Engine Status, algorithms and recommendation limits
How core engine settings control what Liva 7 shows, how it behaves and how far personalisation can go
Overview
Engine Status, algorithm selection and recommendation limits define the boundaries of how Liva 7 personalises your storefront. These settings determine whether the engine is active, which decision model it uses and how many products it is allowed to return in different situations.
This article explains each setting in clear, merchant‑friendly terms so you understand exactly how they shape recommendations across widgets, Smart Content and Smart Bundles.
Engine Status
Engine Status controls whether personalisation is active across your storefront.
When the engine is on, Liva 7 evaluates signals, rules and product relationships in real time. When it is off, widgets and bundles still appear, but they use static fallback logic instead of personalised output.
Engine Status affects:
- whether recommendations adapt to shopper behaviour
- whether Smart Content uses behavioural qualification
- whether bundles use AI‑driven combinations
- whether fallback rules take priority
Engine Status must be on for real personalisation to happen.
What happens when the engine is off
Turning the engine off switches all systems to predictable, non‑personalised behaviour.
This mode is useful for testing, theme changes or troubleshooting. When off:
- Widgets show default or pinned products
- Smart Content uses only segment rules, not behaviour
- Bundles use manual combinations only
- No real‑time adaptation occurs
This ensures your storefront remains stable even when personalisation is paused.
Engine off means static output, not broken output.
Algorithm selection
The algorithm setting controls the style of recommendations the engine produces. Each option prioritises a different logic pattern, giving you control over how your storefront behaves without needing anything technical.
The available algorithms are:
- Smart Mix (Recommended) - A balanced, adaptive model that blends behaviour, product relationships and performance. Best for most stores.
- Customers Also Bought - Prioritises items frequently purchased together. Strong for cross-sell patterns.
- Similar Products - Focuses on items closely related to what the shopper is viewing. Ideal for deep catalogues.
- Based on Browsing History - Leans heavily on what the shopper has viewed in the session. Good for discovery-led stores.
- Based on Purchase History - Prioritises items related to past orders. Useful for stores with strong repeat behaviour.
- Trending Now - Surfaces products with high recent engagement or sales. Great during peak periods or promotions.
- New Arrivals - Highlights the newest products in your catalogue. Ideal for stores with frequent drops or seasonal refreshes.
Merchants typically switch algorithms when they want to:
- push newness or seasonal items
- strengthen cross-sell behaviour
- tighten relevance on product pages
- encourage discovery on the homepage
- lean into trending items
- support repeat-purchase patterns
Algorithms change the engine’s priorities, not its intelligence.
How algorithms behave across different page types
The same algorithm behaves differently depending on where the shopper is. Page intent shapes how the engine interprets the chosen logic.
- homepage - broader, discovery-focused ranking
- collection pages - category-level relevance
- product pages - similarity and complementarity
- search pages - keyword and semantic matching
- cart - high-intent add-on logic
This ensures recommendations always match the shopper's stage in the journey.
Algorithms adapt to page intent automatically.
Minimum recommendation limits
Minimum limits define the fewest products the engine must return before fallback logic applies. If the engine cannot find enough relevant items, fallback behaviour fills the gaps.
Minimum limits help you:
- avoid empty widgets
- maintain layout consistency
- ensure predictable storefront behaviour
If the minimum is set to 4 and only 2 relevant items exist, fallback rules fill the remaining slots.
Minimum limits prevent gaps in your storefront.
Maximum recommendation limits
Maximum limits cap how many products the engine is allowed to return. This keeps recommendations focused and prevents overcrowding.
Maximum limits help you:
- control layout density
- avoid overwhelming shoppers
- keep widgets visually consistent
- maintain performance on mobile
The engine always returns the highest-ranking items within the limit.
Maximum limits keep recommendations tight and intentional.
How limits interact with rules
Limits define boundaries, but rules define priorities. When conflicts occur, rules win.
- pinned products always appear, even if they exceed the maximum
- excluded products never appear, even if the engine wants to show them
- fallback rules fill gaps when minimums are not met
This creates a predictable, merchant-controlled environment.
Rules take priority over limits when conflicts occur.
How limits affect widgets, Smart Content and bundles
Each system uses limits differently based on its purpose.
- widgets use limits to control how many products appear
- Smart Content uses limits only when content includes product lists
- bundles use limits to define how many items can be included
This ensures each system behaves consistently without interfering with the others.
Limits shape output differently depending on the system.
When to adjust engine settings
Adjust settings when your goals, catalogue or traffic patterns change. You may want to update settings when:
- your catalogue grows or shrinks
- you launch seasonal campaigns
- you want more diversity or more precision
- you want to test different merchandising strategies
- you need predictable behaviour during theme changes
Small adjustments can significantly change how personalisation feels.
Engine settings should evolve as your store evolves.
How these settings shape the overall experience
Engine Status, algorithms and limits define the framework the engine works within. They determine:
- how adaptive the engine is
- how recommendations are ranked
- how many products appear
- how rules and fallback logic behave
Together, they ensure personalisation is consistent, predictable and aligned with your goals.
