Question 1

What does trustd look for in Takealot reviews?

Accepted Answer

trustd fetches every review from Takealot's public API and runs two layers of anomaly detection: duplicate reviews (the same account posting multiple reviews on the same product) and identity manipulation (the same account posting reviews under different display names on the same product).

Question 2

Is every detected anomaly considered fraud?

Accepted Answer

No. trustd classifies each detection by severity. Duplicate reviews are treated as innocent — extra copies are removed but the first review is kept. Name changes with the same initial are likely profile updates (downweighted to 70%). Different-initial name changes are suspicious (downweighted to 50%). Three or more different names from one account is a strong manipulation signal and those extra reviews are removed entirely.

Question 3

How is the Trustd Rating calculated?

Accepted Answer

The Trustd Rating uses weighted averaging. Every review starts with a weight of 1.0. When an anomaly is detected, the first review from that account always keeps full weight. Extra reviews are either removed (weight 0) or downweighted (weight 0.5–0.7) depending on severity. The result reflects genuine customer sentiment more accurately.

Question 4

Where does trustd get its data?

Accepted Answer

All data comes directly from Takealot's public product review API. Reviews are fetched in real time when you submit a URL. You are seeing the same reviews visible on the Takealot product page, just analysed for manipulation.

Question 5

What are the limitations of trustd's analysis?

Accepted Answer

The analysis detects duplicate reviews and identity manipulation based on customer account data visible in the public API. It does not yet detect incentivised reviews, coordinated review networks across multiple products, or bot-generated text. Detection methods are continuously expanding.

Takealot Review Checker