Klaviyo Says Email Drives 30% of Your Revenue. Here's How Much It Really Does

The retention agency's monthly report opens with a big number: email and SMS drove 38% of revenue. The proposal on the next slide follows naturally. If campaigns work this well, send more of them: six a week instead of four, plus a sitewide code to the engaged segment every other Friday. The founder is about to sign off.

Then someone asks a simple question: what happens to revenue if we stop sending campaigns for six weeks to a random tenth of the list? The answer, when they finally run the test, is that the business would lose about 8% of revenue, not 38%. Most of the rest was customers who were going to buy anyway, often from a phone that "opened" the email on its own. Email was still valuable. It was just a fifth as valuable as the dashboard claimed, and the plan to send more was about to make it worse.

Every founder has seen the Klaviyo dashboard number. Email and SMS drove 28%, 35%, sometimes 40% of revenue last month. It is the most comforting number in the business, because it seems to prove the channel you pay the least for is doing the most work.

The better question is not "how much revenue did email touch?" It is "how much revenue would we lose if we stopped sending tomorrow?" Those two numbers are rarely the same, and the gap between them decides how many campaigns you send, how much discount you give away, and whether your retention hire is worth what you pay.

How Klaviyo decides a sale belongs to email

Klaviyo uses a last touch model with a time window. If a person opened or clicked a message and then placed an order inside the window, the order is credited to the most recent message they engaged with. For accounts created after October 9, 2024, the default window is 5 days for email and SMS and 24 hours for push. Older accounts may still be on different defaults, and every account can change them.

Three details in that rule matter far more than most teams realise.

• Opens count, not just clicks. A customer does not need to click anything. Opening the email and buying through any other route within the window is enough.
• Machine opens count by default. Apple Mail Privacy Protection preloads emails on Apple devices, which registers as an open whether or not a person read it. Klaviyo counts these opens toward conversions unless you switch that off.
• The order can come from anywhere. Klaviyo does not check whether the buyer arrived from a Google ad, a Meta ad or by typing your URL. If the open happened inside the window, email gets the credit.

So a customer on an iPhone who receives your Monday newsletter, never reads it, sees a Meta ad on Wednesday and buys, is counted as an email conversion. Meta counts the same order as a Meta conversion. Google may count it too if they searched your brand name before buying.

The four kinds of revenue hiding inside the email number

When you break email-attributed revenue down, it almost always falls into four buckets. Only one of them is fully incremental.

The pull-forward bucket is the one founders miss most. Look at daily revenue around a big campaign send. If the campaign day spikes and the following four to six days sit below your normal run rate, a large part of the spike was timing. Total revenue for the fortnight barely moved.

Flows and campaigns behave very differently

Flows are triggered by what a customer just did: abandoned a checkout, browsed a product, went quiet for 90 days. Campaigns are sent because it is Tuesday. That difference shows up clearly when you test them.

• Abandoned checkout and browse abandonment usually hold up best in holdout tests. The customer showed intent and the message reaches them while it is still warm.
• Welcome series often lift first purchase meaningfully, but check what the welcome discount costs on customers who would have bought at full price. We covered that in Is your welcome discount building customers or bargain hunters?
• Winback flows to genuinely lapsed customers are often incremental, because nothing else was going to bring those people back.
• Post-purchase flows tend to look strong in attribution and weaker in holdouts, because they go to people who just bought and many were going to buy again anyway.
• Campaigns to your engaged segment are where the gap is widest. These people open everything and buy often. Klaviyo credits their purchases to whatever they opened last.

How to run a holdout on any list size

Klaviyo offers global holdout groups, which remove a share of your audience from all marketing messages and report the lift. Klaviyo recommends holding out around 5% and running for about three months, and the feature needs an account with at least 400,000 profiles. Most brands under $10 million in revenue will not qualify. You do not need the feature to run the test.

The simple version (lists above roughly 20,000 active profiles)

1. Create a random 10% segment of your engaged, marketable profiles. The simplest way is to export them, give each profile a random number from 1 to 10 in a spreadsheet, re-import that as a custom property, and build a segment for everyone with the value 10.
2. Exclude that segment from every campaign for six to eight weeks. Keep transactional messages (order confirmations, shipping updates) running for everyone.
3. Leave flows running for both groups at first. You are testing campaigns.
4. At the end, pull Shopify orders for both groups, not Klaviyo-attributed revenue. Use revenue per person across all channels.
5. The difference in revenue per person, multiplied by the size of the mailed group, is your incremental campaign revenue.

Testing a flow

Inside a flow, add a conditional split on the same random property so that, for example, values 1 to 5 go down the normal path and values 6 to 10 exit without a message. Run it until each side has at least a few hundred triggered customers. Compare the conversion rate across all channels over the following 14 to 30 days.

In this example email campaigns are worth something real, about $32,400 over eight weeks. They are worth roughly a quarter of what the dashboard said. That changes how many campaigns you send, how much you pay an agency on a percentage of attributed revenue, and how you value your email list in a fundraise.

Illustrative cases

What to change in your reporting

You do not need to stop using Klaviyo's numbers. You need to stop using them to make budget decisions. A practical setup:

What to do this week

1. Check your attribution window and whether Apple MPP opens and bot clicks are included. Write down the current settings before changing anything.
2. Rerun last quarter's email revenue with click-only, 2-day attribution for your own reference.
3. Add up the revenue Klaviyo, Meta and Google each claimed last month. Compare it with Shopify net sales.
4. Split email-attributed orders into first orders, repeat orders and recurring subscription orders.
5. Set up a 10% random campaign holdout and let it run for six to eight weeks.
6. Decide in advance what result would make you cut, keep or increase campaign frequency.

Email is usually one of the most profitable channels a DTC brand has. The point of measuring it properly is not to cut it. It is to stop paying for the part of the number that was never email's to begin with, and to put the effort where messages actually change what customers do.