Custom Formula is a ranking strategy available in Decision Policies that replaces a single-variable ranking signal with a weighted combination of multiple parameters. You choose which parameters matter and assign a percentage weight to each. The engine multiplies each parameter score by its weight, sums the results, and ranks all eligible Offerings for every customer accordingly.
The result is a ranking that reflects both what is strategically important to your business and what is personally relevant to each customer — recalculated in real time, without manual intervention.
Custom Formula = (Attribute 1 score × Weight%) + (Attribute 2 score × Weight%) + … + (Performance score × Weight%) + (Priority score × Weight%)
How it compares to other ranking strategies
MoEngage offers three ranking strategies in Decision Policies. Understanding when each one fits helps you choose the right approach or combine them across different policies.
| Manual priority | Merlin AI | Custom formula | |
|---|---|---|---|
| How ranking works | Author assigns a rank (100, 90, 80…) to each offering manually | AI learns from clicks and conversions to find the best-performing offering for each user | Marketer defines a weighted formula combining Offering attributes, performance, and priority |
| Who sets the score | Author | Algorithm (learns over time) | Marketer defines weights; engine calculates score per user at runtime |
| Personalisation | None — same rank for all users | Implicit — learns aggregate patterns across users | Explicit — score changes per user based on their profile (via dynamic attributes) |
| Business goal control | High — author fully controls rank | Low — optimises for engagement/conversion, not specific business goals | High — marketer decides how much weight each goal carries |
| Responds to strategy shifts | Manually — re-rank all offerings | Gradually — model needs time to relearn | Immediately — adjust weight in the Decision Policy; all Offerings re-rank instantly |
| Best for | Small, stable catalogs with clear priority | Large catalogs where engagement is the primary goal | When business value AND user relevance both need to influence ranking |
| Requires Offering Attributes | No | No | Yes |
| Limitation | Does not account for customer intent or business value trade-offs | Cannot protect high-value, low-CTR Offerings from being buried | Requires upfront configuration of attributes and weights |
| info |
Information Manual Priority does not account for the customer. Merlin AI does not protect high-value, low-CTR offerings. Custom Formula is the only strategy that lets you simultaneously express what an Offering is worth to the business and how relevant it is to this specific customer. |
Parameters available in Custom Formula
A Custom Formula can include up to 5 dynamic parameters in addition to static parameters. Parameters fall into two categories.
Dynamic parameters — Offering Attributes
These are attributes you configure under Settings > Offering Decisioning > Offering Attributes. You can include up to 5 offering attributes — Fixed Value, Dynamic Value, or a mix — as parameters in a formula.
Fixed Value attributes capture the strategic importance of the Offering itself. The score is set by the author at Offering creation time and is the same for every customer. Examples: business objective, margin tier, campaign urgency.
Dynamic Value attributes capture relevance to the individual customer. The score is calculated at decisioning time by matching the offer's associated value against the customer's live MoEngage profile. Examples: category affinity, churn propensity, purchase recency.
Static parameters
Two built-in parameters are always available regardless of attribute configuration:
- Offering Performance — the historical Click-Through Rate (CTR) or Conversion Rate (CVR) of the offer.
- Offering Priority — the manual priority rank assigned to the Offering (1 to 100). The priority number itself becomes the base score.
| info |
Information All weights across all selected parameters must sum to exactly 100%. The formula builder will prevent saving if the total is not 100. |
How to set up Custom Formula ranking
Before you begin
If you have not yet set up attributes, navigate to Settings > Offering Decisioning > Offering Attributes and create at least one attribute before proceeding. Refer to the Create offering attributes guide for instructions.
Step 1 — Create or edit a Decision Policy
Navigate to the Decision Policy you want to configure, or create a new one. In the Ranking Strategy section, select Custom Formula from the available options.
Step 2 — Add parameters
Click Add offering attribute (0/5) in the Custom Formula builder. You can add:
- Any offering attributes you have configured (Fixed Value or Dynamic Value)
- Offering Performance (CTR or CVR)
- Offering Priority
You can add up to 5 parameters in total. If the Add Parameter button is disabled, you have reached the maximum.
Step 3 — Assign weights
For each parameter, enter a weight percentage. Weights represent how much each parameter contributes to the final score. All weights must sum to exactly 100%.
Example weight configuration
- Business Objective (Fixed Value attribute) — 40%
- User Intent / Category Affinity (Dynamic Value attribute) — 30%
- Offering Performance (CTR) — 20%
- Offering Priority — 10%
Step 4 — Save and publish
Once all weights sum to 100%, save the Decision Policy. The Custom Formula becomes active immediately for all Offerings within that policy. You do not need to edit individual Offerings — the engine uses the formula to score and rank them automatically for each customer.
How the score calculation works
When a customer qualifies for multiple Offerings in a Decision Policy, the engine calculates a Custom Score for each eligible Offering and ranks them in descending order. The highest-scoring Offering is shown to the customer.
Calculation formula
Custom Score =
(Attribute 1 Score × Weight%) +
(Attribute 2 Score × Weight%) + … +
(Performance Score × Weight%) +
(Priority Score × Weight%)
Worked example
A Decision Policy contains two offers: Offering A (iPhone launch) and Offering B (phone case). The formula is configured with the weights shown above. Customer X has "Electronics" as their primary category affinity; Customer Y does not have an electronics affinity.
Scores for Customer X:
| Parameter | Weight | Offering A score | Offering B score | Weighted contribution |
|---|---|---|---|---|
| Business Objective (Fixed) | 40% | 90 | 75 | 35.00 (90 × 0.40) | 30.00 (75 × 0.40) |
| User Intent (Dynamic) | 30% | 100 | 0 | 30.00 (100 × 0.30) | 0.00 (0 × 0.30) |
| Offering Performance (CTR) | 20% | 60 | 80 | 12.00 (60 × 0.20) | 16.00 (80 × 0.20) |
| Offering Priority | 10% | 70 | 70 | 7.00 (70 × 0.10) | 7.00 (70 × 0.10) |
| Total Custom Score | 100% | Offering A: 84.00 | Offering B: 53.00 |
Result for Customer X
Offering A scores 84.00, Offering B scores 53.00. Offering A is ranked first — the iPhone launch is shown to Customer X despite having a lower CTR than the phone case, because the business objective weight and the matched category affinity score more than compensate.
For Customer Y, whose category affinity does not match Electronics, the User Intent score for Offering A would be 0 instead of 100. Offering A's total score would drop to 54.00 (35 + 0 + 12 + 7), and the ranking would shift accordingly — all without any changes to the Offerings or the policy.
What happens when two offerings have the same score?
If two or more Offerings in a Decision Policy calculate to the same Custom Score for a given customer, the engine applies a sequential tie-breaking logic to determine which Offering is ranked first:
| Step | Condition | Logic |
|---|---|---|
| 1 | Earliest end date & time | The Offering expiring soonest wins — it is more time-critical. |
| 2 | Earliest start date & time (if end dates match) | Among equally expiring offers, the one that went live first wins. |
| 3 | Random selection (if start & end dates match) | One Offering is chosen at random. |
Next Steps
- To configure the Offering Attributes used by this strategy, refer to Managing Offering Attributes.
- To learn how to create a Decision Policy, refer to Create Decision Policy.
- To assign Offering Attributes to your offerings, refer to Create Offerings.