Repurchase Strategy in Game Item Recommendation: Scenarios, Challenges, and Implementation

Repurchase strategy, which recommends items a user has previously purchased, is widely used in recommendation systems and underlies collaborative filtering algorithms.

In everyday e‑commerce and game shopping, repurchase recommendations often feel delayed or redundant, leading to user confusion and wasted recommendation opportunities, especially when the user already has a stable repurchase tendency.

1. Recommendation Scenarios

Game item recommendation includes several distinct scenarios:

Whether the recommendation opportunity is permanent (always visible on the shop homepage or sub‑pages) or non‑permanent (gift packs with time limits) or refreshable (players can refresh the displayed items).

Whether the recommended item carries a discount.

Whether the content is limited to appearance items, which generally have lower purchase rates and higher volatility.

The purpose of the scenario, such as simplifying the purchase flow, driving revenue, or introducing new items.

Typical scenarios are:

Shop homepage permanent recommendation – usually without discount and may be further divided by whether only appearance items are shown.

Random limited‑time gift packs – typically include discounts and the items cannot be changed after release.

Refresh recommendation – often presents ability‑related items, includes discounts, and the content is fixed for each refresh.

2. Repurchase Strategy and Each Recommendation Scenario

Shop homepage permanent recommendation

Low user attention: many players ignore static recommendation slots, making non‑discounted suggestions ineffective.

Appearance item pairing: appearance items often require combination with other items, so users tend to browse deeper into the shop rather than purchase from the homepage.

Attribution difficulty: purchases made after viewing the homepage may be hard to attribute to the recommendation.

Diversity and new items: over‑reliance on repurchase can make the homepage monotonous and hurt player experience.

Random limited‑time gift packs

Reduced surprise and potential recommendation penalty: the surprise factor drives impulse purchases; if the gift pack does not meet expectations, users may ignore the feature.

Timing impact: different game contexts create varying demand, so triggering the repurchase strategy at the wrong moment harms effectiveness.

Gift‑pack composition: when a player is likely to buy a specific item, adding complementary items can restore surprise.

Refresh recommendation

Real‑time requirement: the system must analyze recent purchase behavior instantly and present the optimal item at the moment of refresh.

Item ordering: because refresh opportunities are scarce, high‑value items should be placed prominently while low‑interest items occupy peripheral slots.

3. Application of Repurchase Strategy

Items are classified into four categories based on two dimensions: continuous purchase density (A) and repurchase recommendation purchase rate (B).

Category 1: A high + B high – stable repurchase items.

Category 2: A high / medium + B medium – items with moderate repurchase demand.

Category 3: A low + B low – items with low repurchase demand, requiring audience expansion.

Category 4: A missing + B low – items with minimal repurchase potential.

For each category, specific tactics are suggested:

Category 1: consider whether “using” the item counts as a purchase, set purchase cooldowns, define a repurchase window, possibly raise purchase thresholds, and analyze failures to update user profiles.

Category 2: combine with bundling strategies, pairing with Category 1 items.

Category 3: identify whether the item is one‑time or repeatable, refine user portraits, run experimental pushes, and eventually remove from repurchase strategy.

Category 4: monitor when the item upgrades to a higher category, especially during version updates.

4. Summary

The article outlines the constraints and considerations of repurchase strategies across major recommendation scenarios and shares practical ideas for applying the strategy in the random limited‑time gift‑pack context. In practice, repurchase strategy serves as a supplemental tactic after algorithmic models, helping to fine‑tune recommendation performance.