Italy's Carousel Spread: Instagram's Feed Out-values Its Facebook Twin at a Lower CPM

The loudest fight in Italian paid social is over short video. Which feed placement earns its CPM, whether Reels are underpriced, whether TikTok's auction is worth the creative overhead. While that argument runs, the richest view on the Italian benchmark sheet is sitting in a format most creative teams treat as a chore: the Instagram feed carousel.

This is a case study of one line in the portfolio:2026-08 benchmark set for Italy, plus its closest twin. Look at it.

The line

On Instagram, in the feed, as a carousel, in Italy:

• CPM: €2.48 (2.4751)
• Modeled value per view: 0.0391, the highest view value on the Italian sheet
• Value per reaction: 0.255
• Value per comment: 0.8; value per share: 1.5; value per save: 0.6
• Underlying evidence: 1,181 units in the source portfolio, the deepest sample on the sheet

Nothing about that list screams opportunity until you place it next to the one line that is supposed to be its equal.

Meet the twin

Facebook runs the same format in the same market, and the sheet prices it like a sibling that looks identical until you check the pockets.

Read the table twice, because two things are going on:

1. The model prices active engagement identically on both platforms. A comment is 0.8 either way, a share is 1.5, a save is 0.6. There is no platform premium on deliberate actions.
2. Everything contested, Instagram wins. View value runs about 6% higher (0.0391 versus 0.0368), reaction value about 42% higher (0.255 versus 0.180), and the CPM is about 10% lower (€2.48 versus €2.76). The platform where attention is modeled as more valuable is also the platform where it costs less.

How big the gap is in context

Both carousel lines sit far above the short-video lines on the same sheet. The modeled view values for TikTok in-feed (0.0026 for video, 0.0036 for image), YouTube Shorts (0.0053), and Facebook Reels (0.0023) run seven to seventeen times lower. Whatever the model sees in a carousel view, it sees very little of it in a skippable video impression. A carousel view implies dwell, a swipe, a decision. A video view can be half a second of autoplay.

We have covered Facebook's cut-price Reels in this space before, cheap video that the model values cheaply. The carousel spread is the opposite story: a placement where the low price and the high modeled value point the same direction.

Why the spread might exist

This section is interpretation, not benchmark data. Treat it accordingly.

• Carousels select for intent. Swiping is effort. The people who reach card three are closer to a buying mindset than the average feed scroller, and the model's view value may be capturing that self-selection.
• The format earns saves. Carousels are where how-to content, spec sheets, and comparisons live. A save is a deferred purchase signal, and a format that collects saves at 0.6 each compounds value that last-click dashboards attribute later or never.
• Creative supply is thin. Good carousel creative is harder to produce than a single static or a repurposed vertical video, so fewer advertisers compete seriously for the placement. Under-contested inventory with strong engagement economics is exactly where CPMs lag value.

What to do with it

• If you run Meta prospecting in Italy, spec at least one carousel per campaign and judge it on saves and swipe depth, not click-through rate.
• Build carousel creative that deserves a save: comparisons, checklists, before-and-after sequences, spec breakdowns. The valuation constants (0.8 comment, 1.5 share, 0.6 save) tell you what the model pays for.
• Do not abandon the twin. Facebook carousels are the second-richest view on the sheet at €2.76, and if your audience skews older the twin may be the better buy despite the spread.
• Replicate the math on your own account before moving real budget. These value figures are modeled, both lines are blended across objectives and verticals (both fields are unset), and everything comes from a single vintage. The spread is a hypothesis with unusually good numbers behind it, not a law.