How WeChat Serves Tailored Ads: Inside the Recommendation Algorithm
This article explains the content‑based recommendation technique behind WeChat Moments ads, illustrates how user behavior is matched to ad attributes, and offers practical tips for influencing the system to display high‑value ads such as BMW.
1. Content Recommendation
Content‑based Recommendation extracts common attributes from a user’s historical behavior, compares them with ad attributes, and pushes the most similar ad.
Principle : Identify shared attributes between user actions and ad metadata, compute similarity, and recommend the highest‑scoring ad.
Steps :
Collect user historical behavior.
Identify shared attributes related to the behavior.
Calculate similarity between these attributes and all ad attributes (e.g., tag overlap).
Recommend the ad with the highest similarity score.
Example :
User A posted: “Youth is great, wish I stay 24 forever.”
Visited: “Chef O2O, want to try today.”
Commented: “My best friend is #2.”
The system infers a profile such as {twenties, female, trendy, kitchen}. This profile matches best with a Vivo phone ad tagged {white‑collar, selfie}, so the Vivo ad is shown.
Content‑based recommendation is common in ad delivery, offers good interpretability, can be computed offline, and has moderate implementation difficulty.
2. How to Increase the Chance of Seeing a BMW Ad
To receive a BMW ad, your profile must be similar to the BMW ad’s feature set, which might include tags like {wealthy, luxury car}. Suggested actions:
Post WeChat moments mentioning luxury cars, e.g., “It’s time to buy a BMW.”
Read articles about BMW models.
Subscribe to public accounts related to luxury cars.
Reply to messages about BMWs.
Additionally, signal wealth:
Deposit a large amount in WeChat Wallet.
Transfer sizable amounts with friends repeatedly to increase transaction volume.
Use car‑related features in QQ Music.
Change your WeChat signature to something like “Looking for a ten‑million‑yuan BMW, recommendations welcome.”
3. Conclusion
Recommendation is a complex process that considers gender, age, preferences, consumption history, location data, and even appearance. Building a user‑tag system, collecting data, training models, and delivering personalized ads is a challenging technical problem.
It is also rumored that sharing a “thousand‑yuan red packet” with friends can trigger a BMW ad.
Signed-in readers can open the original source through BestHub's protected redirect.
This article has been distilled and summarized from source material, then republished for learning and reference. If you believe it infringes your rights, please contactand we will review it promptly.
21CTO
21CTO (21CTO.com) offers developers community, training, and services, making it your go‑to learning and service platform.
How this landed with the community
Was this worth your time?
0 Comments
Thoughtful readers leave field notes, pushback, and hard-won operational detail here.