How we normalize growth metrics

Normalization is what makes trend metrics comparable and trustworthy. It isolates a trend's genuine evolution from noise: shifts in posting behavior, changes in the panel of accounts we track, social media seasonality, and other data-acquisition biases.

The reference we normalize against depends on where a trend sits in our taxonomy.

There are three cases, which you can read about below.

1. Macro trends and main categories

Macro trends are colors, prints, and fabrics - for example Leopard Print, Forest Green, or Denim.

Main categories are the broadest level of the taxonomy - for example Tops, Dresses, Boots, or Sneakers.

Both are normalized across all clothing items detected for Apparel, and all shoe items detected for Shoes.

This means their signal reflects their share of everything we see on social media within their universe (all Apparel, or all Shoes) - not their share of any one shape or crossing.

2. Main shapes and category crossings

Main shapes are the main silhouettes within a category - for example Jumpers is a main shape of Tops, and Cowboy Boots is a main shape of Boots.

Category crossings are a crossing between a broad category and a vertical (a color, print, fabric, or other attribute) - for example Red Tops (Tops crossed with a color) or Long Sleeve Tops (Tops crossed with a sleeve type).

Both main shapes and category crossings are normalized against the volume of their parent category (Tops, Bottoms, Boots, Sneakers, Bags, etc.).

For example, Jumpers and Red Tops are both measured against the total volume of Tops. If Tops overall are growing, that growth is factored out of both signals - what you see is how Jumpers, or Red Tops, are performing relative to the rest of Tops.

3. Shape crossings

Shape crossings are a crossing between a main shape and a vertical - for example Red Jumpers crosses the shape Jumpers with the color Red, while Cropped Jumpers crosses the same shape with a fit type.

Like category crossings, shape crossings are normalized against the volume of their own parent - that's what makes any crossing comparable. It's just that here, the parent is the shape itself, not the category above it.

Red Jumpers is measured against the total volume of Jumpers, not Tops. So broader shifts in the Jumpers market don't influence the crossing signal - what you see is purely how the attribute (Red) is performing within that shape.

The same logic applies across categories:

• Turtleneck Jumpers (shape crossed with a neckline type) is measured against the total volume of Jumpers.

• Ripped Jeans (shape crossed with a treatment) is measured against the total volume of Jeans, not all Pants.

• Fringe Cowboy Boots (shape crossed with a detail) is measured against the total volume of Cowboy Boots, not all Boots.

• Lace T-Shirts (shape crossed with a fabric) is measured against the total volume of T-Shirts, not all Tops.

Why this matters: the same attribute, two different readings

Because the reference changes depending on the level, the same vertical can tell a different story depending on what it's crossed with:

• Red Tops is a category crossing, so it's measured against all Tops. Its signal reflects how Red is performing across the entire Tops universe.

• Red Jumpers is a shape crossing, so it's measured against Jumpers only. Its signal reflects how Red is performing specifically within Jumpers, regardless of how Jumpers themselves are trending within Tops.

If Jumpers are declining as a shape but Red is holding steady within Jumpers, Red Jumpers will show a stable signal - while a broader crossing like Red Tops would still carry some of that Jumpers-driven movement, since it sits at the category level.

Summary

If you have questions about how a specific trend or crossing is normalized, don't hesitate to reach out to your account manager. Our product and fashion experts will be happy to walk you through it.