Attributed Revenue vs. Moving the Needle

It’s a frustration many ecommerce store owners will recognise. Your marketing dashboards tell you things are going brilliantly. Different platforms claim impressive amounts of revenue, with graphs heading reassuringly upwards. Yet you look at the business as a whole and wonder: where is all this success?

One platform attributes revenue to paid search, another to paid social, another to email — and each may be making a perfectly legitimate claim according to its own attribution rules.

The problem comes when attributed revenue is interpreted as additional revenue. A dashboard can tell us which marketing activity gets credit for a sale. It doesn’t necessarily tell us whether that activity added revenue — or profit — to the business.

I came across a particularly striking example recently while reviewing a client’s Klaviyo account.

At first glance, you might look at this Klaviyo dashboard and conclude:

“Klaviyo generated $642,000 in revenue.”

Look more closely, however, and Klaviyo is quite specific: it reports $642,222 in attributed revenue, representing 37.29% of total ecommerce revenue.

That distinction matters.

Klaviyo isn’t claiming that email created $642,222 of revenue that would otherwise never have existed. It’s applying its attribution rules to determine which orders should receive credit from email.

And the wider reporting looks impressive too. Klaviyo attributes $405,382 to automated flows and $236,840 to campaigns, with several measures showing growth of more than 999% compared with the previous year. Campaign performance also benchmarks strongly against similar businesses.

There is plenty here to suggest that email is performing well.

But it still leaves a much more important question.

Did it actually move the needle?

Did the business actually generate substantially more revenue — and ultimately more profit — because of its email marketing? Or is Klaviyo receiving attribution for revenue that might have occurred anyway?

Those aren’t the same thing.

And this isn’t specifically a Klaviyo problem. Every marketing platform needs some way of measuring performance, whether that’s Meta, Google Ads, an email platform, an affiliate network or something else. Attribution is how they do it.

Attribution tells us who gets the credit

If someone receives or engages with an email and subsequently places an order within Klaviyo’s defined attribution window, Klaviyo may attribute that order to the email.

That’s useful. Without attribution, we’d have very little immediate feedback about which campaigns, flows and channels appear to be contributing to sales. It gives us a consistent way to compare performance, identify patterns and decide where to investigate further.

But attribution is a model for assigning credit. It isn’t a scientific experiment proving what caused the purchase.

A customer receiving an abandoned-cart email AND subsequently placing an order does not necessarily mean they placed the order BECAUSE of the abandoned-cart email.

They may have come back anyway. The email may have reminded them, the discount may have persuaded them, or they may have been waiting until payday. Perhaps they saw a Meta ad in the meantime, searched for the product on Google, or simply returned to a browser tab they’d left open.

Klaviyo can’t know the counterfactual: what would that customer have done if the email had never been sent?

That’s the question attribution can’t answer.

AND is not the same as BECAUSE

This distinction becomes obvious when you take marketing out of the equation.

Imagine I made a major change to a client’s Klaviyo account on Monday and on Tuesday it started raining. The Klaviyo changes happened AND rainfall increased, but nobody would seriously conclude that it rained BECAUSE I changed Klaviyo.

The absurdity is obvious because we know those two things aren’t causally related. Marketing is more difficult because there is a plausible relationship: emails influence purchasing behaviour, ads generate sales and SEO brings customers to websites.

But a plausible relationship still doesn’t prove that every sale attributed to a marketing activity was caused by that activity.

Attribution is to incrementality roughly what correlation is to causation.

Or, in less marketing-ish language:

The dashboard tells you who got the credit. It doesn’t necessarily tell you what moved the needle.

Revenue can move between dashboards without the business moving

This becomes particularly important when you’re optimising an existing marketing programme.

Suppose an ecommerce business has a cart-recovery flow that normally receives $20,000 in attributed revenue each month. You introduce a second flow and the following month the original flow receives $12,000 while the new flow receives $10,000.

The new flow appears to have “generated” $10,000. But total attributed cart-recovery revenue has only increased from $20,000 to $22,000.

So did the new flow really add $10,000 to the business? Or did it add $2,000 while capturing $8,000 of revenue that would previously have been attributed to the original flow?

That’s the difference between creating additional value and moving attribution from one bucket to another.

The same thing can happen across marketing channels. A customer might click a Meta ad, receive a Klaviyo email, return through Google and eventually buy. Depending on the attribution models and windows being used, several platforms may legitimately claim credit for the same purchase.

None of those attribution claims necessarily has to be wrong. They are simply answering different questions according to different attribution rules.

But add them together and you can apparently “generate” far more marketing revenue than the business actually generated.

The attribution moved. That doesn’t necessarily mean the business did.

Attributed revenue and incremental revenue answer different questions

This is where the distinction between attributed and incremental revenue becomes useful.

Attributed revenue asks:

Which marketing activity should receive credit for this purchase?

Incremental revenue asks:

How much additional revenue occurred because of the marketing activity?

The second question is much harder because it requires us to consider what would have happened without the marketing intervention.

Suppose 100 customers purchase after receiving an email. If 90 would have purchased anyway and the email persuaded another 10 to buy, the platform might legitimately attribute revenue from all 100 customers to the email.

But only the additional 10 purchases are incremental — because those are the purchases that wouldn’t otherwise have happened.

That’s what actually moved the needle.

In a real ecommerce business, of course, we rarely know precisely which customers would have purchased anyway. That’s why incrementality is harder to measure than attribution.

But harder to measure doesn’t make it less important.

Attribution isn’t the problem

Attribution is useful. It’s fast, simple and usually built directly into the tools we’re already using. It helps us identify campaigns, flows and channels that appear to be performing well or badly, spot patterns and decide where to investigate further.

In many situations, it’s the most practical measurement available.

The problem begins when:

“Klaviyo attributed $100,000 of revenue”

quietly becomes:

“Klaviyo generated $100,000 of additional revenue.”

Both statements might sound like they’re describing the same thing. They’re not.

The first tells us how a marketing platform assigned credit for revenue that occurred. The second makes a much stronger claim: that without the marketing activity, that $100,000 of revenue wouldn’t have happened.

Attribution alone can’t establish that.

And the same principle applies whether we’re looking at email, paid search, paid social, SEO or any other marketing activity.

Attribution isn’t the problem. Misinterpreting attribution is.

So how do you know whether you’re moving the needle?

The strongest evidence usually comes from a properly designed control or holdout group: comparable customers are treated differently so we can measure what happens with — and without — the marketing intervention.

But ecommerce businesses can’t turn every marketing decision into a controlled experiment. We can still get much closer to the truth by looking beyond the attribution report and considering the wider commercial context.

I tend to work through it in this order:

Business activity → Opportunity → Marketing exposure → Purchase behaviour → Attribution → Incrementality

Start with the business itself. What happened to total revenue, orders and profit? Did traffic, conversion rate or average order value change?

Then consider the opportunity available. A cart-recovery programme can’t recover customers who never reached the cart. An email programme can’t generate more purchases from an audience that doesn’t exist. Changes in traffic, audience size and customer behaviour can dramatically change the opportunity available to marketing.

Next, look at exposure and purchase behaviour. Who actually received or interacted with the marketing, and what did they do afterwards?

Only then put the attributed revenue into context and ask the question that matters: is there evidence that the marketing created additional value, rather than simply receiving credit for revenue that might have happened anyway?

This still doesn’t prove incrementality without an appropriate control. But it gives us a much better basis for judging whether the marketing appears to be moving the needle for the business.

Look outside the marketing platform

This is particularly important when a marketing report shows spectacular growth.

If Klaviyo-attributed revenue increases by 50% while total ecommerce revenue remains unchanged, that’s worth investigating. It doesn’t automatically mean the Klaviyo work achieved nothing: perhaps email protected revenue while another channel declined, customer acquisition fell or trading conditions changed.

But equally, we shouldn’t assume that a 50% increase in attributed revenue means the marketing created 50% more value.

The wider business provides the context. That’s why performance and measurement needs to extend beyond Klaviyo attribution alone. Did total revenue grow? Did you acquire more valuable customers? Did conversion or repeat purchase behaviour improve? And ultimately, did profit improve?

Because the objective isn’t to make the graph inside a marketing platform go up.

It’s to make the economics of the business improve.

Which brings me back to that $642,000

The Klaviyo reporting I showed at the beginning is real. Klaviyo really did attribute more than $642,000 of revenue to email, alongside some very strong engagement, conversion and revenue-per-recipient benchmarks.

Those are encouraging numbers. But they don’t tell me that Klaviyo added $642,000 to the business.

To establish whether the work genuinely moved the needle, I need to look outside Klaviyo: at what happened to the business overall, how much opportunity was available and how performance changed compared with an appropriate period before the work was done.

Even then, without a proper control group, I should be careful about claiming causation.

That’s not being pessimistic about email marketing. It’s simply being honest about what the numbers can — and cannot — tell us.

Attribution is useful. But it’s not the finish line.

Marketing platforms need attribution, and ecommerce businesses need a practical way of understanding which marketing activities appear to be contributing to revenue.

So the answer isn’t to throw attribution reporting away. It’s to understand what it tells us — and what it doesn’t.

The next time a marketing dashboard tells you how much revenue was “generated”, look beyond the attribution figure and ask what happened to the business as a whole.

Did revenue actually grow? Did profit improve? Are you creating additional value, or simply assigning credit for revenue that might have happened anyway?

That’s the difference between reporting success and actually moving the needle.

Because ultimately, your business doesn’t bank attributed revenue.

It banks revenue.