Contextual personalization
Contextual personalization is a Loomi AI-powered engine that finds the best variant for each individual customer — not just the best variant on average. It analyzes each customer's unique situation and continuously learns from their interactions to improve over time.
This is a Premium tier feature powered by Loomi AI.
How contextual personalization works
When a customer encounters a weblayer or receives a message, contextual personalization:
- Evaluates their context — device, location, time of day, browsing behavior, and purchase history.
- Selects the variant most likely to achieve your business goal.
- Registers whether the customer completed your goal — for example, making a purchase.
- Uses that outcome to make better decisions for future customers.
The model is continuously retrained in the background as new interactions become available.
Supported channels
You ou can use contextual personalization across weblayers and all scenario channels — email, mobile push, browser push, SMS, RCS, WhatsApp, App Inbox, and in-app personalization.
Webhooks and Ad Audiences don't support automatic interaction tracking and can't be used as optimization targets.
Contextual personalization versus A/B testing
A/B testing finds the single best-performing variant for your entire audience. Contextual personalization finds the best variant for each individual customer.
Example: 70% of your customers respond to discount offers and 30% respond to free shipping. A standard A/B test declares the discount variant the winner and shows it to everyone — including the 30% who would have converted better with a free shipping message. Contextual personalization serves each customer the variant that fits their situation.
Instead of "Which variant is best for everyone?" ask, "Which variant is best for each customer?"
Contextual personalization also:
- Uses AI instead of human-defined rules.
- Learns continuously without waiting for statistical significance thresholds.
- Delivers personalization for every audience segment — not just the majority.
Case study: Deploying contextually personalized banners for bimago led to a 44% increase in subscription conversion rate compared to standard A/B testing. Read the full case study.
Key concepts
| Term | Definition |
|---|---|
| Variant | The weblayer, email, or scenario channel option served to a customer. You can define up to 10 variants. |
| Context | A customer's specific situation — device, browser, country, time of day, and behavioral history. |
| Contextual feature | A data point the model uses to understand context, such as a customer's country or number of purchases. |
| Goal | The customer action that signals a variant succeeded — for example, a purchase or a banner click. Loomi AI uses your goal to learn which variants drive the outcomes that matter to your business. |
| Value | A per-variant signal that reflects its business worth to you. Loomi AI uses value ratios to learn which variants to prioritize. |
| Reinforcement learning | The AI approach Loomi AI uses to continuously improve by learning from each interaction and its outcome. |
How variants, value, and goal work together
These three concepts work as a system. Your variants are the options Loomi AI chooses from. Your goal tells Loomi AI when a variant succeeded — for example, when a customer made a purchase. Your value tells Loomi AI how much each variant's success is worth to your business. Together, they give Loomi AI everything it needs to find the best variant for each individual customer.
Use cases
Contextual personalization works across weblayers and all scenario channels. For detailed examples by industry and channel, see Contextual personalization use cases.
Get started
Updated 3 days ago