The Meta Ad We Almost Turned Off

I joined Hai Mag, CEO of Eva Commerce, for episode 2 of his Connected Commerce podcast. The premise of the show is that this stopped being an Amazon business a while ago, and most operating problems now live in the space between channels rather than inside any one of them.

I have a story that makes the point better than the theory does.

The Meta ad we almost turned off

I work as a fractional e-commerce director for a couple of brands at a time. That means sitting above the agencies rather than inside one of them, including above my own team, asking how all the vendors fit together and whether the overall direction makes sense.

On one of those brands, the person who runs Google and Meta came into our weekly call with a reasonable request. She wanted to cut the Meta ads for one listing. The spend was not returning, and she was not seeing conversions.

Everything about that is correct from inside Meta. If I had only been looking at her dashboard I would have agreed with her.

What I happened to know, because I am in the weeds on the Amazon side too, is that the listing she wanted to stop advertising was one of our top sellers on Amazon.

So the question changed. Instead of "why is this ad not converting", it became "where are these people actually buying".

I went through the usual checks first, the ones you would run on any non-converting ad. Is the price on the landing page the same as the ad. Is the cart working. Is anything broken between the click and the checkout.

Then I found it. This product had essentially no branded search history. And on Amazon, branded search for it had appeared.

People were seeing the Meta ad and going to search for the product on Amazon rather than buying on the website. There was also a small price difference that made Amazon the better deal, which quietly encouraged exactly that behavior. The volume was not enormous, but it was more than enough to matter, and it explained the zero on her report.

Had we turned that ad off, based on a completely defensible reading of the Meta account, we would have removed the thing feeding the Amazon momentum. And the Amazon team, sitting in their own silo, would have watched a top seller decline with no explanation. No listing violations. No bad reviews. Stock was fine. Nothing to point at.

That is how you fix one problem and create another without ever understanding your own business.

How do you actually measure the halo effect?

This is the hard part, and I want to be honest that nobody has it perfectly solved.

There are real tools. Amazon Attribution gives you a brand referral bonus on traffic you send from Google or Meta, and it works some of the time. Buy with Prime is another route. Those are the known methods.

But the signal I trust most is simpler. If a product has no branded search and then branded search starts appearing in your PPC data or your search query performance report, something off Amazon caused that. If you are also running Meta or TikTok, you know what it was.

The reason this is difficult in practice is rarely technical. It is organizational. If you manage Amazon and someone else manages Meta, you need that other vendor to share information and play nicely, and you need to understand both platforms well enough to find where they overlap. When a brand has separate agencies per channel who never speak, nobody is positioned to see it.

Amazon has rewarded off-platform traffic for a very long time. Before any of this was called connected commerce, people were running search find buy tactics, taking someone from off Amazon, having them search for the product and buy it, because the algorithm rewarded it. The mechanism is not new. What is new is that the white hat version, running TikTok Shop or Meta properly and letting the halo land, is now a legitimate core strategy rather than a launch trick. We treat it that way across our multi-marketplace work.

The goal is not perfect attribution. It is understanding what is driving what, knowing which parts you can track, and being honest about which parts are unknowns.

Why does price parity decide where the sale lands?

Hai raised the other side of this and it is worth spelling out.

If your Shopify price drops below Amazon, Amazon notices and can suppress your buy box. If your Shopify price sits above Amazon, customers who wanted to buy direct will simply buy on Amazon instead, and the direct to consumer team is left wondering why their conversion rate is falling.

Both of those look like a channel problem to whoever owns that channel. Neither is. In our own case the price gap was mild and had a purpose, but it was still routing sales in a direction nobody had explicitly decided on.

The point is not that gaps are always wrong. It is that you should not have something happening in your business and not know why. Once you can see it, you can be intentional: push the landing page to Amazon when that is what you want, or build an offer that genuinely earns the sale on your own site.

What does inventory have to do with any of this?

If I had to name the top three things to focus on, it would be inventory and supply chain three times. There is nothing worse than launching a good product and running out of stock, because you then spend months fighting back to where you already were.

What I would push people on is treating fulfillment as a place to be creative rather than a fixed cost.

Amazon's multi-channel fulfillment, where your FBA inventory fulfills Shopify or TikTok Shop orders, has been running a promotion at around 35% off for TikTok Shop fulfillment. I own a warehouse, so I would rather have those orders myself. But if moving a brand to MCF for three to six months takes a third off their shipping rates, that is the right call for the brand and I will make it.

What I usually land on is a hybrid. Some inventory in FBA, some FBM, some in AWD. Walmart and TikTok Shop are both pushing sellers toward their own fulfillment to get the most out of the platform, so the question becomes how complicated you are willing to make it, and whether the answer is best for margin or for customer experience. There is rarely one answer.

The pandemic is what taught me to keep options. FBM listings ranked higher for a period while FBA was overwhelmed and restricted to essentials, and anyone not already in a warehouse or 3PL struggled to get into one. Having only one way out is not a strategy. That thinking is why we run our own warehouse and 3PL.

Be creative in the warehouse

This is the part people skip, and it is where a surprising amount of connected commerce actually happens.

You can put an insert in the box that offers your Amazon customer something on a product they have not tried. You can move an Amazon buyer onto your email list, or use your email list to launch a product on Amazon. When TikTok Shop affiliates drive purchases, you can bring those buyers into a community you own rather than losing them in TikTok Shop DMs where you will never find them again.

All of that happens physically, in the warehouse. An insert, a sticker, a bundle, an influencer package.

Our operation in Kansas City is about 20,000 square feet, which is small on purpose. It means I can try something a hundred at a time, or fifty, or ten, and see whether it works. With a large 3PL that same test needs a new scope of work, a quote, and an approval. By the time you have that, the moment has passed.

I call the whole category post-purchase optimization, and it is one of the few areas left where a small brand has a structural advantage over a big one.

Frequently asked questions

How do you know if Meta ads are driving Amazon sales?

The clearest signal is branded search. If a product had little or no branded search and it starts appearing in your search query performance report or PPC data while you are running Meta or TikTok, that off-platform spend is creating the demand. Amazon Attribution and Buy with Prime help, but they capture only part of it.

Should I turn off an ad that shows no conversions?

Not before checking your other channels. An ad can show zero conversions on your own site while sending buyers to Amazon to complete the purchase, especially when a price gap makes Amazon the better deal. Switching it off can quietly remove the demand feeding a top Amazon seller.

Why does a price difference between Shopify and Amazon matter?

It decides where the customer finishes the purchase, and Amazon watches it. If your own site is cheaper, Amazon may suppress your buy box. If your site is more expensive, buyers who intended to buy direct will move to Amazon instead, and your direct to consumer conversion rate falls for reasons that have nothing to do with your site.

Is Amazon multi-channel fulfillment worth using instead of a 3PL?

It depends on the promotion and the moment. Amazon has been discounting TikTok Shop fulfillment through MCF at around 35%, which can be the right call for a brand for several months. Most operations end up hybrid, with inventory split across FBA, FBM and AWD so that no single failure removes every option.

What is post-purchase optimization?

It is everything you do after the customer receives the product: inserts offering another item, moving an Amazon buyer onto your email list, or bringing TikTok Shop affiliate buyers into a community you own. Most of it happens physically in the warehouse, which is why a smaller operation that can test a hundred units at a time has an advantage.

What to take from this

1. Before you cut an underperforming ad, check the other channels. Especially if the product it promotes is doing well somewhere else.
2. Use branded search as your early halo signal. New branded search on a product that never had any means an off-platform channel is working.
3. Get your channel owners into the same conversation. Not a shared dashboard, an actual conversation, because the insight lives in the overlap.
4. Audit your price gap between Amazon and your own site. Decide where you want the purchase to land instead of letting the gap decide.
5. Run one warehouse test this month. An insert or a bundle, a hundred units, and measure what comes back.

Thanks to Hai Mag and the Eva Commerce team for having me on. If your channels are each performing fine on their own report and the overall business still is not growing, the answer is usually sitting in the space between them.