Two things are becoming unmistakably true in performance marketing: the walled gardens are getting more expensive, and the open web is getting better at converting intent into revenue.
Meta CPMs have risen sharply over the past three years. TikTok audiences, while large, skew toward browsing behavior rather than immediate purchase decisions in most categories outside of impulse-buy price points. Google's Performance Max and Demand Gen campaigns give brands less control over placement and audience segmentation than performance teams are comfortable with. And for brands that don't sell on Amazon, retail media networks—the other dominant intent channel—aren't an option at all. Against that backdrop, a growing number of DTC e-commerce brands are rebuilding their channel mix around commerce media on the open web. Not as a supplementary test, but as a core performance channel with its own publisher strategy, budget allocation, and measurement framework.
Open web commerce media refers to performance-driven advertising placements across non-platform-owned environments where purchase intent is already present. That means editorial commerce sites, product comparison platforms, deal aggregators, cashback networks, and contextual placements alongside actively researching, deciding shoppers—rather than passive social feed scrollers.
The distinction from walled garden advertising isn't cosmetic. Meta and TikTok monetize attention—your ad competes with content the user chose to engage with. Open web commerce media, at its best, shows up in environments where the user is already in buying mode: reading a product roundup, comparing alternatives, searching for a deal, or consulting a review. That behavioral context changes conversion economics materially.
The backbone of open web commerce media for most DTC brands is affiliate network infrastructure. AWIN, CJ, Impact, Tradedoubler, and TradeTracker each maintain publisher marketplaces with thousands of editorial, deal, and comparison sites already set up to drive commerce traffic. Rather than building bilateral publisher relationships from scratch, affiliate networks let brands plug into an existing ecosystem of performance-oriented publishers and pay on a cost-per-sale or cost-per-lead basis.
That pricing structure is where the risk profile diverges from walled garden channels. Meta charges CPMs whether your ad converts or not. Open web commerce media through affiliate channels charges primarily on outcome—a publisher only earns when they deliver a verified sale. For brands under margin pressure, the difference between paying for impressions and paying for revenue is not academic.
The open web doesn't have the scale of Meta or TikTok. But it doesn't charge you for that scale if it doesn't convert. That asymmetry is exactly what performance-driven brands should be pricing in.
Open web commerce media isn't limited to affiliate networks. For brands with larger budgets or more complex media mixes, programmatic and native placements extend the reach significantly:
The measurement challenge with open web commerce media is the inverse of the walled garden problem. Walled gardens tend to overcount their own contribution through closed-loop attribution that credits the platform generously. Open web placements, spread across multiple publishers and networks, can be undercounted if a brand's analytics stack isn't stitching together data from affiliate networks, programmatic DSPs, and on-site conversion tracking simultaneously.
Getting this right requires:
The walled gardens will always have broader reach. But reach isn't the bottleneck for most DTC brands—efficiently converting high-intent audiences into customers is. Open web commerce media, structured correctly and measured rigorously, consistently outperforms awareness channels on that dimension. The brands compounding their performance results over time aren't chasing scale on Meta. They're building publisher relationships and measurement infrastructure on the open web, where intent is native and the cost structure rewards outcomes, not activity.