Retail media vs. Meta
& Google: where DTC
dollars actually convert.

Here's the number that reframes the whole conversation: US retail media ad spend reaches roughly $71 billion in 2026, about 30% of all US digital advertising. That makes retail media the third major force in DTC media, sitting behind Google and Meta but no longer a rounding error next to them. And inside that pool, the concentration is extreme. Amazon controls close to 80% of it, with Walmart Connect a distant second near 8%.

So the question every DTC operator is now asking sounds simple: should I move budget out of Meta and Google and into Amazon Ads and Walmart Connect? It's the wrong question, and the way it's usually framed (retail media versus Meta and Google, as if you have to pick a side) leads brands to make expensive mistakes in both directions. Some pile into retail media chasing the high reported ROAS and starve the demand engine that fed it. Others ignore retail media entirely and watch competitors intercept their own buyers at the moment of purchase.

I've allocated eight-figure annual budgets across all four of these channels at WIN Brands Group, and the thing I learned the hard way is that they don't compete for the same job. Meta and Google mostly create and steer demand. Retail media mostly captures demand that already exists. A sponsored Amazon ad that "converts" at a 6x return is often intercepting a sale that was going to happen anyway, which is a very different thing from a Meta ad that found a customer who'd never heard of you. Confuse those two and your dashboard will lie to you, profitably, for a long time.

The stakes on getting this right have climbed, because the dollars at play are no longer marginal. When retail media was a 5% line item, a misallocation cost you a little efficiency. Now that it's a third of digital and growing double digits a year, the channel you over- or under-fund determines whether you're building a durable brand or renting volume from a marketplace. The decision used to be a tactical knob; it's become a question about what kind of company you're building. That's the lens I'll carry through this whole comparison.

This is the head-to-head, written from inside the budget rather than from a vendor's pitch deck. Where each channel actually converts and why, what the reported numbers hide, how measurement differs, what the real margin looks like once marketplace fees are counted, when to shift budget and when not to, and how to split a real allocation across all of them. By the end you'll stop asking which channel is best and start asking the only question that matters: which channel is earning the growth I'm paying it for?

One promise up front. This is the comparison post, the head-to-head on where dollars convert. The separate question of whether your specific brand should be spending on retail media yet, the decision framework, the readiness checklist, lives in its own piece. Here we're settling the matchup. There we settle your move.

Retail media is now
the third force, and
it's wildly concentrated.

Retail media stopped being a niche the moment its spend crossed $71 billion and hit roughly 30% of US digital advertising in 2026. For context, that's grown from around $60 billion in 2025, and it's the fastest-expanding major category in the entire ad market. When a channel adds more than ten billion dollars of annual spend in a single year, it's no longer something a DTC brand can treat as optional experimentation. It's a pillar.

But the headline hides how lopsided the inside of that pillar is. Amazon's retail media business pulled in roughly $68.6 billion in ad revenue in 2025, growing about 22% year over year, which gives it close to 80% of the US retail media pool. Walmart Connect is the clear number two at around $6.4 billion and 8% share, and then the drop-off is steep: Target's Roundel sits near 1.5%, and everyone else fights over the scraps. When people say "retail media," they overwhelmingly mean Amazon, with Walmart as the only other player at real scale.

The concentration is getting worse, not better. Of every incremental retail media dollar spent in 2026, Amazon and Walmart together absorb more than 89%. So the growth isn't spreading across a healthy field of competing networks. It's pooling into two retailers who happen to own the two largest buying audiences in the country. For a DTC brand, that's both an opportunity and a warning: the audiences are enormous, but you're renting access to your own customers from a landlord who also competes with you.

Why did this happen now? Three forces converged. Privacy changes degraded the targeting that made Meta and Google prospecting so efficient, so advertisers went looking for first-party purchase data, and retailers were sitting on the richest pile of it. At the same time, the retailers realized that ad revenue is the highest-margin line on their P&L, and they got aggressive about monetizing the shelf. Amazon's ad revenue now equals roughly 8.3% of its marketplace GMV, a structural tax it collects on the brands selling there. Walmart's is around 4.3% and climbing fast. The shelf became a media channel, and the brands have to pay to be visible on it.

So when you frame the decision as "retail media versus Meta and Google," understand what you're actually weighing. You're not comparing three interchangeable ad platforms. You're comparing a demand-creation system (Meta, Google's broad reach) against a demand-capture toll booth (Amazon, Walmart) that sits exactly where the purchase happens. The toll booth posts gorgeous return numbers, because of course it does, it's standing at the cash register. The real question is how much of that return it actually created versus simply collected.

What Meta, Google, and
retail media are each
actually good at.

The cleanest way to think about these channels is by where they sit in the customer's journey. Meta is mostly demand generation: it interrupts people who weren't looking for you and creates want. Google splits the difference, capturing demand that already exists through search while also generating discovery through shopping and broader formats. Retail media is almost pure intent capture: it reaches a shopper who is already inside a store, wallet out, choosing between options. Different jobs, different stages, different math.

Meta is the demand engine. Nobody on Instagram is searching for your product. Meta's job is to interrupt a scroll, earn attention with creative, and plant a brand into a head that didn't know it existed an hour ago. That's why creative quality matters more on Meta than anywhere else, and it's why the channel is so dependent on the kind of native, scroll-stopping ad styles that actually hold attention. When Meta works, it's the only one of these channels reliably manufacturing new customers from nothing. When it's measured badly, it gets robbed of the credit for doing exactly that.

Google captures intent and seeds discovery. Someone typing "best magnesium supplement" into Google has already decided to buy something; the question is just which something. That's high-intent capture, and branded search in particular is often capturing demand that Meta or word of mouth created upstream. Google Shopping and the newer demand-generation formats push further up the funnel, but the core of Google's value for most DTC brands is being there at the exact moment a shopper converts intent into a query. It's a closer more than an opener.

Google is genuinely two channels wearing one logo, and conflating them is a common allocation error. Branded search (people typing your name) is almost pure harvesting of demand other channels created, and it converts at a rate that flatters the whole Google line. Non-branded search and Shopping reach in-market shoppers who don't know you yet, which is closer to demand capture against your category. Treating those as a single "Google ROAS" hides the fact that your branded terms are coasting on Meta's work while your non-branded terms are doing real acquisition. When operators tell me Google is their best channel, the first thing I check is how much of that performance is just branded search collecting a toll on awareness they paid for elsewhere.

Retail media captures intent at the point of purchase. A sponsored ad on Amazon or Walmart reaches a shopper who is not just intent-laden but physically inside the checkout environment. They've opened the store, they're searching the category, and your ad puts you at the top of the result. This is the highest-intent placement in all of digital advertising, which is exactly why it converts so well and why the ROAS looks so good. The catch, which we'll spend a whole section on, is that high conversion and high incremental conversion are not the same thing.

So is one channel better? Only in the way a hammer is better than a screwdriver, which is to say, only for its specific job. A brand with no awareness can pour money into Amazon and watch it underperform, because there's no demand to capture yet. A brand with strong awareness can ignore retail media and watch competitors intercept its buyers at the shelf. The mature answer isn't to rank the channels. It's to understand that Meta and Google build the demand that retail media then harvests, and to fund them as a system rather than a contest. That systems view is the heart of any serious channel mix strategy.

The channel scorecard:
intent, cost, margin,
and data ownership.

Here's the comparison laid out the way I'd lay it out for an operator deciding where the next dollar goes. The columns that matter aren't just cost and return; they're intent level, who owns the customer data, and what margin survives after the channel takes its cut. A channel can look cheap on CAC and still be expensive on margin once you count fees, and it can convert beautifully while owning the customer relationship that should have been yours.

Read that data-ownership column carefully, because it's the one founders discount and later regret. When you acquire a customer through Meta or your own Google campaign that lands on your Shopify store, you get the email, the behavioral data, the ability to retain and resell them at near-zero acquisition cost. When you "win" that same customer on Amazon, Amazon keeps the relationship. You got the unit sale and a sliver of margin; Amazon got a customer it can market to forever. That asymmetry doesn't show up in ROAS, and it's worth real money over a customer's lifetime.

The CAC story is also more nuanced than the dashboards suggest. Reported cost per acquisition on retail media usually looks lower than Meta prospecting, sometimes dramatically. But that's because retail media is converting people who were already going to convert. Your true blended cost to acquire a new customer, the one who'd never have found you otherwise, is almost always on Meta or upper-funnel Google, and it's higher. If you only optimize toward the cheapest reported CAC, you'll defund the exact channel that creates the customers, which is the most common self-inflicted wound I see in DTC media. Getting this right means knowing your maximum allowable CAC as a real ceiling, not chasing whichever channel reports the prettiest number.

The margin column is where retail media's shine dulls the most. A Meta-acquired customer who checks out on your own store costs you the ad spend plus your normal fulfillment. An Amazon-acquired customer costs you the ad spend plus Amazon's referral fee, often 15%, plus fulfillment fees if you use FBA. Two sales at identical revenue and identical reported ROAS can leave wildly different dollars in your pocket, and the retail media version leaves less. We'll quantify that in the margin section, but keep it in mind every time a retail media ROAS makes you want to shift budget.

One column I'd add if the table had room is competitive density, because it quietly reprices everything else. Meta and Google auctions are crowded but the inventory is effectively unlimited; there's always another impression. Retail media auctions sit on top of a finite shelf, so when your competitors decide to defend the same keyword, the CPC on a single category term can double in a quarter. The cost lines in that scorecard aren't fixed; they drift with how aggressively the rest of your category is bidding. A channel that looked cheap when you entered can become expensive purely because three competitors woke up to the same opportunity, which is why I treat retail media CPCs as something to re-baseline every quarter rather than set once.

The ROAS illusion:
captured demand wearing
a halo.

Reported retail media ROAS routinely looks 2x to 4x stronger than Meta prospecting, and that gap is the single most misleading number in DTC media. The reason is simple: retail media intercepts a shopper who already decided to buy your category, so it claims conversions that would have happened with or without the ad. High conversion, yes. High incremental conversion, often not. The dashboard credits the channel for demand it merely collected.

Picture the actual journey. A customer sees your brand on Instagram, ignores it twice, sees it a third time, gets curious, and a week later opens Amazon and searches your product name. Your sponsored ad sits at the top, they click, they buy. Amazon's dashboard reports a beautiful conversion at a low cost. But who actually created that sale? Meta did the work; Amazon stood at the door and collected a toll. If you reallocate budget from Meta to Amazon based on the reported ROAS, you'll slowly choke off the demand that was feeding Amazon in the first place, and then watch the Amazon numbers mysteriously soften too.

This is the same attribution trap that distorts the Meta-versus-Google fight, just one level deeper. Branded search on Google captures demand that Meta created; retail media captures demand that the whole upper funnel created. The most accountable-looking channels are accountable precisely because they sit closest to the transaction, which is also exactly why they over-claim. The closer a channel is to the cash register, the more credit it steals from the channels that did the persuading.

There's a specific version of this that catches sophisticated operators, not just beginners: the defensive-branded-keyword trap on Amazon. You bid on your own brand name so a competitor's sponsored ad doesn't sit above your organic listing. The conversion rate is enormous, because these are people searching for you by name, and the reported ROAS can look like 10x or higher. It feels like a no-brainer. But run the holdout and you often discover that the vast majority of those shoppers would have scrolled two inches down and bought your organic listing anyway. You were paying a premium to convert demand you already owned, defending against a threat that mostly wasn't there. Some defensive bidding is genuinely worth it; far more of it is the ROAS illusion in its purest, most seductive form.

So how do you see through it? Incrementality testing, not last-click attribution. The only honest way to know what retail media actually added is to turn it off in a controlled way (a geo holdout, a category pause, a scaled-back test) and watch what happens to total sales, not just sales on that channel. If you cut Amazon ad spend 40% and total revenue barely moves, you were paying to convert demand you already had. If total revenue drops hard, the channel was genuinely incremental. Brands spending real money (above roughly $50K a month per channel) should be running these holdout tests quarterly, because it's the only thing that cuts through the self-reported halo.

Measurement and
attribution: three channels,
three different truths.

The hardest part of comparing these channels is that they don't even measure the same thing, and none of them measure it honestly on their own. Most DTC purchases involve 3 to 5 touchpoints before checkout, yet every platform reports as if it owns the whole conversion. Meta over-credits view-throughs, Google over-credits last-click branded search, and retail media over-credits the in-store interception. Add the three dashboards together and you'll "find" 150% of your actual revenue.

Meta's measurement leans on view-through attribution, claiming credit for purchases from users who merely saw an ad without clicking it. That's not fraud, it's a defensible model for a demand-generation channel, but it inflates Meta's reported return and makes it look more like a closer than it is. Google's default last-click does the opposite distortion: when a customer sees a Meta ad and then converts through branded search, Google takes the whole conversion. Two platforms, two self-serving models, both pointed at the same sale.

Retail media is its own measurement island, and that's the part operators underestimate. Amazon and Walmart report inside walled gardens you can't easily reconcile against your own data, because the customer never touches your site. You see a sponsored-ad conversion in their console, but you can't tie it to your email list, your cohort data, or your first-party traffic and conversion data the way you can with on-site channels. The retailer hands you a number and asks you to trust it, and that number is naturally generous to the retailer.

So what does an operator actually do about it? You stop trusting any single platform's self-report and you triangulate. Multi-touch attribution tools that pull Meta, Google, Amazon, and Shopify into one model give you a directional view of the assisted journey. Your own blended math (total new customers divided by total ad spend across all channels) gives you a brutally honest north star that no platform can game. And incrementality testing, again, gives you the ground truth on what each channel actually added. The brands that allocate well in 2026 don't argue with the dashboards; they build a measurement stack that sits above all of them.

There's a deeper point hiding here about who controls the truth. On Meta and Google you at least own the conversion data on your own store, even if the platforms spin the attribution. On retail media you own almost nothing; the retailer owns the shopper, the data, and the reported result. So the measurement problem and the data-ownership problem are the same problem wearing two hats. The further your spend moves toward retail media, the more you're flying on instruments the retailer built and calibrated. That's not a reason to avoid it. It's a reason to never let it become the only channel you can see by.

The margin reality:
the tax of selling on
someone else's shelf.

Retail media's reported efficiency hides a margin tax that on-platform dashboards never show you. On Amazon, total advertising cost of sale (TACoS) commonly runs 15% to 25% of revenue for established brands, and that ad load sits on top of marketplace referral fees of around 15% plus fulfillment fees if you're on FBA. The same revenue dollar earned on Amazon keeps far less margin than one earned on your own store, even when the ROAS looks identical.

Let's make it concrete. Say you sell a $40 product. On your own Shopify store, an order costs you the ad spend plus your normal pick-pack-ship. On Amazon, that same $40 order pays an Amazon referral fee (frequently around 15%, so roughly $6), FBA fulfillment fees, storage, and then your advertising cost on top. Stack those and the marketplace can quietly take 30% to 40% of the sale before your COGS even enters the picture. A 6x reported ROAS on Amazon can leave you with thinner contribution dollars than a 3x ROAS on your own store, because the 3x version isn't paying the marketplace tax. This is exactly why I push brands to model the full contribution margin per channel rather than trusting a return-on-ad-spend headline.

TACoS is the metric that actually matters here, and most sellers calculate their break-even wrong. The common mistake is computing break-even ACoS off price instead of margin, which leaves brands tolerating an ad load 10 to 15 points higher than their economics can support. The honest version: take your real contribution margin after all marketplace fees, and that's the ceiling for what advertising can consume before the sale turns unprofitable. A TACoS under 15% is healthy; 15% to 25% is the normal operating zone for an established brand; above 35% you're usually buying revenue you can't afford.

Now compare the fee structures honestly. Meta and Google charge you for the click or the impression and nothing else; the margin you keep is your full price minus your own costs. Retail media charges you for the placement and takes a cut of the transaction through marketplace fees you'd never pay on your own site. The same advertising dollar buys a structurally worse margin outcome on retail media, which is the part the ROAS comparison erases. It can still be worth it, because the volume and intent are real, but only if you go in with the full-fee math, not the dashboard math.

The takeaway isn't that retail media is a bad deal. It's that "great ROAS" and "great margin" are different claims, and retail media usually wins the first while losing the second. A disciplined operator holds both numbers in view at once: the return that says the channel converts, and the contribution margin that says how much of that conversion you actually keep. The brands that get burned are the ones who saw the 6x, shifted half their budget, and only later noticed the bank account didn't grow the way the dashboard promised.

Amazon vs Walmart
Connect: not the same
retail media bet.

Retail media isn't one channel, it's at least two very different ones, and treating Amazon and Walmart as interchangeable is a mistake. Amazon is the mature, saturated, expensive incumbent: massive audience, fierce ad competition, and a CPC that reflects it. Walmart Connect is the faster-growing, less-saturated challenger, where ad revenue grew about 46% in 2025 to $6.4 billion, more than twice Amazon's growth rate, precisely because it's earlier in its monetization curve.

That growth gap tells you where the arbitrage is. Amazon's ad revenue already equals roughly 8.3% of its marketplace GMV, which means the shelf is heavily monetized and the auction is crowded. Walmart's sits around 4.3%, roughly half, which means there's still slack in the system: less advertiser competition, lower average CPCs, and more room to win share before the auction prices it in. For a brand with a product that fits Walmart's audience, Connect can deliver intent capture at a better cost than Amazon simply because fewer advertisers have crowded in yet.

The performance picture backs this up, with nuance. Walmart Connect has shown strong incremental results in the right categories, with health and beauty among the standouts, posting some of the highest reported ROAS on the platform. At the same time, Walmart's full-year 2025 ROAS softened by roughly 18% as CPC pressure built and average order values dipped, which is exactly what you'd expect as more advertisers discover the same arbitrage. The window where Walmart is meaningfully cheaper than Amazon is real, but it's a window, not a permanent state. The advertisers piling in are closing it.

There's a category-fit dimension that decides which network is even worth testing. Amazon is the default for almost everything, the place shoppers go to search any product category. Walmart skews toward its core demographics and strongest verticals (grocery, household, health and beauty, value-oriented categories), so a premium or niche brand may find its audience simply isn't shopping there in volume. The right move isn't "Amazon or Walmart" in the abstract; it's matching your product and price point to where your actual buyers already shop, then testing the network where the intent genuinely exists.

The operational burden also differs, and it's a real cost most channel comparisons ignore. Amazon and Walmart aren't just ad platforms; they're marketplaces with their own listing optimization, content requirements, review dynamics, buy-box rules, and inventory logistics. Winning on retail media means winning at retail media operations, which is a genuinely different skill set than running Meta creative or Google campaigns. A brand that's excellent at DTC paid social can still flounder on Amazon because its listings are thin, its reviews are weak, or its inventory keeps going out of stock and tanking its rank. The ad spend is the visible cost; the operational lift to make that spend convert is the hidden one, and it's the reason plenty of strong DTC brands underperform on retail media for a year before they figure it out.

If I'm advising a brand entering retail media today, the sequence is usually this: prove the Amazon listing converts organically first, because paying to drive traffic to a listing that doesn't convert just subsidizes Amazon. Then layer in sponsored ads to defend and expand the listings that already work. Then, if the audience fit is there, test Walmart Connect specifically to capture the cheaper intent before the auction catches up. Treating the two networks as one budget line, or worse, as one decision, leaves the Walmart arbitrage on the table and overpays into Amazon's crowded auction.

How the channels work
together instead of
against each other.

The brands that win in 2026 stopped treating this as an either-or and started treating it as a relay. Meta and upper-funnel Google create the demand; branded search and retail media catch it at the moment of purchase. Run as a system, the channels compound: every dollar of demand Meta creates becomes cheaper, higher-converting volume for Google and Amazon to harvest downstream. Run as a contest, you defund the front of the relay and wonder why the back of it slows down.

The mechanism is the part founders miss. When Meta builds brand awareness, two things happen on your retail media performance. Your branded searches on Amazon climb, because people who saw you on Instagram go looking for you in the store, and those branded conversions are the cheapest, highest-converting ad units in retail media. And your overall listing conversion improves, because shoppers arriving with prior brand familiarity buy at a higher rate than cold category browsers. Meta doesn't just create its own sales; it makes your retail media spend work harder. That cross-channel lift never shows up in either platform's dashboard, which is exactly why the relay gets misunderstood.

This is why the cut-Meta-feed-Amazon move backfires so reliably. A brand sees Amazon posting a 6x and Meta posting a 1.8x, concludes Meta is inefficient, and shifts budget. For a quarter the blended numbers look fine, because there's a reservoir of existing demand still flowing into Amazon. Then the reservoir drains. With nothing refilling the top of the funnel, branded searches flatten, cold-category competition on Amazon gets more expensive, and the once-beautiful retail media ROAS quietly erodes. The brand didn't find efficiency; it ate its seed corn.

The healthy pattern runs the other direction. You fund demand generation deliberately, accept that its reported ROAS will always look worse because it's doing the hard work of creating customers, and you let retail media and branded search harvest the demand at high efficiency. The blended economics of the whole system are what you optimize, not the line-item return of any single channel. This is the same logic behind serious CAC payback analysis: you measure the system's ability to acquire and pay back customers, not whichever channel reports the lowest cost in isolation.

I watched this play out almost exactly at WIN. On more than one brand, the team would get excited about a retail media line posting a return Meta could never match, and the instinct was always to feed the winner. The discipline we built was to ask a different question first: if this channel is so efficient, what's actually filling it? Nine times out of ten the answer traced back to a demand-generation push that had been running for months. The channel posting the trophy ROAS was almost never the channel doing the persuading. Once you internalize that, you stop celebrating the harvester for the farmer's crop, and you start protecting the part of the system that's quietly doing the expensive work.

One practical way to operationalize the relay: judge upper-funnel channels on their contribution to total new-customer growth and blended CAC, and judge lower-funnel channels on efficiency. Don't hold Meta to Amazon's return standard, and don't let Amazon's return convince you it's doing Meta's job. When you grade each channel against the job it's actually built for, the allocation decisions get a lot clearer, which is what the next section is about.

The allocation framework:
how to split a real
budget across all four.

There's no universal split, but there is a sound way to build one. Start from the funnel, not from the platform. A useful default for a demand-dependent DTC brand puts roughly 50% to 60% of budget into demand generation and upper-funnel capture (Meta plus broad Google), and 40% to 50% into high-intent harvesting (branded search plus retail media), then tunes that mix to your awareness level, category, and margin. The split moves with brand maturity, not with whichever channel posted the best ROAS last week.

Notice what's constant across all three stages: demand generation never goes to zero, and retail media's role grows only as fast as the demand it can harvest. The mistake is reading those stage tilts as a license to cut the front of the funnel once the back is performing. The percentages shift, the relay stays intact. A brand that lets demand generation drift toward zero in Stage 3 is one bad quarter away from discovering that its "efficient" retail media was running on borrowed demand all along.

Two adjustments override the default. First, margin: if your category carries thin contribution margins, the marketplace fee load may make heavy retail media spend unprofitable no matter how good the ROAS looks, so you tilt back toward your own-store channels where you keep more of each dollar. Second, data strategy: if owning the customer relationship and the email list is core to your model (subscription brands, high-repeat categories), you deliberately weight toward channels that build your first-party base even at a worse reported return, because the lifetime value of an owned customer dwarfs the one-time margin on a marketplace sale.

And the override that trumps all of them: incrementality. Whatever the framework suggests, the holdout tests tell you the truth. If a quarterly test shows your retail media spend is mostly harvesting demand you'd capture anyway, the framework says shift those dollars back upstream regardless of the dashboard ROAS. The allocation isn't a fixed recipe; it's a hypothesis you keep pressure-testing against what actually moves total revenue. That discipline (build from the funnel, then let incrementality correct you) is the whole game.

The honest tradeoffs:
brand control and
who owns the customer.

Even when retail media's economics work, it carries two costs that never appear on any dashboard: you give up brand control and you give up the customer relationship. On your own store you control the experience, the data, the merchandising, the email follow-up, and the lifetime value. On a marketplace you control a product listing inside someone else's environment, and the customer you "won" belongs to the retailer. Those tradeoffs don't show up in ROAS, but they show up in enterprise value.

Start with brand control. Your Amazon listing lives inside Amazon's template, next to your competitors' identical-looking listings, with Amazon's recommendations steering shoppers toward alternatives, sometimes toward Amazon's own private-label version of your product. You can't build the kind of differentiated experience that justifies a premium, and you're one algorithm change away from a very different result. The brands that lean hardest into retail media often find they've become a commodity SKU in a search result rather than a brand a customer chose, which erodes exactly the pricing power that makes a DTC business valuable.

Then there's the data and relationship cost, which compounds over years. A customer acquired on your own store enters your world: you have their email, their purchase history, their behavior, and the ability to retain and resell them at near-zero cost. That owned base is a huge part of what a buyer pays for when a brand sells; durable, first-party customer relationships are an asset. A customer acquired on Amazon never enters your world at all. You got the unit margin once, and Amazon kept the relationship to monetize forever. Shift too much of your acquisition onto retail media and you can grow revenue while quietly hollowing out the asset that actually makes the brand worth buying.

This is where retail media's convenience becomes a strategic trap if you're not deliberate. It's easy, it converts, the ROAS looks great, so a brand keeps tilting toward it, and three years later it's a marketplace-dependent business with no email list, no brand equity, no pricing power, and a customer base it doesn't own. That's a structurally less valuable company than one of the same revenue built on owned demand, and acquirers price it accordingly. The same channel-concentration risk that a profitability teardown would flag in diligence is the one a brand walks into voluntarily by over-indexing on retail media.

None of this is an argument against retail media. It's an argument for using it with your eyes open. Retail media is a powerful intent-capture channel and an enormous one, and ignoring it cedes your buyers to competitors at the shelf. But treat it as one instrument in a system you control, not the system itself. Keep building owned demand, keep growing your first-party base, keep retail media in its lane as a harvester rather than the whole farm. The brands that hold that discipline get the volume without surrendering the value.

So where do DTC dollars actually convert? Retail media converts highest, because it stands at the cash register, but a large share of that conversion is demand someone else created. Meta and Google convert lower on paper, because they're doing the harder, earlier work of creating customers who didn't exist yet. The honest answer to "retail media versus Meta and Google" is that they're not rivals, they're a relay, and the brands that allocate well in 2026 fund the whole relay rather than chasing the channel with the prettiest single number. Retail media earns its place. It just shouldn't be allowed to crowd out the demand engine that makes it work, or quietly take ownership of the customers that make your brand worth owning.

If you're staring at a budget and trying to decide whether the next dollar belongs on Meta, Amazon, or Walmart, that's exactly the allocation work I've done across an eight-figure portfolio. The consumer commerce practice exists for this kind of decision, and the companion piece on building a full channel mix goes deeper on the system around these four channels.

Questions operators ask
before moving the
budget.