The StampedIQ Sentiment & Topic Analysis features both use advanced algorithms to analyze the text of reviews and provide you with additional tools for managing your reviews. Identify reviews that need further investigation with StampedIQ Sentiment, or use the Topics Filter feature to highlight frequently mentioned keywords and phrases to other customers.
Follow this guide to learn more!
Key Notes
- The following languages can be used:
Arabic Vietnamese
Japanese Italian
Dutch Thai
English Spanish
French Portuguese
German Polish
Indonesian Korean
Chinese (Simplified and Traditional)
In This Guide
Introduction To StampedIQ: Sentiment & Topics Analysis
Our AI-enabled feature analyzes review content to determine whether a review matches the rating given by customers and provides a detailed report of the sentiments expressed. This feature helps you identify reviews that need further investigation, so you can take appropriate action.
An example of an analyzed review in the dashboard:
The feature also helps prevent false-positive reviews from the Social Push and Ads Centre, ensuring that only the best reviews get shown to customers/shoppers.
Sentiment Analysis will flag the review to highlight that the review requires attention and it will be shown in the 'Flagged' folder. These flagged reviews will have to be manually published by the owner before they can show up in the display widgets.
How to Enable
- In the Stamped Dashboard, go to Settings > StampedIQ
- Click Enable beside the StampedIQ: Sentiment & Topics Analysis feature
- Click Save in the bottom right corner
Topics Filter
The Topics Filter feature allows you to display frequently mentioned keywords and phrases in the reviews widget so customers can filter reviews on your product pages. These keywords are automatically generated on each review through our AI algorithm.
Topics only generate on a product-scoped widget. The main widget must be displaying reviews for a single product. If your widget is set to show all reviews across products, such as a full-page or all-reviews layout, topics will not be generated and the filter will not appear. Topics are also built only from your own local reviews for that product, so syndicated and grouped reviews are excluded, and the filter shows only once at least three keywords have been generated.
Example of what a review in your dashboard will look like:
Here's an example of the keyword filters in the main widget on product pages:
Customizing Filters
You can set a topics filter manually with the data-keywords attribute on the main widget code. A maximum of 10 keywords is supported, each separated by a comma. These customised filters apply to the main widget on all product pages.
This is the recommended approach for all-reviews or full-page widgets, where automatic topics are not generated.
<div id="stamped-main-widget"
data-widget-style="standard"
data-keywords="example tag, example tag2, example tag3">To exclude certain words from the Topic Filter, enter them here:
https://go.stamped.io/v3/#/settings/insights
How to Enable
- In the Stamped Dashboard, go to Settings > StampedIQ
https://go.stamped.io/v3/#/settings/insights - Click 'Enable' beside the StampedIQ: Topics Filter feature
- Click Save in the bottom right corner
Topics filter not appearing
Refer to the troubleshooting steps below:
- Check the widget scope. Topics only generate when the main widget is showing a single product. On an all-reviews or full-page widget they are not generated. Use data-keywords instead, as described in Customizing Filters.
- Check the keyword count. The filter only appears once at least three keywords have been generated for that product. Products with few reviews often will not reach three.
- Check where the reviews came from. Keywords are generated only from your account's own local reviews for that product. Syndicated reviews and reviews shared in from a Product Group are excluded, so a product whose reviews are mostly shared or syndicated may not generate topics.
- Allow for caching. Widget content is cached, please allow a few hours and check in an incognito window.