I have argued that ecommerce marketers may be overlooking catalog ads while focusing on static images and UGC videos. In a September 30, 2025 post, I repeated a claim that 70% of ecommerce ad spend on Meta goes to catalog ads, described as ads pulled straight from a product database with no logo or design. I also repeated a claim that catalog ads account for 70% of Meta's $134B revenue, and that an unnamed startup raised $10M to make them 2x more profitable. These are claims in the posts, not figures I independently verified; their sources, scope, and measurement are not provided.
On September 30, 2025, replying to @mattepstein and addressing @oliverbrocato, I said that higher intent, lower friction, and a more eye-catching catalog with Marpipe contributed to “4 ROAS.” I gave no campaign, time period, or measurement details, so this is a reported claim rather than a verified result.
On October 24, 2025, I described a recently made demo store where Meta started showing an upsell product taken from the Shopify catalog. To my surprise, the product “started printing”; I gave no measurable sales or revenue figures. I wondered whether sometimes we should let Meta cook more, but did not report making that decision.
A post I shared reported that live data from 650+ accounts showed 4.0 ROAS for image and video ads and 6.86 ROAS for plain catalog ads. It explained the difference as catalog ads capturing existing purchase intent, like Amazon's “you may also like.” The post does not specify the accounts, measurement period, calculation, or whether the ad types were compared under equivalent conditions, so I treat the figures as reported claims rather than verified results.
The scale problem in the post was that regular ads can showcase only 4–6 products max, while Shein has 1.3 million SKUs and Zara adds 36,000 new products annually; the post argued that individual ads cannot be created for thousands of products. These figures and the scale argument were quoted from @ariesnotebook, not independently established by me. The challenge connects to ai fixes catalog ad scale, which covers Marpipe as a proposed way to address catalog-ad scale.
On September 30, 2025, I agreed with @VibeMarketer_ that “necessary evil” described traditional catalog ads well, and said I was hyped for “this.” The post does not identify what “this” referred to or include the original characterization.
Related: ai fixes catalog ad scale, unit economics over roas, meta ads versus alternatives