New ecommerce performance data released on June 18, 2026, by CommerceIQ reveals a paradox building inside Amazon's marketplace: brands arrived at the doorstep of Prime Day 2026 carrying more inventory than a year ago, yet losing materially more revenue to stockouts. The numbers tell a story of broadly improving efficiency that contains, within it, a concentrated and growing vulnerability.

The headline numbers from the CommerceIQ report

Retail media ROAS reached its highest level in nearly two years. According to CommerceIQ, ad spend increased approximately 10% year over year while return on ad spend climbed to roughly $5.80 - 15% above the same period in 2025, and the strongest reading in approximately 24 months. That gain arrived even as cost-per-click (CPC) remained fairly consistent year over year, which means the additional spend translated into better outcomes rather than simply higher pricing.

Conversion rates moved sharply. The average conversion rate across CommerceIQ's platform data reached 29.8% through the first five months of 2026, up from 26% over the same period in 2025 - representing approximately 15% growth year over year. April stood out as a particularly strong month, with conversion hitting 31%, the highest point of the year so far. These figures are drawn from aggregated and anonymized performance data spanning brands selling on Amazon, according to CommerceIQ.

The inventory picture appeared healthy at the headline level. On-hand inventory increased approximately 12% year over year, giving brands what CommerceIQ describes as improved supply-chain confidence heading into one of the year's largest shopping events. Purchase order fill rates held steady at approximately 83%, flat compared to the prior year but below the six-month highs reached earlier in 2026. According to the report, stronger supply-chain execution is translating directly into fewer missed sales at the PO level.

Then there is the stockout figure. Despite all of that positive inventory movement, revenue lost to out-of-stock events rose approximately 24% year over year. The out-of-stock rate itself remained relatively stable. The divergence - flat OOS rates producing sharply higher revenue loss - points to where the inventory gaps are occurring: not across a broad swath of products, but concentrated inside the highest-revenue ASINs in each brand's catalogue.

What the ROAS numbers actually mean

A ROAS of $5.80 is not simply a bidding efficiency number. It reflects how effectively the advertising spend is closing purchase decisions in an environment where the nature of those decisions is changing. According to CommerceIQ's press release, Amazon has reported that Alexa for Shopping users are 60% more likely to convert than other shoppers.

The glance view data in the report adds another layer to this interpretation. Glance views - the metric that counts visits to a product detail page - remained essentially flat year over year even as conversions climbed. CommerceIQ notes this efficiency gain "warrants monitoring to distinguish better targeting from a flat top-of-funnel." Amazon introduced Alexa for Shopping on May 13, 2026, merging the product expertise of Rufus with the personalization layer of Alexa+. The pattern in the conversion data - more purchasing with roughly the same browsing activity - is directionally consistent with AI-assisted decision-making that compresses the consideration phase.

"Brands are seeing meaningful efficiency gains across ecommerce heading into Prime Day," said Guru Hariharan, Founder and CEO of CommerceIQ. "Conversion rates are up 15% year-to-date, advertising performance is improving, and inventory positions are healthier than earlier this year. The data also reveals a deeper story about how shopping behavior is fundamentally changing."

Hariharan elaborated on the structural shift. "Recently, Amazon reported Alexa for Shopping users are 60% more likely to convert. When AI guides the shopping journey, consumers arrive at a product already decided. And if your product isn't in stock at that moment, you don't just miss out on a sale, you miss out on a customer relationship."

That mechanism matters for how brands interpret both their ROAS gains and their stockout exposure. When a shopper arrives at a product already committed to buying - guided by a conversational AI system rather than browsing multiple options - an empty shelf is not a minor inconvenience. It is an exit from the consideration set entirely.

The stockout problem in technical detail

The divergence between OOS rates and OOS revenue loss is the most technically interesting finding in the CommerceIQ report. According to the data, the Rep OOS percentage (the rate at which the representative listing for a product is out of stock) stayed flat. The revenue loss index, however, spiked to 122 in May 2026 - a 24% increase compared to the same period a year earlier.

CommerceIQ's analysis traces this to what it calls LBB Indexed - a metric tracking lost sales through the Lost Buy Box, which occurs when a brand's product is unavailable and a competitor captures the transaction. When the OOS rate is stable but revenue loss rises sharply, the divergence indicates that the products going out of stock are disproportionately high-revenue ones. A stockout on a low-velocity item costs very little. A stockout on an ASIN generating a significant share of a brand's monthly revenue costs a great deal - and the current pattern suggests the latter is where the gaps are concentrating.

According to the report, prioritizing replenishment by revenue rank could materially close the gap. That is a different approach from managing inventory by unit count or by OOS rate. A brand could have a perfectly stable OOS rate at the portfolio level and still be losing disproportionate revenue if its top five ASINs are the ones cycling in and out of availability.

Purchase order fill rates at approximately 83% are steady but below the six-month highs seen earlier in 2026. Fill rate is the metric that measures how fully Amazon's purchase orders are being fulfilled by brand supply chains - an 83% fill rate means roughly 17% of ordered units are not being delivered as requested. That gap, even at a stable level, creates structural inventory pressure that is amplified when demand accelerates around an event like Prime Day.

Prime Day 2026 is confirmed for June 23-26, a shift from the traditional July timing that PPC Land covered in depth after Amazon confirmed it on April 29. The June calendar compresses lead times relative to a mid-July event, making the inventory management challenge described in the CommerceIQ data more acute. As PPC Land reported, a unit sitting in a receiving queue in mid-June is not Prime-eligible. The FBA inventory cutoff for Prime Day 2026 passed on May 27 - meaning the data in this CommerceIQ report reflects the pre-event state that brands are now locked into.

Pricing dynamics and margin pressure

Average selling prices (ASPs) improved compared to earlier quarters but remained below the prior year. According to the CommerceIQ report, discounts fell to approximately the 15% range - a two-year low - yet that reduction in promotional intensity did not produce strong ASP recovery. The explanation in the data is a mix shift: consumers appear to be purchasing more lower-priced items, which pulls the average price down even as discounting eases.

Ordered product sales (OPS) rose approximately 5% over Q1 2026, and unit margins improved slightly, by approximately 0.5 percentage points. The margin number remains below the prior year, however, even as the sequential improvement is positive. CommerceIQ notes that OPS improvements over both Q1 and the prior year are a favorable signal, even with margin below the 2025 baseline.

The CPC consistency year over year is worth examining against the ROAS improvement. In a normal competitive auction, rising ad spend should push CPCs higher as more brands compete for the same placement. The fact that CPC held steady while ROAS improved by 15% suggests brands are getting more value from the placements they are winning rather than simply paying less for them. That is a meaningful distinction - it implies the quality of purchase intent reaching sponsored listings has increased, consistent with the AI-driven conversion pattern described elsewhere in the report.

How the data sits inside the broader retail media picture

The CommerceIQ findings land within a retail media landscape that has been expanding rapidly on Amazon specifically. Amazon's advertising business crossed $70 billion on a trailing twelve-month basis, with Q1 2026 generating $17.2 billion - a 24% year-over-year increase. The ROAS improvement measured in the CommerceIQ data is consistent with a platform that is both growing in spend and improving in yield efficiency.

The conversion rate numbers in the CommerceIQ data also intersect with broader industry reporting on AI-driven shopping. Amazon's billable AI shopping prompts launched on March 25, 2026, transitioning Sponsored Products and Sponsored Brands prompts from free beta to paid general availability. These prompts, which engage shoppers with conversational product information inside the Rufus AI interface, add a layer of conversion-oriented interaction that did not exist at the same scale in the comparison period a year earlier. The 15% year-over-year conversion increase documented by CommerceIQ covers a period when that infrastructure was being introduced and scaled.

Amazon's unified reporting exited beta in June 2026, standardizing attribution methodology across Sponsored Products, Sponsored Brands, and Amazon DSP. The standardization of attribution methodology carries implications for ROAS comparisons: when CommerceIQ's aggregated data shows ROAS improving to $5.80, part of that movement may reflect genuine efficiency gains, part may reflect attribution methodology changes, and part may reflect the AI-assisted conversion dynamics described above. Separating these threads from aggregate platform data is not a simple exercise.

The retail media attribution question is a persistent one. PPC Land's reporting on the hidden ROI problem in retail mediahas documented how single-retailer ROAS calculations can create feedback loops that undervalue upper-funnel activity. The ROAS gains CommerceIQ reports are real in their own measurement frame - but interpreting what is driving them requires context beyond the headline number.

The operational stakes for Prime Day

The data arriving from CommerceIQ covers April and May 2026. Prime Day 2026 runs June 23-26. The gap between the two periods is where operational execution either closes the stockout gap or doesn't.

According to CommerceIQ, "Prime Day outcomes will hinge on closing the inventory execution gap in highest-revenue products before the event opens." That framing reflects a specific structural reality: the brands that entered the event period with the best aggregate inventory positions may still face concentrated stockout risk on the ASINs that matter most for revenue performance during peak demand.

"The broader ecommerce environment appears healthier than it did earlier this year," Hariharan added. "The opportunity for brands now is making sure operational execution keeps pace. The difference between winning and losing during major retail moments often comes down to having the right products available at the right time."

The 24% increase in OOS revenue loss is particularly striking because it arrived not during a period of inventory weakness - aggregate inventory was up 12% year over year - but during a period of relative strength. That counter-intuitive outcome reinforces the granularity point: total inventory levels are not the relevant metric when stockout losses are driven by ASIN-level gaps in specific high-revenue products.

PPC Land's May coverage of Prime Day inventory deadlines laid out why the June timing creates specific pressure. An early-June Prime Day compresses the supply chain lead times that brands typically rely on, and the FBA inventory deadlines already passed before this data report was published. Brands reading the CommerceIQ findings are, in most cases, now working with the inventory positions they have - not positions they can materially change before June 23.

The report covers aggregated and anonymized performance data from CommerceIQ's platform across brands selling on Amazon, spanning April 2024 to May 2026 for trend purposes. The company describes its client base as more than 2,200 of the world's leading brands including Nestlé, Colgate, and Whirlpool.

Why the marketing community should read the numbers closely

The combination of metrics in this report - improving ROAS, rising conversion without rising traffic, and stockout losses concentrated in top ASINs - sketches a specific competitive environment. Better-converting listings benefit brands whose products are available. But when conversion efficiency is driven partly by AI-assisted purchase decisions that skip traditional multi-product browsing, availability gaps are more decisive than they were when a shopper would typically browse alternatives and potentially return.

In that environment, the inventory execution gap described by CommerceIQ is not merely an operational problem. It has direct advertising implications. A brand spending 10% more on retail media while its top-revenue ASINs cycle out of stock during peak demand is spending money to drive traffic to unavailable products. That dynamic wastes media budget and, as Hariharan notes, forfeits not just a transaction but a customer relationship at a moment when the customer arrived already decided.

The retail media efficiency story in this data is real. So is the stockout risk story. Both can be true simultaneously - and for Prime Day 2026, one determines how well the other pays off.

Timeline

Summary

Who: CommerceIQ, a retail AI platform serving more than 2,200 brands including Nestlé, Colgate, and Whirlpool, released this analysis. The data covers brands selling on Amazon's marketplace.

What: The Pre-Prime Day Pulse Report for April and May 2026 documents four key findings: retail media ROAS reached approximately $5.80, its highest level in nearly two years, with ad spend up approximately 10% year over year; average conversion rates hit 29.8%, up approximately 15% year over year, despite flat glance views; on-hand inventory increased approximately 12% year over year with PO fill rates steady at approximately 83%; and revenue lost to out-of-stock events rose approximately 24% year over year despite stable overall OOS rates, with losses concentrated in high-revenue ASINs.

When: The data covers the period from April through May 2026. The report was released on June 18, 2026, five days before Prime Day 2026 is scheduled to begin on June 23.

Where: The findings cover brands operating on Amazon's marketplace. The data is aggregated from CommerceIQ's platform, which operates across Amazon's US and international operations.

Why: The report matters because it captures the pre-event state of Amazon marketplace performance at a moment when inventory positions are locked and advertising plans are finalized. The divergence between improving efficiency metrics and rising stockout losses in top ASINs identifies a specific operational risk that will shape Prime Day outcomes for brands across the marketplace.