies that look like data errors. Where possible, return to the underlying counts and retain a reproducible calculation sheet.
Avoid annualising a very short observation window without a compelling reason. The resulting rate assumes a pace that may not persist. A short burst caused by a launch, promotion or detection change is especially unsuitable for presentation as an established annual growth trajectory.
Check whether the series changed its meaning
Before comparing endpoints, verify active versus historical status, domain versus merchant counting, country coverage and treatment of multiple technologies on one site. A switch from websites to merchant accounts can change the count even when the underlying businesses have not changed at all.
Record retrieval dates separately from observation dates. A page fetched today may describe an earlier month, and a historical series may have been revised since the previous report. Keep the original extract or provider snapshot so later reconciliation does not depend on a changing web page.
When a provider documents a methodology break, do not draw an uninterrupted growth line across it without explanation. Use comparable segments, request a restated series, or label the break. A precise compound rate across incompatible populations is mathematically calculable and commercially misleading.
Our guide to store-count and GMV-weighted market share explains another distinction: counting stores and weighting economic activity answer different questions. A platform can grow in store count while its mix shifts towards smaller merchants.
Turn a growth claim into a useful buying question
Ask what the statistic changes about your decision. Growth may suggest a larger ecosystem to investigate, but it does not prove that a required integration exists, support will be responsive, or implementation costs will fall. Translate the headline into questions that can be checked against your requirements.
Separate observed results from forecasts in the same report. A projected market CAGR is conditional on a model and its assumptions. It is not a measured platform adoption rate, and it should not share an unlabeled chart line with historical observations. Cite the forecast period and publisher if you use one.
Keep churn and gross additions separate when available. A net increase of 1,000 could reflect 1,100 additions and 100 departures, or 6,000 additions and 5,000 departures. The endpoint growth rate is identical, while the operational interpretation may differ considerably.
See our historical usage and migration guide for the limits of inferring churn from technology observations. Combine that evidence with an actual implementation trial, integration review and support assessment before committing to a platform.
The EcomToolkit view
A percentage becomes useful when its base, dates and population are visible. Show counts beside growth rates, percentage points beside share changes, and intermediate observations beside CAGR. Treat a large number as the start of due diligence, not the end of it.
Bring a vendor growth claim to an EcomToolkit platform assessment. We can help turn the headline into a comparison grounded in your catalogue, workflows and decision criteria.
For the next analytical step, read our guide to choosing linear and logarithmic axes for platform growth charts.
