“What is the average e-commerce conversion rate?” is a question asked thousands of times a month, and it almost always leads to the wrong place. Someone else’s average mixes industries, baskets and business models, so it measures everyone and no one. This guide shows why benchmarks disappoint, what to replace them with (your own norm plus a few hard thresholds) and how such a table of norms turns performance reviews from comparing yourself with a legend into reading your own shop.
Demonstration
The same conversion rate, two different businesses
One number, two shops, zero shared conclusions:
Shop A · accessories, €15 basket
Shop B · furniture, €225 basket
The same value is an alarm signal in one shop and a source of pride in the other. A benchmark that knows nothing about the basket, industry and traffic says nothing; and published averages know none of these things.
Five variables
Five variables that invalidate the comparison
Industry and purchase cycle
A supplement is bought in minutes, furniture over weeks. Session conversion measures entirely different processes.
Basket value
The more expensive the basket, the lower the conversion rate by nature. Comparing without AOV is comparing apples with invoices.
Brand share
Traffic from branded queries converts several times better. A shop with a strong brand “wins” the benchmark without a better site.
Device and source mix
Mobile converts differently from desktop, social differently from search. A shop’s average is a function of the mix more than of quality.
Report methodology
Whose shops, which market, what session definition and period: usually unknown. You compare yourself with a sample you do not know.
The replacement
Your own norm: the only benchmark that knows your shop
Working definition
Your own norm is the range of typical values for a metric, derived from the shop’s history: the comparable period a year earlier and the average of the most recent full periods, taking season and weekday into account. The shop compares itself with itself at a comparable moment, because only this “competitor” has the identical industry, basket, brand and traffic mix.
A table of norms in practice (lingerie shop, September, illustrative data):
| Metric | My norm (range) | Current month | Reading |
|---|---|---|---|
| Mobile conversion rate | 1.1–1.4% | 1.3% | within norm |
| Desktop conversion rate | 2.2–2.6% | 1.8% | below: check checkout after the update |
| Average basket | €24–27 | €29.50 | above: check the mix (has a cheap bestseller dropped out?) |
| MER | 4.4–5.0 | 4.7 | within norm |
| Unpaid orders | 5–7% | 6.2% | within norm |
| GA4 coverage | 93–95% | 94% | within norm |
Notice how the nature of the conversation changes: instead of “is 1.8% good?” (a question with no answer), the question becomes “why has desktop fallen below its own range?” (a question with a procedure: decomposition and diagnosis from the guide on finding the causes of a sales drop). Deviations from your own norm are also the alert mechanism described in the guide on monitoring and anomalies.
Exceptions
Hard thresholds: external comparisons that do apply
A few external reference points make sense, because they follow from arithmetic that applies to everyone:
- Break-even ROAS and MER from your own margin: 100 / margin% is your shop’s mathematical limit. The full calculation is in the guide “What is a good ROAS”.
- Delivery cost relative to basket value: the carrier’s price list is objective; a basket whose margin cannot carry the delivery cost is a problem regardless of benchmarks (threshold calculation: guide to free delivery).
- Measurement completeness: GA4 coverage close to your own norm and Merchant Center product statuses with no disapprovals are the conditions of measurement hygiene.
- Market trend as directional background: if public data shows a growing market while the shop stands still year on year, that signals a loss of share; the direction is enough for this, with no need to compare specific metrics.
The most expensive use of a benchmark
Decisions forced by someone else’s average: cutting prices “because conversion is below average”, rebuilding the shop “because the basket is too small”. If your own norm is stable and the physical thresholds are met, such a move fixes a result nobody broke and breaks one that was working.
DataOrganizer
How DataOrganizer compares a shop with its own norm
Comparing with yourself requires one thing: the history of all metrics in one place. DataOrganizer keeps it from connected sources, so norm ranges, year-on-year and season-on-season comparisons, and anomalies against your own pattern are calculated from the data. A monthly review looks like this:
September against your norm:
| Metric | Last year | 3-month avg. | Now | Reading |
|---|---|---|---|---|
| Revenue | €103k | €107.8k | €117k | +14% YoY: above norm |
| Desktop conversion rate | 2.4% | 2.3% | 1.8% | below range: diagnose checkout |
| MER | 4.6 | 4.8 | 4.7 | within norm |
| Category: nightwear | €16k | €14.5k | €12.3k | −23% YoY despite the season: to review |
Overall above norm thanks to two categories, but two downward deviations deserve a diagnosis: desktop conversion has been below its own range for three weeks (coinciding with a theme update), and the nightwear category is weaker year on year despite its peak season. None of these observations required a single external average.
Methodological note: the figures in the examples are illustrative. Your own norm requires comparable periods: after major changes (a new theme, a change in pricing policy, a new channel) the ranges are set again from the date of the change. Published benchmarks can serve as a curiosity over coffee; decisions are made on the norm and the physical thresholds.
Key takeaway
The only shop with an identical industry, basket, brand and traffic mix is you a year ago. A table of your own norms plus the physical thresholds answers every “is this good?” that someone else’s average cannot.
FAQ
Frequently asked questions about e-commerce benchmarks
What is the average e-commerce conversion rate?
Published values hover around 1–3%, but they mix industries, baskets and markets, so they do not answer the question about your shop. The practical reference point is your conversion rate from a comparable period plus the trend; someone else’s average will not make any decision for you.
Why are benchmarks misleading?
Five variables skew the comparison: industry and purchase cycle, basket value, share of branded traffic, device and source mix, and the report’s unknown methodology. The result of the comparison says more about the sample than about the shop.
What should you compare yourself with?
With your own norm (ranges from comparable periods: year on year, season, days of the week, per category and device) and with physical thresholds: break-even ROAS from margin, delivery cost vs basket, measurement completeness. These two frames of reference are enough for decisions.
How do you build your own norm?
For each key metric, set a range from history: the comparable period a year earlier plus the average of the latest full periods, broken down by day of the week where the rhythm matters. Current values are read as within norm / above / below, and deviations get a diagnosis.
How does DataOrganizer help?
It keeps the history of metrics from all sources, so norm ranges and year-on-year comparisons are calculated from the data, anomaly detection works on the shop’s own pattern, and the chat answers “is this month good?” with a comparison against your norm per category.
Cheat sheet for the whole guide
Instead of benchmarks: two frames of reference
1. Your own norm. Metric ranges from comparable periods (year on year, season, days of the week), per category and device. The only benchmark that knows your basket, brand and traffic.
2. Physical thresholds. Break-even ROAS from margin, delivery cost vs basket, measurement completeness: values derived from calculation that apply to everyone.
3. Decisions without external averages. DataOrganizer keeps the history of all metrics together, so the norm table, YoY comparisons and anomalies against your own pattern refresh automatically, and the question “is this good?” is answered from your data.
DataOrganizer · a shop compared with its own norm
Connect your sources: norm ranges, year-on-year comparisons per category and anomalies against your own pattern will be calculated from the data. Demo account available instantly; you pay only for connected sources.