Segmentation is the process of dividing a broad customer base or product catalog into smaller groups that share a meaningful characteristic — behavior, purchase history, margin, seasonality, or product category — so that marketing and feed strategy can be tailored to each group instead of applied uniformly across the board. In feed management specifically, it applies on both sides at once: audiences get segmented by who they are and how they act, while the catalog itself gets segmented by which products deserve more aggressive bidding, tighter margin control, or seasonal prioritization. It's the organizing principle behind almost every advertising decision that goes beyond a single, one-size-fits-all campaign.

Why Segmentation Matters for Feed Performance

Treating an entire catalog as one undifferentiated block means a retailer's best-margin, highest-converting products get the same budget and bidding treatment as slow-moving stock, which flattens performance instead of concentrating spend where it earns the most. Splitting the catalog into segments — by margin tier, category, or expected seasonal demand — lets a retailer push budget toward what's actually working and pull back from what isn't, often with only a modest restructuring of existing feed data. This is close to the exact mechanic behind how Note Cosmetics increased ROAS by 127% while spending 10% less: reorganizing the catalog into performance-based segments and adjusting strategy per group rather than campaign-wide.

How Segmentation Works

Segments are usually defined through custom label fields or product type values in the feed — tagging products by margin band, bestseller status, or seasonal relevance — which campaign structures then read to apply different budgets, bid strategies, or promotional rules per group. On the audience side, segmentation draws on behavioral targeting to split shoppers by demonstrated intent, and increasingly on hybrid matching, which blends structured identifiers with fuzzy attribute matching to group similar products or similar shoppers even when the underlying data isn't perfectly clean. Segments aren't static either: demand forecasting often determines when a product should move from a low-priority segment into a high-priority one ahead of an expected demand spike, so the logic needs to update on a schedule rather than being set once and left alone.

Example: Segmentation in a Product Feed

<item>
  <g:id>SKU-41823</g:id>
  <title>Matte Liquid Lipstick - Terracotta</title>
  <link>https://example-shop.com/products/matte-liquid-lipstick-terracotta</link>
  <g:price>18.00 USD</g:price>
  <g:availability>in stock</g:availability>
  <g:custom_label_0>high-margin</g:custom_label_0>
  <g:custom_label_1>bestseller</g:custom_label_1>
  <g:custom_label_2>autumn-forecasted-demand</g:custom_label_2>
</item>

Each custom_label places this item into a different segment simultaneously — margin tier, sales rank, and seasonal forecast — so campaign rules can layer bid adjustments for all three at once instead of managing them as separate, conflicting campaigns.

Related Concepts

Segmentation rarely operates in isolation: it's the foundation that behavioral targeting builds on to narrow audiences further, and it increasingly relies on hybrid matching to group products and shoppers accurately when identifiers alone don't tell the full story. Because segments should shift with expected demand rather than stay fixed, demand forecasting plays a direct role in deciding which segment a product belongs to at any given point in the season.