Custom labels (custom_label_0 … custom_label_4) are five free-form fields that let you write values of your own choosing into your product feed. Google does not show these fields to shoppers; you use them to split listing groups in Performance Max and product groups in Standard Shopping. A good custom label strategy means dedicating each field to a single question (e.g. "is the product selling?", "what is its stock status?", "which season does it belong to?") and updating the values automatically, based on data. That way, instead of splitting the campaign by hand every week, the feed moves products into the right group on its own.
Why do you need custom labels?
In a campaign you can only split products by category or brand. But advertising decisions usually rest on other questions:
- Has this product been selling recently, or is it getting impressions without selling?
- Is the profit margin high or low?
- Is stock plentiful, about to run out, or is the size run broken?
- Is it a seasonal product, on promotion, or a new arrival?
Google does not have this information; once you write it into the feed, it becomes usable in your campaign structure.
How should you allocate the five fields?
Dedicating each field to a single dimension makes reporting and campaign splitting easier. An example allocation:
| Field | Question | Example values |
|---|---|---|
custom_label_0 |
Performance | star, opportunity, low-performance |
custom_label_1 |
Stock status | stock-high, stock-low, broken-sizes |
custom_label_2 |
Price / margin band | price-0-500, price-500-1500, price-1500+ |
custom_label_3 |
Season / collection | summer-26, winter-26, all-season |
custom_label_4 |
Promotion | on-sale, new-arrival |
A few rules:
- Keep the number of values small. A handful of values per field is enough; overly fragmented values shrink the groups until they are meaningless.
- Write values consistently.
Star,starandSTARmay behave as different values; settle on one spelling. - One value per field per product. If two different rules write a value to the same field, decide in advance which one wins.
Using them in Performance Max
Custom labels are useful in Performance Max in two ways:
- Splitting listing groups: Within an asset group you can include or exclude products by custom label value. For example, you can take products labeled
low-performanceout of the main campaign and test them in a separate, low-budget campaign. - Separate campaigns: Moving high-performing products into their own campaign and running them with a higher budget and a different target ROAS is a common approach. This way a small number of best sellers do not swallow the entire budget, or, the other way round, products with potential are not left without any impressions.
Avoid having the same product in more than one Performance Max campaign; product groups should be set up to be mutually exclusive.
Using them in Standard Shopping
In Standard Shopping you can split product groups by custom label and bid separately on each group. A classic structure: split all products by custom_label_0, bid higher on the star group and lower on the low-performance group. Combined with campaign priority, you can also have the same product served from different campaigns for different queries.
Updating label values automatically
The power of custom labels comes from values that update themselves. Labels written by hand go stale within a few weeks: yesterday's best seller may be out of stock today. Typical data sources for automatic updates:
- GA4 metrics: Views, add-to-carts and purchases over a given period.
- Stock data: Total stock quantity and how many of the sizes are in stock.
- Price and discount rate: The price band and whether the product is discounted.
- Product tags / category: Attributes you keep on the store side, such as season, collection or fabric.
We covered the logic of grouping by product performance in the guide which products to advertise using GA4 data, and stock-based grouping in the guide on out-of-stock products and broken size runs.
How to set up custom labels in HexFeed
In HexFeed, custom labels are defined through feeds. On a platform (e.g. Google) one feed is chosen as the main feed; the other feeds on the same platform can each carry a custom label field (Custom Label 1–5) and a value. Products that match the rules of those feeds are written to the main feed's output with that value in the corresponding custom_label field.
Example: you create a feed called "Best sellers", set its rule to "more than 5 Google purchases in the last 30 days", choose Custom Label 1 as the custom label field and best-seller as the value. In the main Google feed, every product that matches this rule gets best-seller written to its custom_label_0 field. Because the rules are re-evaluated against current data every time the feed refreshes, the label stays up to date on its own. If two feeds on the same platform use the same custom label field, the panel warns you.
On the rule side you can use stock, size availability, price, discount rate, category, brand, product tag and 7–90 day GA4 metrics (views, add-to-carts, purchases, conversion), as well as the product matrix segment. For details see the feed filtering and product matrix pages; depending on your e-commerce platform, see the Google Shopping with Ticimax page.
Checklist
- I dedicated each of the five custom label fields to a single question and prepared a written table.
- I use a small number of consistently spelled values in each field.
- Label values update automatically from data (GA4, stock, price).
- I verified in Merchant Center that the custom label values reach the products.
- Performance Max listing groups or Shopping product groups are mutually exclusive.
- There is a separate, controlled test structure for low-performing products.
- I monitored for a few days how label changes were reflected in the campaign.
Frequently asked questions
Are custom label values visible to users?
No. Custom label fields are used only for campaign configuration and reporting; they are not shown in the ad or on the product card. Even so, keeping the values readable makes them easier to use within your team.
What if five fields are not enough?
Most of the time they are; if not, rather than squeezing several dimensions into one field, ask which question actually changes a campaign decision. A label dimension that is not used only produces clutter.
Does campaign learning suffer when a custom label changes?
When a product moves from one listing group to another, it may have to collect data again under a different campaign or group. Choose reasonable thresholds and do not make the period too short, so that labels do not change hands too often.
Can custom labels be used on Meta too?
Yes. The Meta catalog also reads the custom_label_0–custom_label_4 fields, and they can be used as filters in product sets. Meta's own internal_label field can be used for the same purpose.
