Create product sets

​

Why segmentations matter

​

How it works

1. Export a CSV mapping each relevant product to its custom label (e.g., “High margin,” “Low impressions”).
2. Use these labels in your Google or Meta feeds for more targeted campaigns or bidding strategies.

​

Creating a product set

1. Go to “Segmentations.”
   • In the Dema platform, look for the Segmentations section to manage product-level groupings.
2. Click “New product set.”
   • Give your product set a name (e.g., “Slow movers” or “High margin products”) to easily recognize it.
   • Pick a custom label—this label will appear in your output CSV and can be mapped to a custom_label_x field in Google or Meta.
3. Save.
   • After entering a name and custom label, click “Save.”
   • You can always come back and edit these details later.

​

Adding rules to your product set

1. Click “Add rule.”
   • Provide a description for the rule (e.g., “Products without impressions”).
   • Configure filters to identify which products match your criteria.
   • Remember that the timeframe in the top-right corner (e.g., last 28 days) controls the data window for all metrics.
2. Choose filters.
   • You can filter on metrics (e.g., impressions, gross sales, net gross margin) or dimensions (e.g., category, channel).
   • For example, a rule might say: Impressions < 500 AND Channel = “Google.” Any product meeting all filters in this rule qualifies for that label.
3. Assign a label.
   • In the rule configuration, set a Label (e.g., “Low impressions”). This label will be applied as the custom label in your exported feed whenever a product matches these filters.
4. Optional: set an expiry date.
   • If you want this rule to expire on a certain date—e.g., after a seasonal promotion—include an expiry date. After that date, the rule disappears.
5. Rule order matters.
   • If a product matches multiple rules, it inherits the label from the first matching rule in the list.
   • Reorder rules if you want different prioritization.
6. Save.
   • Each time you add or modify a rule, click “Save.” Dema then updates the product set accordingly.

​

Viewing the results

• % of products in the set (by count)
• % of marketing spend (feed-based ads) these products account for
• % of gross sales they represent
• % of inventory value they hold

​

Practical examples

• Rule 1: Low impressions
  • Filters: Impressions (Product ads) < 500 AND Channel = Google
  • Label: “Low impressions”
  • Captures products with fewer than 500 impressions on Google in the last 28 days.
• Rule 2: Slow moving clothes
  • Filters:
    • Forecast sell-through (Rest of season) < 50%
    • Category = CLOTHING
    • Gross sales < 20,000
  • Label: “Slow movers”
  • Targets items with low sell-through rates and modest sales.
• Rule 3: High margin
  • Filter: Net gross margin 2 > 50.00%
  • Label: “High margin”
  • Ideal for identifying highly profitable products to promote further.

​

Exporting your product sets to feeds

1. Find the export option.
   • In the Segmentations view, locate the three dots (⋮) next to your product set and select Copy link to CSV file.”
2. How the CSV is structured.
   • The CSV contains product IDs at a higher “product” level, rather than size or variant level.
   • Dema also pulls in all relevant feed-based ad IDs to properly label items in Google or Meta. For instance, if product 1920 in Dema has size variants 1920-1, 1920-2 in Google, the CSV will list all those variant IDs under the same custom label.
   • Important: If a product hasn’t had any marketing spend in feed-based ads (i.e., it never appeared in Google or Meta), Dema doesn’t have a corresponding external ID. Consequently, that Google ID won’t appear in the exported file.
3. Mapping IDs.
   • Since Dema aggregates marketing data, it knows how product 1920 in your e-commerce platform relates to IDs 1920-1, 1920-2, etc. in Google, and possibly 1920 in Meta.
   • All these IDs get the same custom label in the export, meaning you can easily update your Google or Meta feeds to reflect the segmentation.
4. Use in Google or Meta.
   • In your advertising platforms, map the exported custom label column to custom_label_x or whichever attribute is appropriate. This allows you to create specialized campaigns or rules (e.g., increase bids on “High margin” items, limit spend on “Slow movers,” etc.).

​

Tips and best practices

1. Keep labels short and descriptive.
   • “LowImpr” or “HighMargin” are easier to manage in ad platforms than lengthy names.
2. Timeframe awareness.
   • Rules use the data from the timeframe in the top-right corner (e.g., last 28 days). If you need a broader overview, adjust the date range.
3. Mind the rule order.
   • The first matching rule is applied, so prioritize rules carefully to avoid unintended overlaps.
4. Monitor the charts.
   • The segmentation charts help you see how each product set affects overall business metrics.
5. Regularly refresh rules.
   • Product statuses change over time (e.g., a slow mover could become high margin if costs drop). Revisit your rules periodically.
6. Check product IDs.
   • If a product has never had any feed-based ad spend, Dema can’t export an external ID for it. It won’t appear in the CSV.

​

Next steps

1. Create product sets aligned with your goals—e.g., handling slow movers, capitalizing on high-margin items, or grouping low-impression products for improvement.
2. Refine rules and watch the segmentation charts to understand the impact on marketing spend, gross sales, and inventory value.
3. Export the CSV once you’re satisfied with your segment definitions.
4. Update your feeds: Upload the CSV to Google or Meta, mapping Dema’s custom labels to the correct attribute fields.
5. Optimize campaigns based on these new labels—target, bid, or pause segments strategically for higher profitability.