Yotpo Insights FAQ

By default, all users of your account can access Insights. If you want to create users with exclusive access to Yotpo Insights, our account admins can designate an Insights Only user role, which restricts access to the Yotpo workspace.

To use Insights, your store's default language must be English and your reviews must be predominantly in English.

You’ll get value from Yotpo Insights whether you’re a seasoned merchant with 1000s of reviews or if you’re a new merchant just starting out. Once you reach 10 reviews, you’ll start getting topics generated by our AI algorithm.

The Insights dashboard displays topics which were mentioned in customer reviews according to the selected filters. Topics can also be searched by topic name via the Insights search bar.

If you're unable to locate a particular topic, try to reset your filters.

Note: Topics with fewer than 10 opinions will not be displayed or returned in search results.

Yes, products on your blocklist are shown in the Insights dashboard.

The number of topics displayed in your Insights dashboard is based on your review volume. You can expect to see 10-20 topics (with at least 10 opinions each) for every 1k reviews.

Insights only analyzes organically generated reviews. 

As syndicated reviews are cloned, they are not analyzed.

Currently, it is not possible to manually change sentiment.

Currently, it is not possible to manually change keyword grouping.

Yes! You may filter products by one or more product collections to get insights on a specificcollection or compare multiple collections.

Note: New product collections may be created via product catalog import. If you're missing the Product Collections field in your product catalog, please contact your Yotpo representative for assistance. 

Insights also allows you to view a topic breakdown by product collection.

The overall sentiment accuracy of Insights stands at 91%.

New data is refreshed and analyzed every 24 hours.

Yotpo analyzes all historical data from textual customer reviews.

Data that is not analyzed:

• Currently, Yotpo does not analyze content from Custom Fields or Custom Review Forms, Q&A, Emojis, and Photos. 

• Non-English Reviews:
  Currently, Insights is only able to gather insights from English reviews exclusively. To use Insights, your store's default language must be English. Stores with English as the default language, but non-English reviews can use Insights, however, data accuracy and integrity may be adversely affected by non-English reviews. 

Yotpo's algorithm identifies real opinions on specific topics within the text of reviews.

Sentiment is then calculated for each paired topic and opinion using a Deep Learning statistical model, and is classified as a Negative or Positive sentiment.

The total sentiment score of a topic is then calculated by subtracting the percent of negative opinions from the percent of positive opinions (like a Net Promoter Score/NPS).

The scale of topic sentiment score ranges from -100 to +100 where -100 is the most negative topic sentiment and +100 is the most positive topic sentiment. Like the star-rating distribution (1-5 stars), most topics have a positive sentiment score.

Sentiment Change compares the change in a topic's sentiment score between two date ranges. The difference of the two scores is represented as the sentiment change per the selected date range. For example, if your date range is set to Past Month, the sentiment change will reflect the change in sentiment when comparing data from the past month to the month before it.

Note: Sentiment change may appear as N/A in cases where there is insufficient data to calculate the sentiment score.

Insights automatically compares topic sentiment with over a dozen Yotpo accounts within your store's industry. This is calculated by looking at the average sentiment score for a given topic within each industry.

Yes! The Insights algorithm can identify more than one opinion and sentiment per topic. This also includes conflicting sentiments towards different topics within the same review.

For example, "The socks were very comfortable but the delivery was slow".

When analyzing Insights for a topic (which includes more than one product), only organic reviews are factored. Grouped (cloned) reviews are excluded.

Note: Grouped (cloned) products will still appear in the product drilldown but may not have any organic reviews and opinions.

If you've come across a misidentified sentiment or opinion in reviews, we encourage you to report the issue so that we can improve our algorithm accordingly. Such issues may include a sentiment with incorrect polarity, a phrase that was misinterpreted as an opinion, or other discrepancies in sentiment or data integrity.

Discrepancies may be reported by clicking the additional info icon when reading reviews. Note that reporting issues will not result in any immediate changes to your dashboard. We'll review your input and periodically refine our algorithms according to your feedback.

As Insights is able to identify multiple opinions per sentence, each row in your export file represents a unique opinion and sentiment found in a sentence within a review. To view your most positive or negative opinions, simply sort or filter the Opinion Sentiment column according to the desired opinion polarity.

Reviewer location is based on the IP address of the reviewer at the time the review was submitted.