Return on Spend Is Core to Zembula's Long-Term Success

Every marketing dollar gets measured eventually. The question is whether you measure it on the way out or after it lands. At Zembula, we chose to measure ourselves the same way our customers’ CFOs measure their ad spend: return on every dollar invested. Not impressions. Not opens. Revenue back per dollar forward.

This post originally explained that philosophy. Since then, a lot has changed in the market around us, and the case for measuring Zembula on customer return on spend has only gotten stronger. Paid-ad economics have compressed. Attribution on paid channels has gotten fuzzier. And email, which never depended on third-party cookies or platform-mediated identity, keeps producing clean, first-party performance data. Here’s where things stand.

Why Zembula Measures on Customer Return on Spend

The email industry has a favorite stat: $36 in ROI for every $1 spent, per Litmus. It’s a real number, and it’s impressive. But the denominator is ESP subscription cost, which is typically a tiny line item in your marketing budget. Measuring email’s value against your Klaviyo or Braze invoice is like measuring your warehouse’s value against the cost of the padlock on the door.

We reject that framing. The real resource being spent is subscriber attention and inbox space. That’s finite, it degrades with overuse, and it’s the actual constraint on email revenue. Zembula reports customer return on spend on a per-day basis, the way performance marketers report on paid ads, with 7-day click-based revenue attribution at the block level. The denominator is what customers pay Zembula, not what they pay their ESP.

This framing matters because it puts us on the hook. If our Smart Banners and Smart Kickers don’t generate enough click-attributed revenue to justify the spend, we fail by our own metric.

What Changed: Paid-Ad ROAS Fell While Email Economics Held

The macroeconomic backdrop for marketing spend shifted meaningfully. According to Rule1’s ROAS benchmark analysis, the average ecommerce ROAS dropped to 2.87x in 2025, down 4% year over year. That’s the average. The median sits at 2.04x, meaning half of ecommerce businesses are operating below a 2:1 return on ad spend. Once you subtract COGS, shipping (roughly $28 on a $120 order), payment processing fees, and ad operations overhead, the contribution-margin math turns negative quickly at sub-3x ROAS.

On the cost side: Meta CPMs climbed roughly 20% year over year, and Google CPCs rose 12.88% while conversion rates dropped 9.28%. As Rudy’s AI benchmarks put it: “The 10.03% decline in median ROAS reflects increasing ad costs across the platform. With CPCs rising 12.88% and conversion rates dropping 9.28%, advertisers are paying more for less.”

The measurement side got worse too. iOS App Tracking Transparency now suppresses visibility on a significant share of iOS conversions. Opt-in rates sit between 15% and 30% globally, which means ad platforms are making optimization decisions on incomplete data. Email doesn’t have this problem. Every click, every conversion tied to that click, is captured on first-party identity that you own.

How Zembula Measures: 7-Day Click-Based Attribution at the Block Level

Our primary measurement methodology is a 7-day click-based revenue attribution on the specific content Zembula powers. This means we track clicks on each Smart Banner and Smart Kicker variant, then attribute revenue from purchases that occur within seven days of that click.

This is the same measurement window Meta uses as its default attribution setting. We chose it deliberately so customers can compare Zembula performance against their paid-media dashboards in the same language, with the same rules.

The attribution runs at the block level, not the campaign level. That means a customer can see which specific personalization use case (abandoned cart, loyalty reminder, browse abandonment, price drop) is generating the most revenue per impression. It’s creative-level insight, updated daily, modeled on the kind of reporting ad teams take for granted but email teams rarely get.

How We Verify: Longitudinal Audience Testing (and When It’s the Wrong Tool)

Click attribution is excellent for daily monitoring and creative optimization. But to validate that Zembula is causing incremental revenue (not just claiming credit for it), we run longitudinal audience tests.

In a longitudinal test, we split the audience 50/50. Half sees Dimensions-powered content. Half sees the email without it. Audiences are assigned at open time (not send time), which eliminates open-rate differences between the groups and reduces noise. The test runs for several weeks with the audiences held constant, and new subscribers are distributed equally between groups as they join.

At the end, we check for 95% statistical significance and compare revenue and transactions between the two groups. We provide the raw audience data to customers so they can verify independently.

Here’s the part most vendors won’t say out loud: longitudinal testing is the wrong tool for many programs. If your audience isn’t large enough to achieve statistical significance within a reasonable timeframe, the test produces a result that looks precise but isn’t. We tell customers honestly when the audience math doesn’t support it. For those customers, sustained directional attribution across millions of impressions, tracked daily, is enough to build a program on. Chasing statsig you can’t achieve wastes time and resources.

What Customers Earn with Zembula: 11x to 41x Return on Spend

Across our customer base, Zembula Dimensions delivers 11x or greater return on spend. At the top end, J.Crew achieved a 41x return on spend with 19 personalized use cases live across their daily email program, launched in six weeks. As J.Crew’s Director of Performance Marketing Daryn Foster put it: “The level of revenue performance demonstrated in such a short time by Zembula was very impressive. Technical implementation was fast, and their support has been great.”

These returns come from the long tail of personalized use cases: abandoned cart, abandoned browse, loyalty status, shipment tracking, price drops, back in stock, birthday, welcome. Each use case targets a specific behavioral signal, and each gets its own performance data. The compounding effect of running 15+ use cases simultaneously through a single Smart Banner and Smart Kicker placement, with open-time decisioning selecting the highest-priority message for each individual, is what produces the outsized returns.

For context on what these numbers mean relative to the market, our email performance benchmark report breaks down click-to-conversion rates and revenue attribution across use cases.

The Next Dollar Question: Where Email Beats Paid Ads on Contribution Margin

The real comparison between email and paid ads isn’t platform-cost ROI versus ROAS. It’s contribution margin per impression.

Paid ads carry incremental media cost on every impression. Each additional thousand impressions on Meta costs $10.88 in CPMs (and rising). Each Google click costs $5.26 on average. The audience is rented, and the price of that rental goes up as more advertisers compete for the same inventory.

Email runs on owned audience with zero incremental media cost per impression. You already have the subscriber. The email is already being sent. The only incremental cost is the personalization layer, which is what Zembula provides. This means a 3x return on email spend is structurally worth significantly more than a 3x return on paid ads, because there’s no media cost eating the margin underneath.

This is the math that should inform where the next marketing dollar lands. If your paid program is running at 2.87x ROAS (the industry average) and your email personalization is running at 11x+ return on spend, even redirecting 1% of ad budget into email personalization represents a meaningful efficiency gain. That’s a CFO-defensible reallocation, not an act of faith.

The ad industry already knows email data is its most important input. Meta Custom Audiences, Google Customer Match, CDP identity graphs from Segment and LiveRamp, lookalike modeling: all of these are seeded from first-party email audiences. The irony is that the ad ecosystem runs on email data, but the email channel itself has been systematically underfunded relative to its contribution.

What Zembula Is Building Toward: The Most Efficient Next Dollar in Marketing

We want Zembula to be the most efficient next dollar you can spend in marketing. That’s the north star, and everything we build serves it.

Our per-day performance reporting gives CRM teams the same measurement rigor that ad teams use, so both can speak the same language in budget conversations. Our click-to-conversion reporting gives the CFO a metric they actually care about. Our longitudinal testing gives the skeptic a holdout methodology they can trust. And our open-time decisioning across 100+ behavioral use cases means the system keeps getting smarter as more data flows through it.

As paid-ad ROAS continues to compress and ATT-degraded attribution gets noisier, owned-channel measurement doesn’t just keep pace. It widens its lead. The advertiser who can prove their email channel produces $15 of revenue per dollar spent on personalization, with clean first-party attribution, is going to win the budget conversation against a paid channel that produces $2.87 per dollar with 30-40% of conversions invisible.

That’s the future we’re building for.

Key Takeaways

• Zembula measures itself on customer return on spend, not the $36-per-$1 ESP-cost ROI that the industry cites. The denominator matters.
• Average ecommerce paid ROAS fell to 2.87x in 2025, with the median at just 2.04x. Google CPCs rose 12.88%, and iOS ATT continues to suppress conversion visibility on paid channels.
• Zembula uses 7-day click-based attribution at the block level, the same attribution window as Meta’s default, so customers can compare email personalization against paid media using the same rules.
• Longitudinal holdout testing validates incremental lift, but we’re transparent about when the audience math doesn’t support statistical significance. Directional attribution across millions of impressions is a valid foundation.
• Customers see 11x to 41x return on spend. J.Crew achieved 41x with 19 personalized use cases launched in six weeks.
• Email’s unit economics are structurally better than paid ads. Zero incremental media cost per impression means a 3x email return outperforms a 3x paid return on contribution margin.
• The next marketing dollar should land where the math is best. For most enterprise retailers, that’s email personalization, and the gap is growing.