Italian Beauty Instagram Is Buying a Portfolio of Tests, Not a Format Strategy

Italian beauty advertisers on Instagram are not signalling a broad creative-format consensus. They are signalling something more operational: a long tail of small paid experiments, with a relatively small group of ads absorbing meaningful budget.

Across a 90-day cohort of 248 Instagram beauty ads from 10 brands, median spend was just €6.18. The 25th-percentile ad spent €1.44, while the 75th-percentile ad spent €35.51. [1] Total spend across the cohort was roughly €26,744. [1] That is not the spend curve of a market putting a large share of its budget behind a few standardised, repeatable executions. It is a market where many ads get a small allocation before evidence earns them more.

The useful planning implication is that the median is not a target. It describes the market's testing floor.

The gap between a test and a scaled ad is the real story

The distance from €1.44 at the lower quartile to €35.51 at the upper quartile is substantial. An ad at the 75th percentile has received more than five times the median budget, and nearly 25 times the budget at the 25th percentile. [1]

That spread suggests a two-speed operating model:

1. Low-cost creative probes make up a large share of the live library.
2. Selected executions earn additional delivery after the first allocation.

For operators, this changes how a competitor ad library should be read. A crowded library is not necessarily evidence that every concept is important. Many of those ads may be cheap probes. The more relevant question is: which creative propositions keep appearing after a brand has had a chance to cut the weak ones?

This cohort cannot answer that question directly, because format is unclassified for every ad. It does, however, establish the budget behaviour. Italian beauty advertisers appear to be running a testing portfolio rather than visibly concentrating spend around a known dominant format. [1]

Format reporting is the missing layer

Every ad in the paid cohort is labelled unknown for format. [1] That means there is no defensible comparison available between Reels, Stories, static placements, carousels, or other creative containers.

This is more than a reporting inconvenience. It blocks a core optimisation decision. If performance data cannot be connected to format, a team may conclude that a winning message is universally strong when it is actually winning because of its delivery environment. The reverse problem is equally costly: a weak result may be blamed on the product claim when the real issue is executional fit.

The absence of usable CPM data compounds the problem. No ads in this cohort have CPM available, so spend cannot be converted into a cost-of-delivery benchmark. [1] Operators therefore cannot use this dataset to claim that one beauty format buys cheaper reach than another. They also should not infer efficiency from spend alone.

One reach signal is worth retaining. The median EU total reach per ad was 8,645. [1] But this is a cumulative reach measure, not a clean daily-delivery or format-level efficiency metric. The reported reach-per-day percentile values are all zero, even though reach-per-day data is present for 230 of the 248 ads. [1] Treat that field as unsuitable for pacing analysis until its measurement logic is validated.

What to do differently next week

A beauty team competing in this market should build the missing decision structure internally rather than wait for category reporting to improve.

Tag every ad before launch. At minimum, record placement or format, visual mode, product category, talent type, opening hook, claim, offer status, and call to action. A format label alone is insufficient. A creator-led product demo and a polished product film may both be video, but they are different creative hypotheses.

Separate test budgets from scale budgets. The category's lower spend points indicate that small experiments are normal. [1] Make that explicit in the workflow: assign a fixed test allocation, define a minimum observation window, then promote only ads that meet a pre-agreed performance threshold.

Measure promotion rate, not only winner performance. If ten ads launch and one scales, the team needs to know whether the problem is weak hypotheses, poor production, bad audience matching, or an overly strict promotion threshold. The spend distribution makes this especially important, because raw ad count can conceal a very small number of truly scaled executions.

Do not claim a format winner from this market snapshot. The data does not support one. The honest competitive advantage is cleaner metadata and a faster learning loop.

Italian beauty advertisers appear comfortable funding many small paid tests. The brands that turn that behaviour into an advantage will not simply produce more variants. They will know which variant was tested, what it was designed to prove, and why it earned the right to move from a €6-style experiment into a larger budget decision.