Tag Rules are a powerful tool for controlling and filtering the games that appear in your recommendation placements, especially for the Gaming vertical. Unlike the Sports vertical, where rules are more dynamic and tied to inventory field names, Tag Rules in Gaming offer precise, structured control to create highly targeted outputs tailored to your needs.
This guide walks you through creating Tag Rules, understanding available rule types, combining them effectively, and navigating their limitations to build engaging recommendation experiences. Whether you’re promoting a new game or ensuring supplier diversity, Tag Rules give you the flexibility to shape your placements.
Getting Started: Create a New Tag Rule
To begin, navigate to the placement configuration screen in your platform’s dashboard. Click the Add New Row button to create a new rule entry. This opens a form with dropdowns and fields to select the rule type (e.g., IS_IN, RANK), tag (e.g., supplier), values (e.g., Supplier A), and optional parameters like position or percentage.
Available Rule Types
Tag Rules let you filter recommendation results by applying specific conditions to your game inventory. There are five main rule types, each defining a unique way to shape your placements:
| Type | Summary |
|---|---|
RANK |
Pins a game to a specific position or range (e.g., positions 1 to 5). |
RANK_NOT |
Prevents a game from appearing in a specific position or range. |
IS_IN |
Includes only games matching the specified tag values. |
NOT_IN |
Excludes games matching the specified tag values. |
% Tag |
Ensures a percentage of results match the specified tag values (e.g., 50% of games from a supplier). |
Combining Rule Types: AND vs. OR Logic
Tag Rules use AND and OR logic automatically based on how you structure them, giving you flexible control over your placement output.
AND Logic
When you add multiple separate rules (i.e., multiple rows in the configuration screen), they are combined with AND logic. This means a game must satisfy all rules to appear in the placement.
Example:
Suppose you create two rules:
IS_INsupplier=Supplier ANOT_INterritory=Territory B
The placement will only show games from Supplier A AND not available in Territory B.
OR Logic
Within a single rule, you can specify multiple values (e.g., multiple suppliers). These are evaluated with OR logic, meaning a game only needs to match one of the values.
Example:
A single rule like IS_IN territory=Territory A OR Territory B will include games available in either Territory A or Territory B.
AND and OR logic is working as expected.Tag Sources
Tag Rules are built from the tags in your game inventory, such as supplier, territory, game_name, game_code, or custom tags like theme=adventure. Any tag present in your inventory is available for use, giving you the flexibility to create rules tailored to your specific data. For example, you could filter by theme=fantasy if your inventory includes a theme tag.
Rule Type Details & Examples
IS_IN
This common rule type restricts recommendations to games with specific tag values.
Example: To show games only from Supplier A or Supplier B:
- Rule:
IS_IN - Tag: supplier
- Values: Supplier A, Supplier B (uses
ORlogic, including games from either supplier)
In the UI, select IS_IN from the rule type dropdown, choose supplier as the tag, and enter Supplier A and Supplier B in the values field.
NOT_IN
The inverse of IS_IN, this rule excludes games with specific tag values.
Example: To exclude games from Supplier C:
- Rule:
NOT_IN - Tag: supplier
- Value: Supplier C
This ensures no games from Supplier C appear in the placement.
RANK & RANK_NOT
These rules let you pin games to specific positions or prevent them from appearing in certain spots.
Example 1: Pinning a Game
To ensure a new game appears in position 1:
- Rule:
RANK - Tag: game_name
- Value: Name of New Game
- Position: 1 to 1
Example 2: Pinning to a Range
To place promotional games in the top 10:
- Rule:
RANK - Tag: promotion_tag
- Value: promo_games_q2
- Position: 1 to 10
Example 3: Excluding a Position
To prevent a game from appearing in position 1:
- Rule:
RANK_NOT - Tag: game_name
- Value: Old Game
- Position: 1 to 1
RANK rules (e.g., two rules targeting position 1–5), as only one game can occupy each position, causing conflicts. Use exact positions (e.g., 1–1, 2–2) to minimize issues.% Tag
This rule ensures a specified percentage of games in the placement match a tag.
Example: To ensure 30% of a 20-game placement features games from Supplier Evolution:
- Rule:
% Tag - Tag: supplier
- Value: Evolution
- Percentage: 30
- In Top: 20
% Tag rules are “best-effort.” The system tries to meet the percentage but may fall short if other rules (e.g., RANK) or inventory constraints limit matches.Practical Use Cases
Here are two scenarios showing how to combine rules for real-world goals:
Use Case 1: Promoting a New Game with Supplier Diversity
You want to pin a new game to position 1 and ensure 50% of the remaining 19 positions in a 20-game placement feature games from Supplier X or Supplier Y.
- Rule 1:
RANK, game_name=New Game, position=1 to 1 - Rule 2:
% Tag, supplier=Supplier X OR Supplier Y, percentage=50, in top=20
Test in Rec Viewer to confirm the new game is first and roughly 9–10 of the remaining slots feature Supplier X or Supplier Y.
Use Case 2: Balancing Multiple Suppliers
You want a 60-game placement to evenly feature games from four suppliers (A, B, C, D). Instead of one rule with IS_IN supplier=A OR B OR C OR D (which may skew toward A), use separate rules:
- Rule 1:
% Tag, supplier=A, percentage=25, in top=60 - Rule 2:
% Tag, supplier=B, percentage=25, in top=60 - Rule 3:
% Tag, supplier=C, percentage=25, in top=60 - Rule 4:
% Tag, supplier=D, percentage=25, in top=60
Check Rec Viewer to ensure a balanced distribution (around 15 games per supplier).
Advanced Concepts & Limitations
Tag Rules are powerful but have nuances and limitations. Understanding how they’re evaluated and interact is key to getting the results you want.
How Rules Are Evaluated
The system follows a structured process to apply Tag Rules:
- Collects all recommendation items from your inventory.
- Gathers all configured Tag Rules.
- Sorts items by score (ascending).
- Applies
RANKandRANK_NOTrules to pin or exclude specific positions. - Applies
% Tagrules to the remaining slots. - If not enough games match the rules, backfills with other inventory items to fill the placement.
This order matters, especially when combining RANK and % Tag rules.
Example:
In a 20-game placement:
- A
RANKrule pins 10 games to positions 1–10. - A
% Tagrule for supplier=Supplier X at 50% applies only to the remaining 10 slots (positions 11–20), resulting in 5 games from Supplier X, not 10.
If too few games match, the system backfills with other inventory items, which may not align with your rules. Use Rec Viewer to spot backfilling.
Known Limitations
Tag Rules guide the system but aren’t foolproof. Here are key limitations to watch for:
- Best-Effort Matching: Rules aren’t guaranteed. If your inventory lacks enough matching games (e.g., not enough Supplier X games for a 50% rule), the system may return fewer matches or backfill with non-matching items.
- Overlapping Rank Rules: Multiple
RANKrules targeting the same positions (e.g., two rules for 1–5) conflict, as only one game can occupy each spot. For example, if Game A and Game B both target position 1, only one wins. Use exact positions (e.g., 1–1, 2–2) to avoid this. - Uneven OR Distribution: A single rule with multiple
ORvalues (e.g.,IS_INsupplier=A OR B OR C OR D) often skews toward the first value. In a 60-game placement, you might get 25 games from A, 18 from B, 15 from C, and 2 from D. For better balance, use separate% Tagrules (e.g., 25% for each supplier). - Gaming vs. Sports: Tag Rules are most critical for Gaming’s rigid structure. Sports rules are more dynamic, relying on inventory fields, so Tag Rules may have less impact there.
OR conditions into separate % Tag rules. Always verify in Rec Viewer.Summary
Tag Rules offer a sophisticated way to customize your Gaming placements, from pinning new games to ensuring supplier diversity. By understanding the evaluation flow, rule interactions, and limitations—like overlapping RANK rules or uneven OR distributions—you can craft precise recommendation experiences. Test your rules in the Rec Viewer to ensure your placements look perfect before going live.