The State of Ecommerce Retention 2026: What 109 Klaviyo Audits Show

Every agency in this category competes on the same number: what percentage of your revenue comes from email. Ours included. It is on our own money pages.

Having now audited more than a hundred ecommerce Klaviyo accounts and written down what we found, we think that number is the most misleading statistic in retention marketing. Not because agencies are lying. Because the number itself is a measurement artefact, and almost nobody checks how it was produced before quoting it.

This is what our audit archive says. It is the first time we have published it.

Key Findings

• The typical brand arrives at 10 to 15% of revenue from email and SMS. The spread across audits ran from 2% to 49%, which makes any single "industry benchmark" close to meaningless without a category attached.
• At least a quarter of the accounts where we could check were reporting a materially inflated figure. One brand's 45 to 49% collapsed once the attribution window was corrected. Another's 18% was closer to 10 to 15% on honest settings.
• Flows are underbuilt almost everywhere. One brand had automations producing 5% of its email and SMS revenue. The healthy range is far higher, and it is the single most common structural gap we find.
• Abandoned cart open rates are the most reliable early warning. We saw 27.5% and 42% against a 60% benchmark, in accounts where nobody had flagged a problem.
• Subscription penetration is much lower than founders assume. One subscription-led brand had 6.3% of customers subscribed against a 25 to 35% benchmark. Another had 8.9%, meaning 91.1% of its customers were one-time buyers.
• The uncomfortable one: we have never once audited an account where the reported email revenue share was too low because of conservative attribution. The error runs one way.

Methodology, and Why You Should Discount Some of This

Putting this first, because a research piece that hides its method is marketing.

Source. Our own audit archive. Every audit and account review we run is recorded and summarised. We searched the full archive and found 109 recorded audit and account-review calls. Of those, 34 contain quantified retention data, and 20 record an email or email-plus-SMS revenue share figure at the time of the audit. Where a range was recorded, we used the range.

Anonymisation. Every brand below is described by category only. These figures were shared with us in confidence during audits and sales conversations. The only brand named anywhere in this report is Koala Eco, a former client who reviewed and approved their numbers and provided a testimonial. Our other named results are the case studies already published on our site.

Three limitations you should hold against these numbers:

1. Selection bias, and it is severe. These are brands that booked an audit with a retention agency. People do that when they suspect something is wrong. This sample is tilted toward underperformance and you should not read the median as an industry average. It is a median of brands who thought they had a problem.
2. Attribution is not standardised. That is the point of the report, but it cuts both ways: the figures below were produced by whatever settings each brand happened to be running. We note where we corrected for it and where we could not.
3. Sample size. Twenty revenue-share data points is enough to describe a distribution. It is not enough to publish a category-level benchmark to one decimal place, and we have not tried to.

Finding 1: The Range Is So Wide That "Industry Benchmark" Is Almost Useless

Here is the distribution of email and SMS revenue share at the moment we opened each account. Categories only.

The median lands between 10 and 15%. The top of the range is more than twenty times the bottom.

Two things follow. First, when an agency tells you the benchmark is 30%, ask 30% of what, for whom, measured how. A premium meat brand with a seasonal gifting spike and a games brand selling one product to one household are not on the same scale and never will be. Second, the brands at the top of this table are not necessarily better run than the brands in the middle. Which brings us to the actual finding.

Finding 2: The Number Is Systematically Inflated, and It Only Errs in One Direction

This is the part we did not expect to be so consistent.

In every audit where we were able to check the attribution configuration against the reported number, the reported number was too high. Not once did we find a brand under-reporting because of conservative settings.

Four examples from the archive, anonymised:

The apparel accessories brand reporting 45 to 49%. Our audit note records the share as "overinflated due to generous attribution settings." This brand believed email was carrying roughly half the business.

The toys and collectibles brand reporting ~18%. Running a 7-day attribution window. Corrected, the true contribution looked closer to 10 to 15%, which on their volume was an estimated $250,000 a year of difference between the story and the reality.

The functional beverage brand reporting a healthy number until we stripped out non-incremental subscription renewals that Klaviyo was counting as email-driven revenue. A subscription renewing on schedule is not an email you should take credit for. Adjusted, it was around 10%.

The supplements brand at 33%, inflated by a paid media outage. Their Meta account had problems, traffic shifted to owned channels, and email's share rose for reasons that had nothing to do with the email program. When the ads came back, the "gain" would have evaporated.

And a fifth, the one that generalises: a sleep brand reporting 6%, which our note flags as "likely inflated by Klaviyo's standard 5-day attribution window." Even at the bottom of the range, the number was generous.

Why this happens, and why it is not fraud

Klaviyo reports last-click within an attribution window. If a customer clicked an email inside that window and then bought, the order is counted, whether or not the email caused it. That is exactly what the platform says it does. It is measuring its own channel.

A marketing mix or multi-touch tool like Triple Whale or Northbeam is answering a different question: given everything that touched this customer, how much credit does email genuinely deserve. The same month will read materially lower there.

Neither tool is lying. They are answering different questions. The problem is that an entire industry has standardised on quoting the higher one without saying so, including us. Every revenue-share figure on our own website is Klaviyo-attributed, and we now label it that way.

The three questions this gives you

Ask any agency, and ask your own team:

1. What is the attribution window? A 1-day click window and a 7-day open window produce numbers that are not comparable.
2. Does it count opens or only clicks? Apple Mail Privacy Protection auto-opens a large share of email. Open-based attribution in 2026 is not generous, it is fiction.
3. Does it match how you measure your other channels? If email is on last-click and paid is on an MMM, you are comparing two different currencies and email will always look better.

Then compare like with like: Klaviyo against Klaviyo, never an agency's Klaviyo screenshot against your own Northbeam dashboard. A meaningful number of agency relationships have ended over exactly that mismatch, with nobody lying.

Finding 3: Automations Are Underbuilt Almost Everywhere

The clearest structural pattern in the archive. Brands over-invest in campaigns, which are visible and feel like work, and under-invest in flows, which run silently and compound.

From the audits:

• An air purification brand where automated flows produced 5% of email and SMS revenue. The strategic goal we set was 40%.
• The same brand running a 90/10 campaign-to-flow revenue split.
• A premium meat brand at 70/30, campaign-heavy, in a category where post-purchase automation should be doing heavy lifting.
• A haircare brand where campaigns drove only 41% of non-sale revenue, the inverse problem, with an over-reliance on discount events.
• A home and body care brand sitting at 50/50 where the target was 70/30.

The nuance most benchmarks miss: the right split changes with maturity. Early in an engagement, flows should carry the program, because fixing broken automations is the fastest money in the building. In a mature, well-run account the ratio flips and campaign volume becomes the growth engine. In our most optimised accounts, roughly 70% of revenue comes from campaigns and 30% from flows.

So a 90/10 campaign split is not automatically wrong. A 90/10 split with flows producing 5% of revenue is, because it means the automations were never built, not that the program outgrew them.

Finding 4: The Abandoned Cart Open Rate Is the Best Early Warning You Have

If you only check one thing after reading this, check this one.

Two audits recorded abandoned cart open rates of 27.5% and approximately 42% against a 60% benchmark. In both cases nobody at the brand had flagged a problem, because total email revenue looked acceptable.

An abandoned cart email goes to someone who was on your site, with your product in their basket, minutes ago. It is the highest-intent message you will ever send. If it is not being opened at 60% or better, something is broken upstream: the trigger, the segmentation, the sending domain, or the timing. Low engagement here is never a copy problem.

Related welcome flow findings from the same archive: an eyewear brand at 44% open and 2.9% click, and a supplement brand whose discount email opened at 51.7% against a 60 to 70% benchmark while training its customers to wait for a coupon.

Finding 5: Subscription Penetration Is Far Below What Founders Assume

Two data points, both from brands that considered themselves subscription businesses.

• A supplement brand with 6.3% of customers subscribed, against a 25 to 35% benchmark for its model, and an orders-per-customer figure of 1.1 to 1.2. Our audit traced the cause to aggressive, inconsistent discounting: BOGOs and a 20% welcome offer that made the one-time purchase cheaper than committing to a subscription. The subscription offer was competing with the brand's own promotions and losing.
• A home and body care brand with 8.9% of customers subscribed, which is the more useful way of saying that 91.1% of its customers were one-time buyers.

That second framing is the one to internalise. When subscription penetration is under 10%, the opportunity is not optimising the subscribers you have. It is the ninety percent who bought once, and the post-purchase window where you either convert them or lose them.

What Good Actually Looks Like

Drawing on both the audit archive and the accounts we run. Where a number is from our own managed accounts we say so.

Retention economics, the numbers that matter more than channel share:

• Repeat customer rate: 20 to 30% of your customer base having ordered more than once, across DTC generally. Consumables should reach 40 to 55%, apparel 25 to 35%, considered and durable goods 10 to 20%.
• Revenue from returning customers: 35 to 45% for a healthy DTC brand, 50 to 65% with real subscription penetration.
• First-to-second order conversion: typically only 20 to 30% of first-time buyers ever place a second order. Once they do, the odds of a third rise to roughly 45 to 60%. Order two is the hinge, and most post-purchase programs are pointed at the wrong place.
• Subscription churn: 5 to 8% monthly is good. Split voluntary from involuntary, because failed payments are usually 20 to 40% of total churn and are the cheapest thing in the program to fix.
• LTV to CAC of 3:1 on contribution margin, not revenue. Revenue LTV flatters every discount-heavy program ever built.

Channel health, first-party from accounts we manage as of 2026:

• Revenue share: brands typically arrive around 15% and should clear 25% by month six. Klaviyo-attributed, read it against your Klaviyo.
• Open rates: 60 to 65% on well-segmented campaigns. Apple-inflated, so a health signal and never a success metric.
• Click rates: 0.7 to 1.3% after bot-click filtering. An agency bragging about 3% is cherry-picking a segment or counting scanner traffic.
• Abandoned cart open rate: 60%+.
• Popup signup rate: 6%+. The old best practice of 3% is roughly what a default Klaviyo popup converts at. On one account, replacing the popup took Australian signups from about 2% to roughly 16%.
• Campaign cadence: 12 to 30 per month for 7 and 8-figure brands.
• Email to SMS: about 80/20.

The Six Checks You Can Run on Your Own Account This Afternoon

None of these need an agency. All six come from the failures that recur most often in the archive.

1. Find your attribution window. Klaviyo, Settings, then your attribution configuration. Note the window and whether it counts opens. Every revenue number you have ever quoted was produced by this setting.
2. Open your abandoned cart flow and read the open rate. Under 60% means something is broken upstream of the copy.
3. Check flow recipients, not flow revenue. A flow with almost no recipients over 30 days is not underperforming, it is off. Stacked exclusion rules are the usual cause and they are invisible until you look.
4. Calculate what share of last year's customers bought again. Almost nobody knows this offhand and almost everybody who calculates it finds it lower than they assumed.
5. Measure the median gap between first and second order. Then check whether your post-purchase timing fires before it. Most flows are timed on an agency template rather than the brand's actual repurchase curve.
6. If you sell subscriptions, calculate penetration. Subscribers divided by total customers. Under 10% means your opportunity is the one-time buyers, not the subscribers.

What We Changed About Our Own Marketing Because of This

Two things, and we mention them because it would be strange to publish this and not.

We now label every revenue-share figure on our site as Klaviyo-attributed, and tell prospects to compare Klaviyo to Klaviyo and to agree the scoreboard tool before signing.

And we report client results incrementally where we can. Our Koala Eco engagement took email and SMS from 12% to 23% of US store revenue and 19% to 31% in Australia between a September 2025 baseline and April 2026. The number we lead with is not the $1.4M of gross Klaviyo-attributed revenue over that period. It is the $700,000+ of revenue above the pre-engagement baseline, which on a $4,500 per month retainer is roughly a 22x incremental return. The gross figure is the bigger headline and the worse measurement.

That is the question worth asking any agency, including us: what would my number have been without you?

Frequently Asked Questions

What percentage of ecommerce revenue should come from email and SMS?

Across our audit archive the range was 2% to 49%, with a median between 10 and 15% at the time of audit. Well-run programs typically clear 25% by month six, and 30 to 40% is realistic for subscription or high-repeat products. But the range is so wide by category that a single benchmark is close to useless. A premium meat brand with seasonal gifting and a games brand selling one product per household have structurally different ceilings. Ask what the benchmark is for your category and your attribution settings before accepting any number.

Is Klaviyo's revenue attribution accurate?

Klaviyo reports last-click within an attribution window, which is exactly what it says it does. The problem is interpretation. In every audit where we could check the configuration against the reported figure, the reported figure was too high, and we have never found the error running the other way. Common causes: a 7-day or open-based window, subscription renewals counted as email-driven, and paid media outages that shift traffic to owned channels. The same period will read materially lower in a marketing mix tool like Triple Whale or Northbeam.

What is a good abandoned cart open rate?

60% or better. Two audits in this research recorded 27.5% and roughly 42%, in accounts where nobody had flagged a problem because total email revenue looked acceptable. An abandoned cart email goes to someone who was on your site with your product in their basket minutes earlier, so it should be your highest-engagement message. Anything well under 60% points to a broken trigger, bad segmentation, a sending domain issue or wrong timing, not to weak copy.

What share of my customers should be subscribers?

For a brand built around subscription, 25 to 35% is a reasonable benchmark. Both subscription-led brands in this research were far below it, at 6.3% and 8.9%. In one case the cause was the brand's own promotions: aggressive BOGOs and a 20% welcome discount made buying one-time cheaper than subscribing, so the subscription offer was competing with the brand's own discounting and losing. If your penetration is under 10%, the opportunity is your one-time buyers, not your existing subscribers.

What percentage of email revenue should come from automated flows?

It depends on maturity, which most benchmarks miss. Early on, flows should carry the program, because repairing broken automations is the fastest available money. In a mature, optimised account the ratio flips and campaigns become the growth engine, typically around 70% campaigns to 30% flows in the accounts we run. The genuine warning sign is not a campaign-heavy split, it is flows producing a tiny share of revenue in absolute terms. One brand in this research had automations generating 5% of email and SMS revenue, which means they were never built rather than outgrown.

How was this research conducted?

We searched our own recorded audit archive and found 109 audit and account-review calls. Thirty-four contain quantified retention data and twenty record an email or email-plus-SMS revenue share figure at the time of audit. Every brand is described by category only, because these figures were shared with us in confidence. The sample has severe selection bias: these are brands that booked an audit with a retention agency, which people generally do when they suspect something is wrong. Read the median as a median of brands who thought they had a problem, not as an industry average.

Who conducted this research?

The Email Marketers, a retention marketing agency for 7 to 9-figure ecommerce brands and a Klaviyo Platinum Master partner. We publish this dataset knowing it undercuts a statistic our own industry, including us, uses to sell. We think the category is better served by buyers who know how to interrogate the number than by another round of agencies quoting it at each other.

---

About the author

Melanie Balke is the founder and CEO of The Email Marketers, a retention marketing agency for 7 to 9-figure ecommerce brands, and host of the No Mild Takes podcast. She has worked retention from every seat: in-house at an ecommerce brand, as a freelancer, inside another agency, and since 2019 running her own. Her team generated over $103 million in attributed client revenue in 2025. Clients include Grüns, Koala Eco, Gimme Seaweed, Elevate Outdoor Collective (the company behind K2 Skis and Völkl), Open Store, Outer Furniture, The Freeze Pipe, and Llama Naturals. Connect on LinkedIn or X.

Methodology summary: source is The Email Marketers' internal audit archive, comprising 109 recorded audit and account-review calls. 34 contain quantified retention data; 20 record an email or email-plus-SMS revenue share figure at time of audit. All brands are anonymised to category level with the exception of Koala Eco, who reviewed and approved the use of their figures. The sample is drawn from brands that requested an audit and is therefore biased toward underperformance; it should not be read as an industry average. Channel-health benchmarks described as first-party are drawn from accounts actively managed by The Email Marketers as of September 2026 and are Klaviyo-attributed using click-based attribution. Data compiled September 2026.

Journalists and researchers: the underlying category-level dataset is available on request at hello@theemailmarketers.com. We will share it with anyone, including competitors.