Everyone Wrote Off X Ads. Then Mack Weldon Hit 3.1x ROAS. Here's What's Actually Happening.

I'm going to write something that will annoy roughly half the people who read it.

Elon Musk is quietly building the most aggressive performance ad platform in eCommerce in 2026. Almost nobody in DTC marketing is paying attention because they hate him. And the brands willing to ignore the politics and just look at the numbers are hitting 3.1x to 3.8x ROAS at materially lower CPAs than Meta.

This is not a take. This is what's happening on the platform right now. I'm going to show you the receipts and tell you exactly what we'd do if we were running paid media for your Shopify brand this quarter.

The story everyone got wrong

In January 2026, X's UK ad revenue posted a 58% YoY decline. The marketing trade press loved this number. "Musk killed advertising." "Brands are fleeing X." "Bluesky won."

What they missed: that 58% number is about brand-safety advertisers — the Fortune 500 CMOs who pulled their reach campaigns over content moderation concerns. It tells you nothing about performance advertising, which is the thing that actually matters if you're a Shopify brand trying to sell more product.

Meanwhile, in the part of the platform that actually drives revenue for DTC: X has been quietly rebuilding the entire ad stack around performance. Real-time targeting. Reply ads. Dynamic product campaigns. A native Shopify app with one-click pixel setup.

And the brands that tested it got results most paid media buyers would kill for in 2026.

The numbers nobody's quoting

Mack Weldon — 3.1x ROAS on X targeting affluent urban men. (For context, the average Meta DTC ROAS in 2026 hovers around 1.8–2.4x for cold prospecting.)

Another retailer running the X Shopify Pixel — 3.8x ROAS improvement and 62% drop in cost per purchase after one-click setup. Read that again. Not 6%. 62%.

These are not normal numbers in 2026. Meta CPAs have been climbing year-over-year since iOS 14. Google Shopping is saturated for almost every category. TikTok is a black box. The advertiser cost-of-attention is at an all-time high almost everywhere.

Except on X. Where political baggage is keeping CPMs artificially low, and Musk's performance-ad and AI-driven targeting push is actively improving the auction every quarter.

If you're running our paid media playbook at scale, this is the exact pattern you look for: an underpriced channel where the auction is moving in your favour. We covered the same underpriced-channel dynamic when we rebuilt The Phoenix's account — see the Phoenix paid media case study where we ran $23M+ through Google and Meta at 3x ROAS by treating every new audience pool as a temporary arbitrage window.

Why this is happening

Three things, stacked.

1. The brand-safety advertisers left

When the Fortune 500 pulled out, they took their ad budgets with them. Auction demand dropped. Which means CPMs dropped. Which means the brands still bidding — mostly DTC, small-to-mid eCom, performance-focused — are buying impressions at a fraction of what Meta charges for equivalent reach.

This is exactly what happened to YouTube in 2017 after the brand-safety scandal. Performance advertisers who held their nerve got 18 months of dirt-cheap CPMs before the big spenders came back.

X is in that window right now. Whether it lasts 6 months or 24 months, nobody knows. But it's open right now.

2. Musk is genuinely fixing the ad stack

I'll give him credit where it's due. Musk has explicitly committed to performance-first advertising and X has shipped real product:

Dynamic Product Ads that pull from your Shopify catalog and personalise per-impression — same model that made Meta DPA the workhorse of DTC paid media for a decade
Reply ads that drop your product into high-engagement conversations as a contextual recommendation
Real-time targeting powered by Grok's natural-language understanding of what users are talking about right now
One-click Shopify pixel setup that takes the technical friction down to roughly zero

This is a serious performance-ad stack. The DTC brands that already tested it are seeing it perform like Meta did in 2018.

3. Agentic commerce is about to land — and X is positioning for it

This is the part nobody's connecting yet.

On June 9, 2026, Gopuff and xAI launched "Go," an AI shopping agent built on Grok that automatically builds carts based on user prompts ("plan a game-day party" → 15-minute delivery), reads signals from X (yes — your X posts, your weather, your context), and predicts what you'll need next.

That's agentic commerce arriving in the wild. And X is the data layer underneath it.

Now consider: when AI shopping agents start meaningfully buying products on people's behalf (which is the eMarketer forecast for 2026–2027), where do you want your brand presence to be? Right inside the data the agent is reading.

Brands that build a presence on X in 2026 won't just get cheap clicks today. They'll be the brands Grok recommends in 2027. We talked about the same shift in our SEO in the age of AI post — the citation footprint matters more than the backlink profile now. Same principle applies to commerce. And it's the same conclusion we reached covering Shopify's Winter '26 Edition: clean data and a clean store are the prerequisites for everything AI-mediated, including ads.

What we'd do if we were running paid media for your Shopify brand right now

Three-step playbook. Test budget required: $3K–$5K. Not $50K.

Step 1 — Install the X Shopify Pixel this week

The X Business Shopify app is one-click. Takes 15 minutes. Even if you never run an X campaign, get the pixel installed and start collecting audience data now. This is free. There's no reason not to.

Step 2 — Run a dynamic product ad test against your top-5 SKUs

Two weeks. $3K total spend. Target: existing customer lookalikes + interest-based audiences relevant to your category. Use your existing product catalogue feed. Test creative pulled directly from your highest-performing Meta creatives — no reinvention required.

If you see anything north of 2.5x ROAS in two weeks at that spend level, you've found your channel. Scale.

If you see less than 1.5x ROAS, kill it and move on. (You're out $3K. That's the cost of finding out a channel doesn't work for you, which is roughly half what you'd burn testing a single Meta creative angle at scale.)

Step 3 — Layer reply ads on conversations in your category

Reply ads are the underrated tool. Identify the 20–30 X accounts your customers actually follow. Run reply ads against high-engagement conversations on those accounts. Conversion rates on contextually-placed reply ads in our test campaigns have been meaningfully higher than feed ads. Less competition. More signal.

This is the same logic that made influencer-adjacent paid media on Meta work in 2019. The audience is already engaged. You just need to show up.

Two caveats — because this isn't a free lunch

1. Brand safety risk is real. Your ad will appear next to content you don't control. If your brand can't tolerate occasional adjacency to politically-charged or otherwise-edgy content, X is not for you. Most DTC brands selling actual products to actual customers can tolerate this just fine. If you're selling to enterprise procurement, maybe not.

2. Attribution is messy. X's attribution is improving but it's not Meta. Run X campaigns with a clear baseline of pre-existing performance on other channels so you can measure incremental lift, not just last-click ROAS.

We covered the same attribution complexity in our Phoenix paid media case study — where we built a custom attribution layer pulling Affirm and Klarna financing data back into the ad performance dashboards. That kind of work matters more on X than on Meta because X's native reporting is younger. Same logic pairs with email — the Phoenix email program drove $15.8M in email revenue alongside the paid spend, and Meta's pixel never saw most of that journey.

The bigger story

What Musk is building on X is part of a broader shift the smart paid media buyers are already positioning around. The big consumer attention is moving toward AI-mediated discovery — agentic shopping, AI-search, contextual ads inside conversational AI. The brands that build presence on the platforms where that infrastructure lives now will own disproportionate share of that traffic when it goes mainstream.

X is one of those platforms. Reddit (which just launched its own Shopify DPA integration in March 2026) is another. TikTok Shop continues to grow. Shopify's own Shop AI assistant is rolling out broader. The brands that go all-in on Meta + Google in 2026 and ignore everything else will look exactly like the brands that went all-in on Facebook + Google in 2018 and ignored Instagram and TikTok. (The Klaviyo side of this same shift — autonomous AI in your marketing stack — is covered in our Klaviyo Composer + Customer Agent breakdown.)

You know what happened to those brands.

The honest summary

You don't have to like Elon Musk to run X ads. You don't have to agree with him politically. You just have to look at what's happening in the auction and recognise that an underpriced performance channel with native Shopify integration and an AI commerce future is the kind of opportunity DTC brands get maybe once every three years.

The brands that ignore that opportunity because of how they feel about the platform owner will be the brands wondering in 2028 why their competitors have a CAC advantage they can't reproduce.

Politics doesn't sell product. Performance sells product. The numbers are the numbers.

If you want the X Ads test playbook executed end-to-end on your Shopify brand — or you want a sanity check on whether your current paid media mix is leaving this opportunity on the table — that's exactly the audit we do. We've managed $23M+ in paid media spend at 3x ROAS and run paid media across Google, Meta, TikTok, Reddit, and now X for DTC brands at every scale. We covered our broader take on AI in eCommerce marketing elsewhere, but this X angle is the one most agencies are missing right now.

Sources: