Solved: Catalog Ads Dominating? Attribution may be the problem

Hi there,

Thanks for reaching out! It sounds like you've put some serious effort into your catalog ads, which is more than most do, so kudos for that. It's not every day you see someone triple their attributed revenue from a single ad type.

But I've got to be brutally honest with you. When I see an account where 67% of revenue is coming from a bottom-of-funnel DPA (Dynamic Product Ad) campaign, my first thought isn't "wow, success!"... it's "blimey, that's a classic attribution trap."

I know that's probably not what you wanted to hear, but stick with me. I've seen this exact pattern dozens of times, and understanding what's really going on is the difference between thinking you're scaling profitably and actually doing it. You've done some great work on the creative front, now we just need to make sure it's pointed at the right target to actually grow your business, not just shuffle numbers around on a dashboard.

We'll need to look at The Great Attribution Illusion...

Let's talk about what's likely happening here. Your catalog ads are set up to retarget people who have already visited your site, viewed products, or even added items to their cart. These are people at the very, very end of their buying journey. They're already warm, they know your brand, they like your products. They are, for all intents and purposes, going to buy anyway.

Then, your DPA pops up in their feed, showing them the exact product they were just looking at. They click it, land back on your site, and complete the purchase they were probably going to make this evening anyway. And what happens? Meta's ad platform, in its infinite wisdom, slaps a big gold star on your DPA campaign and says, "I did that! All me!".

This is what we call "hoovering up conversions." The campaign isn't creating new sales; it's just getting the final touchpoint, the last click, before a sale that was already in the pipeline. It's like the cashier at the supermarket trying to take credit for your entire weekly shop. They were just the last person you interacted with, they didn't convince you to buy the milk, bread, and eggs.

This is the fundamental flaw with relying on default in-platform attribution. It tells you what got clicked last, not what actually influenced the decision. The real work was likely done by your top-of-funnel static ads that introduced them to the brand in the first place, or your email marketing, or a blog post they read two weeks ago. But the DPA gets all the glory because it's easy for the platform to measure.

I remember one client we worked with, an eCommerce brand selling custom maps and navigation tools. Their account looked a lot like yours. Their catalog ads were reporting an incredible 8x ROAS, generating about $71k in attributed revenue. The previous agency had been pouring more and more budget into it, thinking they'd found a magic money machine. But when we looked at their overall business revenue month-on-month, it was completely flat. The in-platform ROAS was going up, but the bank balance wasn't. The catalog ads were simply cannibalising sales from other channels and taking the credit. Shifting more budget there had zero impact on the overall top-line revenue. It was a classic case of mistaken correlation for causation.

I'd say you need to measure what actually matters: Incrementality...

So, if you can't trust the numbers inside Ads Manager, what can you trust? The answer is a concept called incrementality. The only question you should care about is: "How many sales happened because of this ad that would not have happened otherwise?"

Big brands with huge budgets use something called a Conversion Lift Study, which Meta offers, to measure this properly. But you don't need a massive budget to get a reliable answer. You can run your own version, which we call a holdout test. It's the single most powerful way to understand the true value of any retargeting campaign.

The logic is simple: you take a small portion of your audience (say, 10-20%) and deliberately *prevent* them from seeing your ads. Then you compare their buying behaviour to the 80-90% who *do* see the ads. If the group that saw the ads converts at a significantly higher rate, then you know your ads are having a real, incremental impact. If the conversion rates are pretty much the same... well, then you know your ads are just taking credit for sales that were already baked in.

Setting this up is a bit of a faff, but it's worth its weight in gold. Here's how you'd do it for your website visitors:

1. Create Your Full Audience: In Meta's Audiences section, create a Custom Audience of all your website visitors from the last 30 days. Let's call it "All Visitors - 30D".
2. Create Your Holdout Audience: This is the tricky bit. You need to create a *different* custom audience, but using a rule that captures a random slice. One simple way is to create an audience of "All Visitors - 30D" whose URL visited contains "utm_source=google" (or any other parameter that roughly half your audience might have). This isn't perfect randomisation, but it's a good enough proxy. A better, more technical way involves firing a specific event for a random subset of users, but let's keep it simple for now. Call this "Holdout Group - Approx 10%". You'll need to play with the rules to get the size right.
3. Set Up Your Campaign: In your catalog ad campaign's ad set, target the "All Visitors - 30D" audience.
4. EXCLUDE The Holdout: This is the most important step. In the same ad set, you must then EXCLUDE the "Holdout Group - Approx 10%" audience.

Now your catalog ads will run to most of your recent visitors, but a small, randomly-ish selected group will be shielded from them. Let this run for a few weeks, long enough to get statistically significant data. Then, you need to do the maths.

You'll need to pull the data from your Shopify or analytics platform. You need to know:

• How many unique users were in the group that saw ads, and how many of them purchased?
• How many unique users were in the holdout group, and how many of them purchased?

If you see a big difference (what we call 'lift'), then your ads are working. If the numbers are close, you're just paying to preach to the choir. This simple test gives you the truth that no attribution report ever will.

You probably should re-think your creative strategy...

Now, let's talk about the excellent work you did on your creative. Enriched SKUs, custom backgrounds, adding video... that's fantastic stuff. You've clearly proven that better creative gets better results. The mistake is keeping that high-powered creative locked up in a bottom-of-funnel campaign.

Think about it. The people seeing your DPAs are already sold. They've been on your site, they've seen the product. A nicer background is good, but it's not the thing that's going to tip them over the edge. Their purchase intent is already at 99%.

Where that amazing creative would have a massive impact is at the Top of the Funnel (ToFu). When you're trying to capture the attention of someone who has never heard of your brand before, who is just scrolling through their feed looking at cat videos. That's where a stunning video or a beautifully designed image makes all the difference. That's where creative has to do the heavy lifting of stopping the scroll, creating desire out of thin air, and convincing someone to click and learn more.

My advice would be to take your best-performing, enriched catalog assets and start testing them in your prospecting campaigns. Use them in carousel ads or Advantage+ Shopping Campaigns targeting broad audiences or lookalikes. You've already identified the creative that resonates with your warmest audience; there's a very good chance it will also be effective at attracting new customers who look just like them. You've built a high-performance engine; it's time to put it in a race car, not a golf cart.

Your current setup is like having a world-class salesperson, but you've tasked them with only greeting people as they walk out the door. Let them get out there and find you some new business!

You'll need a healthier marketing mix...

A profitable, scalable advertising strategy isn't about finding one magic campaign type and pouring all your money into it. It's about building a balanced ecosystem where each part of the funnel does its job effectively. Right now, your ecosystem is all bottom and no top, which is unsustainable. You're harvesting the fruit but not planting any new trees.

A healthier structure would look something like this:

• Top of Funnel (ToFu) - Awareness & Demand Generation (Approx 50-60% of budget): This is your growth engine. The goal here is to reach new, relevant audiences who have never heard of you. You'd use your best static and video ads, and your enriched catalog creative in carousels, to target broad audiences, interest groups, and lookalikes of your best customers. The key metric here isn't ROAS; it's driving high-quality traffic, low Cost Per Click (CPC), and building up your retargeting pools.
• Middle of Funnel (MoFu) - Consideration & Nurturing (Approx 20-30% of budget): This is for people who have shown some interest but aren't ready to buy yet. You'd retarget website visitors (who haven't added to cart), video viewers, and social media engagers. The creative here could be different – maybe customer testimonials, reviews, or ads that overcome common objections.
• Bottom of Funnel (BoFu) - Conversion & Closing (Approx 10-20% of budget): This is where your current catalog ads live. They are incredibly effective for a very specific job: closing the deal with people who have shown strong purchase intent (e.g., added to cart, initiated checkout). Their budget should be proportionate to this small, high-intent audience, not 67% of your entire spend. You should also be excluding recent purchasers to avoid annoying them and wasting money.

By rebalancing your budget like this, you create a sustainable system. The ToFu campaigns constantly feed new potential customers into the MoFu and BoFu campaigns. You stop relying on the same small pool of retargeting traffic and start actively growing the total number of people who know and consider your brand. Your overall revenue will start to climb, instead of just your attributed ROAS.

The ultimate goal is to move from obsessing over channel-specific ROAS to looking at your Blended ROAS, often called Marketing Efficiency Ratio (MER). This is the simplest and most honest metric: Total Revenue / Total Ad Spend. If that number is healthy and climbing, your advertising is working. If it's stagnant, something is wrong, no matter what the Facebook dashboard tells you.

I'd say you should focus on the numbers that build a business...

This brings us to the most important metric that most eCommerce businesses completly ignore: Customer Lifetime Value (LTV). Knowing your LTV changes everything. It tells you exactly how much you can afford to spend to acquire a customer (your Customer Acquisition Cost, or CAC) and remain profitable.

Once you know this, you're free from the tyranny of trying to make every single ad profitable on the first purchase. You can confidently spend money on top-of-funnel campaigns knowing that even if they don't convert immediately, you're acquiring a customer who will be worth 3x, 5x, or even 10x your initial investment over their lifetime.

The calculation is simpler than you might think. You just need a few key numbers from your business.

• Average Revenue Per Account (ARPA): How much a customer spends on average per order, or per month if it's a subscription.
• Gross Margin %: Your profit margin on that revenue after cost of goods sold.
• Monthly Churn Rate: The percentage of customers you lose each month. For non-subscription ecommerce, you can estimate this by calculating your 'repurchase rate'. For example, if 20% of customers make a second purchase within 3 months, your retention is 20% and your churn is 80% over that period. It's a bit of an art to calculate for non-subscription models, but getting a rough idea is better than nothing.

The formula is: LTV = (ARPA * Gross Margin %) / Monthly Churn Rate.

For instance, if your average order value is £100, your gross margin is 60%, and your monthly churn is 10% (meaning a customer sticks around for an average of 10 months), your LTV is (£100 * 0.60) / 0.10 = £600.

A healthy business typically aims for an LTV:CAC ratio of at least 3:1. So, with a £600 LTV, you can afford to spend up to £200 to acquire a new customer. Suddenly, that ToFu campaign that's bringing in new customers at a £50 CPA doesn't look like a "low ROAS" campaign anymore; it looks like an incredibly profitable investment. This is the maths that unlocks true scale.

I've built a little calculator for you below to play around with your own numbers. See how small changes in retention or margin can dramatically affect how much you can afford to spend on growth.

This is the main advice I have for you:

I know this has been a lot to take in, and it can feel a bit like I've just poured cold water on your success. That's not my intention. My goal is to give you the framework to build something far more scalable and genuinely profitable for the long term. You've clearly got a knack for the creative side; coupling that with a robust measurement and strategy framework would be unstoppable.

Here's a summary of what I believe your next steps should be, laid out in a simple table. This is the exact process we'd follow to move an account from being reliant on attribution to being driven by true, incremental growth.

Navigating this stuff is complex, there's no doubt about it. The ad platforms are designed to give you vanity metrics that make you feel good and keep you spending, rather than the hard truths that actually help you grow. It takes a contrarian mindset and a focus on first principles to cut through the noise.

You've already done the hard work of creating a great product and developing compelling creative. The next stage is about implementing a measurement and advertising strategy that reflects how customers actually behave, not how an attribution algorithm wishes they did.

If you'd like to chat through any of this in more detail and have a look at your account together, we offer a completely free, no-obligation initial consultation. We could walk through how to set up these tests properly and build a strategy that's geared for real, incremental growth.

Hope this helps!

Regards,

Team @ Lukas Holschuh