This week we're talking about a problem brands face every day without realizing it: cutting their best acquisition channels because ROAS says they aren't working. When in reality, the problem isn't the channel. It's that they're asking ROAS to do something it was never designed to do.
For a subscription business, what matters isn't what an ad returned today. It's what kind of customer it brought in, and how much that customer is worth over time. Welcome to The Halo. Every Thursday: a breakdown of how measurement actually works, a real client result, the best from our blog, and three links worth your time. Written for those who want to understand what's really happening with their marketing investment.
Welcome back to The Halo. In this issue we cover:
The Take: Why ROAS is the wrong metric for subscription brands, and what to measure instead.
From the Blog: Why your organic traffic isn't actually organic.
Discovery of the Week: MaryRuth's Pinterest looked weak. Until they measured it properly.
Prescient Voices: Digital spend doesn't stop working when the session ends.
Three Things: Links worth your time this week.
01 · THE TAKE
Why ROAS is the wrong metric for subscription brands
Subscription brands have a measurement problem that's easy to fall into, and hard to see from the inside.
They're measuring customer acquisition with ROAS. What they don't realize is that ROAS doesn't measure customer acquisition because it measures the return of a transaction.
For a subscription business, those are two completely different things.
A customer who subscribes and stays for 18 months is fundamentally more valuable than someone who buys once and never comes back. ROAS can't tell the difference. It sees a conversion, logs the revenue, assigns credit to the last touchpoint, and moves on.
The result is that subscription brands end up making budget decisions without knowing what kind of customer each channel is actually driving.
WHAT'S ACTUALLY HAPPENING
Unlike search or social channels where users are actively looking to buy, Pinterest is a discovery platform. People browse for inspiration, not to purchase. That's why ads tend not to convert in the first session, they plant a seed.
Someone sees an ad, doesn't click, and moves on. Days later, they search for the brand on Amazon, subscribe, and become a long-term customer.
That customer never shows up in Pinterest's reporting. It shows up in Amazon's reporting. And because ROAS only sees the last session, Pinterest gets no credit - not for the conversion, and not for the kind of customer it brought in.
A concrete example: when Prescient analyzed MaryRuth Organics' data, it discovered that 67% of Pinterest's total impact was invisible to platform reporting. Most of it landed on Amazon, search, and organic traffic, places platform reporting can't see.
You're not just using the wrong metric. You're missing 67% of the picture. And that's how subscription brands end up cutting their best acquisition channels.
A TEST YOU CAN RUN THIS WEEK
Pull your Pinterest spend over the last 90 days. Now look at the trend in new customers - not revenue, new customers - for the same period, with a 1-2 week lag.
If they move together, Pinterest is driving more acquisition than your dashboard shows. If both are moving down, that's worth investigating too, it may mean you've already started cutting a channel that was doing more than you realized.
Then ask: how long do those customers stay? What's their LTV (lifetime value) compared to customers from other channels?
If you can't answer that by channel, you're making budget decisions with an incomplete picture, and that's exactly how brands end up cutting their best acquisition channels without realizing it.
WHY THIS MATTERS AT PRESCIENT
Most MMMs measure revenue. They don't measure customers.
Prescient is the only MMM that can show you new customers acquired within the halo effect - customers who converted anywhere across your business after seeing a Pinterest ad they never clicked.
For a subscription brand optimizing toward a 3:1 LTV to CAC ratio, that's the difference between knowing what's working and guessing.
Our platform also uses campaign-level saturation curves to show marketers how far they can scale each campaign before losing efficiency, so budget decisions are based on predicted incremental lift, not last week's ROAS.
The revenue was always there and so were the customers. But it wasn't clear where new customers were coming from, as they were converting in places your platform couldn't see.
02 · FROM THE BLOG
Why your organic traffic isn't actually organic
Your analytics calls it organic. But most of it has paid roots - awareness campaigns that planted a seed weeks earlier, now showing up as non-branded organic search, branded search or direct visits. Linnea breaks down the specific paths from paid to organic, why standard dashboards hide that connection by design, and what it actually means for how you evaluate awareness spend.
6 min read
03 · DISCOVERY OF THE WEEK
MaryRuth's Pinterest profile seemed weak until they validated it outside the dashboard.
MaryRuth Organics runs a massive omnichannel business - Amazon accounts for roughly two-thirds of total revenue. When they looked at how their Pinterest campaigns were performing inside native reporting, the numbers were underwhelming. Direct attribution was low. It looked like a channel worth cutting.
They ran the same data with Prescient, and the picture changed completely. Only 33% of Pinterest-driven revenue was going where the platform could see it. The other 67% reflected halo effects, incremental impact that occurs when exposure on Pinterest drives conversions later in other channels that attribution systems don't directly connect back to the original ad. This includes 80% flowing into Amazon Selling Partner revenue, 18% into paid search on Amazon, 1.5% into organic search on TikTok Shop, and 0.5% into Shopify halo effects.
Pinterest wasn't underperforming. It was generating demand that was being captured elsewhere in the customer journey, primarily by Amazon.
With that visibility, they didn't cut Pinterest. They scaled spend 8x in November 2025. Pinterest maintained CAC efficiency nearly 50% more efficient than the portfolio average, and materially outperformed Microsoft Ads (92% lower), Amazon Ads (81% lower), Meta (73% lower), Google Ads (63% lower), and TikTok Ads (58% lower), while maintaining stable efficiency even at 8x spend.
The takeaway: the channels that are usually considered for awareness require further analysis because they are likely generating more revenue than what appears on your dashboard based on the Halo effect.
04 · PRESCIENT VOICES
Marketers already know journeys aren't linear. The problem is that intuition doesn't get budget approved. What changes the outcome is being able to show that same intuition, backed by data Finance can trust.
05 · THREE THINGS WORTH READING
1. What actually drives direct and organic traffic?
Direct and organic traffic tell you how visitors arrived - not what caused them to come. A significant portion traces back to awareness campaigns your dashboard can't connect to the original ad.
2. A practical guide to cross-channel marketing strategy
Your channels don't operate independently. Awareness campaigns lift branded search, direct traffic, and retail sales that never show up in platform reporting. If you're allocating budget channel by channel, you're working with an incomplete picture.
3. Your competitor's ad budget might be your best lead gen tool
When competitors run awareness campaigns, they build category demand that doesn't stay with them. Some of those searches end up on your site as organic traffic. The brands best positioned to capture it are the ones with strong organic presence and a tight bottom funnel.
The Halo
Exclusive insights, every week.
Subscribe to The Halo for sharper marketing thinking.