How Klaviyo Calculates Revenue: Why Your $50K Flow Number Needs Context

Somewhere in your Klaviyo dashboard right now sits a number that probably feels a little too good. Maybe it says your flows drove $50,000 last month. Maybe your email revenue share reads 40% of total store revenue, which sounds fantastic until you try to reconcile it with what Shopify and GA4 are telling you at the same time. If you’ve ever stared at that number and thought “there’s no way this is fully accurate,” you were right to be suspicious, and you’re not alone.

This isn’t a story about Klaviyo lying to you. Klaviyo is doing exactly what it was built to do: track clicks, apply a rule, and assign credit. The problem is that the rule it applies, and the assumptions baked into that rule, produce a number that answers a narrower question than most people think it does. Understanding how Klaviyo calculates revenue, and why that calculation is structurally different from “how much revenue your email program actually caused,” is the difference between making confident decisions and making expensive ones.

This guide walks through the mechanics of Klaviyo’s attribution model, the specific reasons its reported revenue runs higher than reality, and a practical way to calculate a number you can actually defend to a CFO, an investor, or your own future self.

How Klaviyo Actually Calculates Revenue

Klaviyo uses what’s called a last-click, or last-touch, attribution model. In plain terms, this means that when a customer completes a purchase, Klaviyo looks backward to find the most recent email or SMS message that customer opened or clicked, and if that interaction happened within a defined lookback window, the full value of that order gets credited to that message. According to Klaviyo’s own help documentation, the platform uses this last-touch model with a default lookback window for all new accounts, and it applies cooperative multi-channel logic when a customer has interacted with both email and SMS, meaning whichever channel was touched most recently within its own window wins the credit.

What Counts as an Attributed Order

The default windows matter a great deal here. As Klaviyo explains on its own blog, the standard setup gives email a 5-day open-or-click window and SMS a 24-hour click window. That means if a customer opens or clicks an email and then completes a purchase anytime within the next five days, that order gets attributed to email, provided no more recent message interaction took precedence. One useful thing Klaviyo does get right internally is deduplication. Since the model always resolves to a single most-recent touchpoint, a purchase never gets counted twice between your own email and SMS sends. The problem, as you’ll see below, is that this internal tidiness says nothing about what happens once you compare Klaviyo’s number to what Meta, Google, or your own Shopify dashboard are reporting for the exact same order.

Correlated Revenue vs. Incremental Revenue: The Distinction That Actually Matters

Here is the single most important idea in this entire guide, and it’s worth reading slowly. Correlated revenue is any purchase that happened near a marketing touchpoint. Incremental revenue is the purchase that would not have happened without that touchpoint. Klaviyo, like every other last-click attribution system, measures correlation. It cannot measure causation, because doing so would require knowing what the customer would have done in a world where your email never arrived, and no dashboard has access to that alternate reality.

This distinction sounds academic until you apply it to your own numbers. If a customer was already planning to buy a product they’d been browsing for a week, and your abandoned cart flow happens to email them the day before they were going to check out anyway, Klaviyo will confidently attribute that entire order to the flow. The flow gets full credit for a purchase decision that was already made. Multiply that pattern across thousands of orders a month, and you start to see why “Klaviyo revenue” and “true email revenue” are two different numbers describing two different things, even though the dashboard presents them as if they were the same.

Why Customers Who Were Going to Buy Anyway Get Attributed to Your Flows

There are three specific mechanical reasons Klaviyo’s reported revenue systematically overstates what your email and SMS program is actually causing, and understanding each one helps explain why the gap is so consistent across brands.

The first is channel overlap. Picture a customer who clicks a Klaviyo email on Monday, sees a retargeting ad on Instagram Tuesday, and completes the purchase Wednesday. Klaviyo attributes 100 percent of that order to the email, because that was the most recent tracked interaction within the window. Meta’s own ad platform, running its own last-touch logic independently, is very likely also claiming that same order as an ad-driven conversion. As Attribution’s analysis of Klaviyo’s model points out, when a customer clicks a Meta ad and then receives a Klaviyo abandoned cart email before buying, Klaviyo claims the full sale, and so does Meta. Two platforms, one order, two 100 percent attribution claims. If you added up every channel’s attributed revenue across your entire stack, the total would almost certainly exceed your actual store revenue, which is a strong clue that something in the accounting doesn’t reconcile with reality.

The second reason is what you might call untracked-click orders, and it works in the opposite direction. A customer opens an email, reads it, closes it without clicking anything, then opens a new browser tab, navigates to your site directly, and buys. Because no click occurred, Klaviyo’s attribution never fires for that order, even though the email may well have been the reason the customer thought to visit at all. This is a real source of undercounting that partially offsets the overcounting described above, which is part of why the net effect on total attributed revenue is genuinely hard to predict without testing.

The third is dark social, sometimes called word-of-mouth leakage. A subscriber receives your email, forwards it to a friend, and the friend clicks through and buys. Subjectlime’s breakdown of Klaviyo attribution notes that in this scenario, Klaviyo attributes the resulting order back to the original recipient, not the friend who actually converted, which means the report technically shows a “successful” attribution while completely misrepresenting who the customer actually was and what convinced them to buy.

What Klaviyo’s Attribution Window Sees, and What It Misses

Scenario	What Klaviyo Attribution Does
Email click, then ad exposure, then purchase	Credits 100% to email. The ad platform likely also claims 100%.
Email opened, closed, direct site visit, purchase	No click occurred, so attribution does not fire. Order is undercounted.
Email forwarded to a friend who buys	Credits the original recipient, not the actual buyer.
Customer was already going to buy regardless	Credits the flow in full. No mechanism exists to detect intent that predates the touchpoint.

How the Numbers Get Inflated, With the Actual Math

The attribution window setting itself has a direct, measurable, and fairly dramatic effect on your reported revenue, and it’s worth seeing the actual numbers rather than taking that on faith. Research from Bardeen’s guide to Klaviyo attribution shows that moving from the default 5-day window out to 30 days can show email driving roughly 20 percent more sales than the default setting would report, simply because more purchases fall inside the longer lookback period. Go the other direction and shorten the window to just 1 day, and attributed revenue can be cut roughly in half, since only purchases happening almost immediately after the click still qualify. Subjectlime’s agency data narrows that further: shortening from the 5-day default down to just 3 days typically drops attributed revenue by 15 to 25 percent on its own. None of these adjustments change how many people actually bought your product. They only change which purchases the software is willing to count.

This is exactly why comparing your Klaviyo number to your Shopify or GA4 number rarely lines up cleanly, and it’s rarely a sign that something is broken. GA4 typically uses UTM parameters and a last-non-direct-click model with its own default 30-day window, while Klaviyo is working off message-level opens and clicks with a much shorter window by default. Subjectlime’s research puts the typical gap between the two platforms at 20 to 40 percent, and both numbers are technically “true,” they’re just answering different questions using different rules.

Putting It in Context: The 70 Percent Rule

One useful way to translate a Klaviyo number into something closer to reality, without running a full test, comes from Subjectlime’s portfolio analysis across many ecommerce accounts. Across their client base, Klaviyo-attributed revenue has consistently landed at roughly 70 to 85 percent of what true, causally-driven email revenue actually turns out to be once more rigorous methods are applied. Here’s how that plays out in practice. Say Klaviyo reports $10,000 in attributed weekly revenue. Applying a 70 percent discount to account for overcounting puts estimated true email revenue closer to $7,000. If total Shopify revenue for that same week was $40,000, the honest email contribution is closer to 17.5 percent of total revenue, not the 25 percent a naive reading of the dashboard would suggest. That gap matters enormously the moment someone in a leadership meeting starts making budget decisions based on the bigger number.

It’s also worth knowing what a healthy range actually looks like once you’re reading the number correctly. For a mature direct-to-consumer brand, Subjectlime’s benchmark data suggests that email and SMS attributed revenue landing somewhere between 25 and 40 percent of total Shopify revenue is a reasonable, healthy signal. Anything below 15 percent points to a genuinely underperforming channel worth investigating. Anything above 50 percent is usually not a sign of an incredible email program. It’s a sign that something in your tracking setup, most likely double counting between overlapping channels or an unusually long attribution window, is inflating the number past what’s plausible.

What Real Incremental Revenue Actually Looks Like, and How to Calculate It

If a dashboard cannot answer the question “did this touchpoint actually cause this sale,” what can? The answer is incrementality testing, and while it sounds technical, the underlying idea is simple and borrowed directly from clinical research. You create two groups of otherwise similar customers. One group continues receiving your normal marketing, whether that’s a flow, a campaign, or an ad. The other group, called the holdout or control group, is deliberately excluded from that specific touchpoint. After enough time passes, you compare purchase behavior between the two groups. Whatever difference remains between them is your actual incremental lift, the revenue that touchpoint caused, isolated from everything else happening in the customer’s world at the same time.

According to the CDP.com glossary on incrementality testing, holdout groups typically represent somewhere between 5 and 10 percent of the total audience, which is usually enough to detect a meaningful difference without sacrificing too much revenue from customers you’re deliberately not marketing to during the test window. For the results to actually mean something statistically, Amplitude’s guidance on experiment design recommends at least 1,000 people per group as a reasonable minimum to confidently detect something like a 10 percent lift, though the exact number depends on your baseline conversion rate and how large a difference you’re trying to detect.

The gap between attributed performance and true incremental performance can be larger than most teams expect. Common Thread Collective’s incrementality work with ecommerce brands documented a real case where Meta’s own reported ROAS for an acquisition campaign sat at 1.57, while a proper geo holdout test revealed the true incremental ROAS was actually 2.42, meaning the platform’s own number was significantly understating performance in that particular instance. The lesson isn’t that platform-reported numbers are always wrong in one predictable direction. It’s that you genuinely cannot know which direction the error runs, or how large it is, until you test.

This same logic applies directly to Klaviyo. Attribution’s platform notes that incrementality testing lets brands run holdout experiments specifically to answer whether paid retargeting is adding real lift on top of what Klaviyo would have converted anyway, or whether that ad spend is simply buying credit for a sale that was always going to happen through email. That is precisely the kind of question a last-click dashboard, no matter how carefully you read it, is structurally incapable of answering on its own.

The Business Decisions Being Made on Inflated Numbers

None of this would matter much if attribution numbers stayed inside a marketing dashboard and never influenced anything. They don’t stay there. Founders present them to investors. Marketing teams use them to justify budget. Leadership uses them to decide which channels get more spend and which get cut. When the underlying number is inflated and nobody in the room understands why, the decisions built on top of it inherit that same inflation.

One particularly dangerous pattern is what happens when platform-reported performance looks stable even as true incremental performance is quietly declining. Research from Stella’s guide for advanced marketing measurement points out that a channel can maintain a steady attributed ROAS on paper even while the actual incremental contribution of that channel is falling, because platform attribution has no way to distinguish a genuinely incremental conversion from one that was going to happen regardless. A team watching only the dashboard would see a healthy, unchanged number and have no reason to investigate. A team running periodic incrementality checks would catch the decline months earlier, while there’s still time to act on it.

The organizational cost of this shows up in a few predictable ways. Budget gets reallocated toward whichever channel’s dashboard looks best, rather than whichever channel is actually driving incremental growth. A retention program can look wildly successful in Klaviyo’s reporting while contributing far less real lift than the number suggests, which crowds out investment in the parts of the customer journey that would benefit more. And perhaps most corrosively, once one team notices that Klaviyo, GA4, and Shopify all tell a different story about the same period, trust in the data itself starts to erode, and decisions increasingly get made on gut feeling instead, which is precisely the outcome good measurement was supposed to prevent.

None of this means you should distrust Klaviyo’s reporting entirely or stop looking at it. Attributed revenue is still genuinely useful for relative comparisons, like tracking whether this month’s flow performance improved compared to last month’s under the same settings, or comparing one campaign against another sent to a similar audience. What it’s not built for is telling you the absolute truth about how much revenue your email program is causing in isolation, and treating it as though it does is where the expensive decisions start.

Klaviyo Attribution Self-Audit Checklist

Run your current reporting setup against these checks. If more than two or three come back as gaps, your team is very likely making decisions on a number that needs more context than it’s currently getting.

1. You know your current attribution window settings for both email and SMS, and you didn’t just leave them on default without checking.
2. You’ve calculated what percentage of total Shopify revenue is attributed to Klaviyo, and you know whether it falls in the healthy 25 to 40 percent range.
3. Anyone presenting Klaviyo revenue to leadership applies a discount factor, such as the 70 percent rule, rather than presenting the raw attributed number as fact.
4. You’ve disabled open-based attribution for email, since Apple Mail Privacy Protection makes open tracking unreliable for most accounts.
5. You’ve run, or have a plan to run, at least one holdout test on a major flow or campaign segment to measure real incremental lift.
6. You understand why Klaviyo, GA4, and Shopify report different revenue numbers for the same period, and you know which number to trust for which specific decision.
7. Your team excludes customers currently active in a flow from receiving overlapping campaigns, to reduce the flow-versus-campaign attribution conflict.
8. You revisit your attribution settings periodically rather than treating them as a one-time setup decision made years ago.

How Klaviyo Calculates Revenue: FAQ

Why doesn’t my Klaviyo revenue match my Shopify revenue?

Klaviyo only counts orders where a customer clicked or opened a tracked message within your attribution window before purchasing, while Shopify counts every order placed on your store regardless of source. Klaviyo revenue will always be a subset of total Shopify revenue, and if the gap seems unusually large or small, the first thing to check is your attribution window and whether open-based attribution is still enabled.

Should I shorten my Klaviyo attribution window?

Many agencies recommend narrowing the default window, often to a click-only 3-day setting, because it reduces overlap with other channels and produces a number that’s easier to compare against true email performance. This will make your attributed revenue number go down, which isn’t a sign of worse performance. It’s a sign of more honest measurement, and it’s worth communicating that clearly to your team before making the change so nobody mistakes the new, lower number for a real drop in sales.

What’s the simplest incrementality test I can run?

A basic holdout test is the most accessible starting point. Take a flow or campaign segment, hold out roughly 5 to 10 percent of eligible customers from receiving it, and after a few weeks compare purchase rates between the group that received the messaging and the group that didn’t. The difference in conversion between the two groups is your real incremental lift for that specific touchpoint.

Does this mean Klaviyo’s reporting is useless?

Not at all. Klaviyo’s attributed revenue is genuinely useful for relative, apples-to-apples comparisons within your own account, such as tracking whether a redesigned flow is outperforming the old version, or comparing similar campaigns against each other. Where it falls short is answering the absolute question of how much revenue your email program is truly causing, which is a different question that requires a different method to answer.

Stop Confusing Attributed Revenue With Real Revenue

Klaviyo isn’t misleading you on purpose. It’s answering a specific, narrow question, which touchpoint was closest in time to each purchase, and it’s answering that question accurately every single time. The mistake happens when that answer gets treated as though it’s answering a completely different question, namely how much revenue your email and SMS program is truly causing in a world where it didn’t exist.

Once your team understands the mechanics behind the number, from the default attribution windows to the channel overlap problem to what a healthy percentage of total revenue actually looks like, you can start making decisions with real confidence instead of borrowed confidence. Start by checking your own attribution window settings this week, then build toward running your first holdout test. The gap between what you think your email program is worth and what it’s actually worth is exactly the kind of gap that’s worth closing before it costs you a much larger budget decision down the line.

Ready to Turn Your Customers Into Loyal Buyers?