Similarweb's State of Ecommerce 2026 report found that direct referral traffic from AI platforms is up more than 200% year over year, but still doesn't drive much traffic to ecommerce sites compared with search. At the same time, AI buying recommendations carry outsized influence on the sale itself, in some cases giving the recommended brand a 2-to-1 advantage over its competitors. The report also found that 89% of the time, shoppers who use AI in their research also use search, meaning consumers are stacking tools rather than switching between them. For eCommerce teams, that combination creates a real blind spot: the channel with the least visible traffic is quietly deciding which brand wins, and most teams don't have anyone whose job is to see it, let alone own it.
Similarweb's newly released State of Ecommerce 2026 report, produced in collaboration with Statista, surfaces a contradiction that most eCommerce marketing and analytics teams aren't built to handle. Direct referral traffic from dedicated AI platforms is up more than 200% over the past year. And yet, according to the report, that traffic still doesn't produce a large volume of clicks to ecommerce marketplaces and online stores compared with search. Judged purely on click-through, AI would look like a minor channel.
It isn't. The same report found that AI buying recommendations exert an outsize influence on purchases, in some cases giving the recommended brand as much as a 2-to-1 advantage over its competitors. A channel can be small in traffic and enormous in influence at the same time, and that's exactly the situation most eCommerce teams are currently unequipped to measure.
Two Numbers That Don't Match, and Why That's the Point
Traditional marketing and web analytics were built around a simple assumption: traffic volume roughly tracks with impact. A channel that sends a lot of clicks matters more than one that sends a few. Similarweb's data breaks that assumption cleanly. AI referral traffic is growing fast in relative terms but remains small in absolute volume next to search. Its influence on the actual purchase decision, however, doesn't scale down to match.
Despite AI referral traffic growing more than 200% year over year, Similarweb found it still doesn't produce a large volume of direct clicks to ecommerce sites compared with search. The influence shows up in who gets purchased, not in who gets clicked, which is exactly why it's so easy for a standard analytics dashboard to miss it.
We wrote earlier this year about the surge in AI shopping traffic and whether eCommerce teams are built to capture it. Similarweb's newer numbers sharpen that argument. The traffic itself was never really the point. The purchase decision an AI conversation quietly shapes, before a shopper ever lands on a site through search or a direct visit, is the part that actually moves revenue.
Consumers Aren't Switching to AI. They're Stacking It on Top of Search.
The report's most useful finding for hiring purposes might be this one: 89% of the time, consumers who use AI in their shopping research also use search. Daniel Reid, Similarweb's Principal Insight Analyst for Consumer Goods and Retail and the report's lead author, put it plainly: consumers are not switching tools, they are stacking them.
Reid described the pattern as shoppers using AI to explore and narrow down options, then turning to search to move toward an actual decision, adding that the most complex journeys, the ones using both, convert the best. That's a direct challenge to any team that's organized its AI strategy and its search strategy as two separate, competing initiatives rather than one connected discovery layer a shopper moves through in sequence.
The Attribution Blind Spot Most eCommerce Teams Have
Here's where the hiring gap actually lives. Most eCommerce marketing organizations still measure channel performance through last-click or last-touch attribution models built for a world where search and paid media were the dominant discovery paths. Under that model, a purchase that started with an AI recommendation and ended with a branded search gets credited entirely to search. The AI conversation that actually decided which brand won disappears from the report.
A channel with a 2-to-1 influence on purchase decisions and almost no visible attribution isn't a small problem to leave unowned. It's the exact kind of gap that shows up in board decks as "search performing well" while the real driver goes uncredited and under-resourced.
We covered a closely related version of this gap in SEO vs. AEO vs. GEO: what eCommerce hiring managers need to know. The practical hiring lesson from Similarweb's data is the same one: AEO and GEO aren't a replacement for a search function, they're a second discipline that has to sit next to it, because shoppers are visibly using both together and converting best when brands show up well in both.
What This Actually Means for Who You Hire
AEO/GEO Specialist
Owns whether a brand actually gets recommended inside AI conversations, tracking share of voice across AI platforms the way an SEO specialist tracks search rankings. This is a distinct discipline from traditional SEO, not a rebrand of it.
Marketing Attribution / Analytics Lead
Builds measurement models that can credit AI-influenced purchases instead of defaulting to last-click, so leadership can see the real performance of a channel that drives outsized influence with comparatively little visible traffic.
Mobile Commerce / App Product Manager
Similarweb's report also found ecommerce app sessions growing at roughly 1.3 times the rate of web visits, with 86.5% of US consumers saying they primarily shop on a smartphone or tablet. A discovery layer built for AI and search still has to convert on the surface where most shoppers actually complete the purchase.
Marketplace Category Growth Manager
The report found clothing, shoes, and jewelry unit sales up 33.7% in the US within marketplaces, while electronics grew just 1.5%. Categories driven by discovery and recommendation are pulling ahead of categories where shoppers already know exactly what they want, which argues for category-specific ownership rather than one generalist marketplace role covering both.
Who Should Actually Own This
At most eCommerce organizations today, AI visibility gets handled informally by whoever already owns SEO, if it gets handled at all, and attribution modeling stays frozen in a last-click framework nobody has revisited since paid search was the dominant acquisition channel. That was a defensible setup when AI-influenced purchases were a rounding error. It stops being defensible once a single recommendation can swing a purchase 2 to 1 in a competitor's favor.
None of this requires abandoning search-focused hiring. It requires recognizing, the way Similarweb's data makes unavoidable, that AI and search are now one connected discovery journey rather than two competing channels, and that the team structure needs an explicit owner for the half of that journey a standard analytics dashboard still can't see.
Global B2C ecommerce revenue is projected to top $4.9 trillion by 2030, per Statista's contribution to the report, an increase of more than 27% from 2026. We've written before about the hiring gap behind AI's growing share of that 2030 number. Similarweb's data adds the missing piece: the AI influence already showing up in today's purchase decisions is real revenue that most eCommerce teams currently have no one assigned to see, let alone grow.
Frequently Asked Questions
What did Similarweb's State of Ecommerce 2026 report find about AI and shopping?
Similarweb's 2026 report found that direct referral traffic from AI platforms is up more than 200% year over year, but still doesn't drive a large volume of traffic to ecommerce marketplaces and online stores compared with search. At the same time, AI buying recommendations carry outsized influence on which brand actually gets purchased, in some cases giving the recommended brand a 2-to-1 advantage over its competitors. The report also found that 89% of the time, consumers who use AI in their shopping research also use search, meaning most shoppers are combining tools rather than replacing one with the other.
Why doesn't AI referral traffic show up more in ecommerce analytics?
Most shoppers don't click straight from an AI conversation to checkout. Similarweb's data shows AI conversations produce far less immediate click-through traffic than search does, even as they shape which brand a shopper eventually chooses. A shopper might get a recommendation from an AI platform, then search for and purchase that exact brand directly, a purchase that traditional last-click attribution credits to search or direct traffic, not to the AI conversation that actually decided the outcome.
What does it mean that consumers are "stacking" AI and search instead of switching?
Similarweb's lead analyst, Daniel Reid, described it directly: consumers are not switching tools, they are stacking them. Shoppers increasingly use AI platforms to explore and narrow down options, then turn to search to move toward an actual purchase decision. The report found the most complex shopping journeys, the ones that use both AI and search, convert at the highest rate. That means brands optimizing for one channel while ignoring the other are optimizing for only part of the actual path to purchase.
What kind of hire fixes the AI attribution blind spot?
It typically takes two roles working together rather than one generalist. An AEO/GEO specialist owns whether a brand actually gets recommended inside AI conversations, distinct from traditional search rankings. A marketing attribution or analytics lead builds measurement models that can credit AI-influenced purchases instead of relying purely on last-click data, so the AI channel's real impact on revenue becomes visible to the rest of the organization.
Does this affect every ecommerce category equally?
No. Similarweb's report found meaningful category variation within marketplaces alone: clothing, shoes, and jewelry saw unit sales grow 33.7% in the US, while the electronics category was nearly flat, up just 1.5%. Categories with more subjective, discovery-driven purchases appear to benefit more from AI-influenced recommendations than categories where shoppers already know exactly what model or SKU they want.