Italian Retail Marketplaces on Facebook: What the Thin 90-Day Data Can and Cannot Show

Summary

This note reviews Facebook activity for retail marketplaces in Italy over a 90-day window ending 18 October 2026. We set out to publish cohort benchmarks for paid and organic performance. The data did not support that. One paid query failed, the organic sample covers a single brand, and the spotlight view returned one ad. This article reports what came back, what did not, and what would be needed before any benchmark is published.

What we tried to measure

We ran three queries for the Italian market (IT), the retail marketplace industry, and the Facebook platform:

• Paid ads, sales objective, 90 days: The query failed. The analytics store returned a server error (ClickHouse code 386), caused by a type mismatch in a comparison between a string field and an unsigned integer ID. No ads were returned.
• Organic posts, 90 days: The query returned data for one brand and 10 posts.
• Spotlight ad, sales objective: The query returned one ad.

None of these results forms a cohort. Each is a sample of one kind or another, and the paid side has no usable sample at all.

Organic: one brand, ten posts

The organic window ran from 20 July to 18 October 2026. Across the 10 posts, interactions per post were:

• 25th percentile: 24
• Median: 37
• 75th percentile: 89

The median number of posts per brand was 10, but that figure describes a single account. Engagement rate and views per post came back null, so we cannot compare reach-based measures at all.

The spread between the middle and top of the distribution is wide. Still, with one publisher, that spread reflects the habits of one account, including its content mix, posting times, and audience. It cannot be separated into effects of format, timing, or the wider market. Reading it as a sector norm would be a mistake.

Paid: one spotlight ad

The single ad returned by the spotlight query was run by Amazon Prime. Its key details are:

• Language: Italian
• Target location: Italy
• Call to action: shop
• Primary audience label: unisex
• First seen: 27 June 2026
• Estimated mid-point spend: 23,364.2, with no lower or upper bound
• Reported EU total reach: 11,682,101

Two points limit how far this can be read. First, the spend estimate has no stated currency, and the reach figure covers the EU as a whole, not Italy alone. Second, the creative fields are empty. Format, caption, hook phrase, hook type, discount, urgency, and influencer flags are all null, and the thumbnail is missing. The only creative signal we can report is the call to action. We cannot say which approach the ad used or whether it was typical of the category.

Why we are not publishing benchmarks

Three gaps block a benchmark:

1. No paid cohort. The ads query failed, so there is no sample of paid campaigns for the market.
2. No breadth on organic. One brand is not a market. Percentiles from one account describe that account.
3. No usable creative detail. The one ad we could see has no format, hook, or caption data, so no creative pattern can be drawn from it.

Publishing percentiles or rankings under these conditions would give readers a precision the data does not have. That is why this review stops at description.

What would make this usable

A benchmark on this topic would need, at minimum:

• A working paid ads query, with the ID type comparison corrected so the cohort can return rows.
• Enough distinct brands in the organic sample to make percentiles meaningful, with a stated minimum before any figure is published.
• Populated creative fields for paid ads, so format and messaging can be compared rather than guessed.
• A confirmed currency and a stated geographic scope for every spend and reach estimate.

Until those conditions are met, the most useful takeaway is procedural. The Italian retail marketplace question on Facebook remains open. The data we have shows that one account published posts with interactions ranging from roughly 24 to 89 in the middle half of its sample, and that one large marketplace advertiser was running an Italian-language shopping ad with an estimated mid-point spend. Those are observations, not benchmarks, and they should be presented that way.

We will revisit this topic once the paid query runs and the organic sample covers more brands.