💬 Quick answer: Across 24 direct-to-consumer stores, returning visitors were 18.8% of sessions but produced 36.5% of revenue — converting at 2.17% against 0.85% for new visitors, a 2.60× difference. Every store in the panel converted returning traffic better than new.
Updated 30 August 2026 · Data window: July 2025 – June 2026 · Source: first-party client data, pooled and anonymised
TL;DR
- Returning visitors: median 18.8% of sessions, 36.5% of revenue. Roughly a fifth of the traffic earning over a third of the money.
- Returning-visitor conversion rate 2.17% vs 0.85% for new — a median 2.60× gap (IQR 1.74×–3.40×).
- All 24 stores converted returning traffic better. The narrowest gap was 1.03×; the widest 5.17×.
- Returning share of revenue ranged from 19.3% to 57.8% — the widest-spread metric in the set, and the clearest signal of whether a brand has a retention engine.
- GA4 classifies visitors, not customers, and cookie loss counts real returners as new — so these figures understate retention rather than flatter it.
- Window: July 2025–June 2026; 24 stores, 18.5M sessions, 249K transactions.
What percentage of ecommerce revenue comes from returning customers?
Across 24 direct-to-consumer stores the median returning-visitor share of revenue is 36.5%, with the middle 50% between 33.0% and 47.5%. That revenue comes from just 18.8% of sessions.
Do returning customers convert better than new ones?
Yes, consistently. Returning visitors converted at a median 2.17% against 0.85% for new visitors — a 2.60× difference. This held in every one of the 24 stores measured, with the gap ranging from 1.03× to 5.17×.
Is a 37% returning-customer revenue share good?
It is typical. 36.5% is the median across the panel; above 47.5% puts a store in the top quartile, and the highest in the panel reached 57.8%. Below 33.0% suggests the brand is buying new customers faster than it is keeping them.
The headline gap
Across 24 direct-to-consumer stores over the twelve months to June 2026, returning visitors accounted for a median 18.8% of sessions but 36.5% of revenue. Roughly one in five visits produced more than one in three dollars.
The mechanism is conversion rate. Returning visitors converted at a median 2.17% against 0.85% for new visitors — a 2.60× difference, with the middle 50% of stores between 1.74× and 3.40×.
What makes this unusually robust is that it held everywhere. All 24 stores converted returning traffic better than new traffic. Not most, not the well-run ones — all of them. The narrowest gap in the panel was 1.03× and the widest 5.17×, but the direction never reversed. Findings that survive in every single account are rare, and this is one.
The practical implication: traffic you have already paid for once converts two to three times better the second time. Any spend that brings a past visitor back is competing against a much lower bar than spend that finds a stranger.
The full distribution
Per-store rates first, then equal-weight percentiles across the 24 stores — the same method used in every report in this series, so no single high-traffic store can set the benchmark.
| Metric | 10th | 25th | Median | 75th | 90th |
|---|---|---|---|---|---|
| Returning share of revenue | 25.7% | 33.0% | 36.5% | 47.5% | 52.7% |
| Returning share of sessions | 9.0% | 14.5% | 18.8% | 23.0% | 24.7% |
| New-visitor conversion rate | 0.18% | 0.62% | 0.85% | 1.39% | 3.26% |
| Returning-visitor conversion rate | 0.58% | 1.23% | 2.17% | 3.78% | 5.50% |
| Returning : new multiple | 1.42× | 1.74× | 2.60× | 3.40× | 4.50× |
Returning share of revenue has the widest meaningful spread — 19.3% to 57.8% across the panel. That range is the single clearest indicator of whether a brand has built a retention engine or is renting growth from paid acquisition. A store at 19.3% is replacing its customer base every year; a store at 57.8% has one.
Returning share of sessions, by contrast, is tightly clustered (14.5%–23.0% across the middle half). Most DTC stores get a similar proportion of repeat visits. What separates them is how much money those visits are worth.
What GA4 actually measures — and why this understates retention
This is the caveat that matters most, so it goes before the analysis rather than buried in the methodology.
GA4's newVsReturning dimension classifies visitors, not customers. A "returning visitor" is someone whose browser GA4 recognises from a previous visit. That is not the same as a repeat purchaser: a returning visitor may never have bought anything, and a genuine repeat customer who switches from phone to laptop, clears cookies, or returns after their cookie expires is counted as new.
Both errors push in the same direction for the headline finding. Misclassifying real returners as new inflates the "new" group with people who already know the brand — which flatters new-visitor conversion rate and understates the gap. The true difference between a first-ever visitor and a genuine returning customer is almost certainly larger than the 2.60× measured here.
So this is a floor, not a ceiling. We publish it as a floor deliberately: it is the number the data can support, and a conservative figure that survives scrutiny is worth more than an ambitious one that does not.
A third bucket, sessions GA4 could not classify at all, ran at a median 6.6% of sessions per store. Those are excluded from the split entirely rather than assigned to either side.
Why returning visitors convert better
The gap is not evidence that returning visitors are a better audience in some abstract sense. It reflects where they sit in the buying process.
- Consideration is already done. A first visit is usually spent deciding whether the brand is credible and the product is right. A return visit starts past that.
- Intent is self-selected. Nobody comes back to a store by accident. The act of returning is itself a purchase signal, which is why it outperforms any targeting a marketer can buy.
- Trust is established. The friction that stops a first-time buyer — is this real, will it arrive, can I return it — is resolved for someone who has bought before.
- Returning traffic skews owned-channel. Email and SMS overwhelmingly reach people who already know the brand, which is why flow performance and returning-visitor performance track each other so closely.
That last point connects directly to our email flow benchmarks, where triggered flows clicked at 3.2× the rate of broadcast campaigns for the same reason: they reach people at a moment of demonstrated intent rather than on the sender's schedule.
How this squares with our conversion benchmark
Two figures from separate reports in this series should agree, and checking that they do is a useful test of both.
Our ecommerce conversion rate benchmark puts the median DTC site conversion rate at 1.17%. This report puts new-visitor conversion at 0.85% and returning-visitor conversion at 2.17%, with returning visitors making up 18.8% of sessions.
Blending those at the median session mix gives an implied overall conversion rate of roughly 1.10% — close to the 1.17% measured independently, and the small residual is explained by the unclassified session bucket and by the fact that medians of ratios do not recombine exactly. Two separately-built figures landing within a rounding step of each other is a reasonable sign that neither is an artefact.
What to do with this
- Judge acquisition spend on more than first-order return. If a returning visitor is worth 2.60× a new one at the same session count, a campaign that breaks even on first purchase is not breaking even overall — it is buying an asset.
- Segment your conversion-rate reporting. A blended site conversion rate hides both numbers. A blended rate that improves because returning traffic grew is a very different story from one that improves because the store got better at converting strangers.
- Treat the returning share of revenue as the retention KPI. It is the one figure here that separates the panel most sharply, and it moves slowly enough to be worth tracking quarterly rather than weekly.
- Fix flow coverage before adding acquisition budget. Owned channels are the main lever that manufactures return visits — see the flow benchmarks for where the gaps usually are.
- Do not read this as "retention beats acquisition". Returning visitors only exist because acquisition created them. The finding argues for valuing acquisition correctly, not for spending less on it.
Methodology and limits
Figures come from the GA4 Analytics Data API, pulled per store property for the trailing twelve months from 1 July 2025 to 30 June 2026, using the newVsReturning dimension against sessions, transactions and purchase revenue. Each store's rates are computed first, then equal-weight percentiles are taken across stores.
Panel: every direct-to-consumer client store with purchase tracking live and at least 10,000 sessions in the window — 24 stores in total. Seven further properties were examined and excluded, all for data-quality reasons stated in advance: five had no purchase tracking or returned no rows, and two fell below the session floor. No store was added or removed on the basis of its results.
Known limits, stated plainly:
- Visitors, not customers. As set out above, GA4 measures browser-level return, so these figures are a conservative proxy for true repeat-customer behaviour — not a repeat-purchase rate.
- Cookie and consent loss. Cross-device journeys, cleared cookies, expired identifiers and consent refusals all push genuine returners into the "new" bucket. This suppresses the measured gap.
- n = 24 stores is a small panel. It is real first-party data rather than a survey, but it is not a census of DTC ecommerce.
- Unclassified sessions. A median 6.6% of sessions per store could not be classified and are excluded from the split rather than allocated.
- Mixed verticals. The panel spans electronics, pet, home, beauty, apparel, food and beverage, CPG and crafts. Purchase frequency differs sharply by category, which widens the revenue-share range.
- GA4 attribution. Transactions and revenue use GA4's own measurement, which will differ from what a brand sees in Shopify.
- Clients of one agency. These are actively managed stores, which may sit above an unmanaged baseline.
How to benchmark your own store
- In GA4, open Reports → Retention, or build an exploration using the New / returning dimension against sessions, transactions and purchase revenue.
- Set the range to the last full 12 months so seasonality and cookie expiry are handled consistently.
- Calculate returning share of revenue first — it is the figure that separates stores most, and the one to compare against the 33.0%–47.5% interquartile range above.
- Then divide returning conversion rate by new conversion rate. If your multiple is below 1.74×, your owned channels are likely under-built rather than your site being broken.
- Exclude the unclassified bucket from both sides rather than assigning it, or your split will not be comparable to this panel.
- Track it quarterly, not weekly. This metric moves slowly and weekly readings are mostly noise.
Sources and methodology
Frequently asked questions
What percentage of ecommerce revenue comes from returning customers?
Across 24 direct-to-consumer stores the median returning-visitor share of revenue is 36.5%, with the middle 50% falling between 33.0% and 47.5%. The full range ran from 19.3% to 57.8%. Those returning visitors were only a median 18.8% of sessions.
Do returning visitors convert better than new visitors?
Yes, and consistently. Returning visitors converted at a median 2.17% against 0.85% for new visitors, a 2.60× difference. This held in all 24 stores in the panel, with the gap ranging from 1.03× at the narrowest to 5.17× at the widest. The direction never reversed in any account.
Is this the same as repeat purchase rate?
No, and the distinction matters. GA4 classifies visitors by whether their browser is recognised from a previous visit, not by whether they have purchased before. A returning visitor may never have bought, and a genuine repeat customer switching devices or clearing cookies is counted as new. These figures are a conservative proxy for repeat-customer behaviour, not a repeat-purchase rate.
Does cookie loss make these numbers unreliable?
It makes them conservative rather than unreliable. Cross-device journeys, cleared cookies, expired identifiers and consent refusals all move genuine returners into the new bucket. That inflates the new group with people who already know the brand, which flatters new-visitor conversion and understates the gap. The true difference is very likely wider than the 2.60× measured here.
What is a good returning-customer revenue share for a DTC store?
36.5% is the panel median, so a store near that figure is typical. Above 47.5% is top quartile. Below 33.0% suggests the brand is acquiring new customers faster than it is retaining them, which usually shows up as rising acquisition costs before it shows up in revenue.
Why is returning share of sessions so much lower than returning share of revenue?
Because returning visitors convert at a much higher rate. A median 18.8% of sessions produces 36.5% of revenue precisely because those sessions convert at 2.17% rather than 0.85%. The revenue share is the conversion gap expressed in money.
How was this benchmark measured?
GA4 Analytics Data API, pulled per store property for the trailing twelve months from 1 July 2025 to 30 June 2026, using the newVsReturning dimension against sessions, transactions and purchase revenue. The panel is every direct-to-consumer client store with purchase tracking live and at least 10,000 sessions — 24 stores, 18.5 million sessions and 249,453 transactions. Each store's rates are calculated first, then equal-weight percentiles are taken across stores so no single large store sets the benchmark. Unclassified sessions are excluded from the split and disclosed.
Does this mean we should spend less on acquisition?
No. Returning visitors only exist because acquisition created them, so the finding argues for valuing acquisition correctly rather than reducing it. The practical read is that a campaign judged solely on first-order return is undercounting its value, because the visitor it acquires converts two to three times better on any subsequent visit.
- Google Analytics Data API — newVsReturning dimension
- DTC Ecommerce Conversion Rate Benchmark
- Email Flow Benchmarks
- DTC Email Marketing Benchmark
- Checkout Abandonment Rate Benchmarks
- Meta Ads Benchmarks
See also our ecommerce conversion rate benchmark for the blended site figure this report decomposes, and our email flow benchmark for the owned-channel mechanics that manufacture return visits. Our Meta ads benchmark covers the paid acquisition side.