Native scores

The Native score is the original score from the review site. Native scores are displayed in the Responses report, and in reports that are limited to data from one source. For example, these reports only use TripAdvisor data, therefore they use the native TripAdvisor scores:

Normalized scores

In reports that combine scores from multiple review sites, social media scores must be adjusted to a common scale, that is, normalized. Social Feedback has two common scales, which are summarized below. The scale chosen for each report depends on the content and purpose of the report.

When creating custom module reports with specific System fields, use the fractional scale if you need to maintain a high level of precision; alternatively, when performing calculations like 5%, use the rounded scale to take care of rounding first. For example, this table shows how Booking native scores (1–10 scale) are normalized to fractional and rounded values:

Suppose that Social Feedback needs to find the average score of two Booking reviews with scores of 3 and 9. In that case, I'd be best to use the fractional scale to reduce rounding errors (as shown in the table below):

Alternatively, suppose that Social Feedback needs to calculate the NPS of the same Booking scores. In that case, you should apply the rounded scale to normalize each score to an integer. Otherwise, the score of 9 won't be recognized as a promoter (that is, 4.5 is not a promoter score, but 5 is a promoter score).

Fractional scores (0.0 to 5.0)

When aggregating scores from multiple review sites, Social Feedback often normalizes the scores with the fractional formula to preserve a high level of accuracy. Social Feedback uses this formula to normalize social scores to a 0 to 5 scale:

round(100*native_score/ max_native_score)/20

For example, in the report below, scores from social reviews are aggregated to present a high-level overview of the data. The scores from sites that do not use a 0–5 scale (such as Booking.com) have been normalized to a fractional scale:

This table shows how scores for popular review sites are normalized to a 0.0–5.0 value, using this formula: round(100*native_score/ max_native_score)/20. For example, a Booking score of 3 (out of 1–10) would be normalized to 1.5.

Rounded scores (1 to 5)

Some calculations, such as frequency or NPS, require integer values. When creating reports that use these calculations, normalizing the scores with the rounded formula can be the best option. Social Feedback uses this formula to normalize social scores to an integer between 1 and 5 (inclusive):

round(5*round(Native_Score) /Max_Native_Score)

For example, this module divides review scores into three categories of integers, so each score must be rounded to the nearest integer:

This table shows how scores for review sites are normalized to a 1–5 rounded value, using this formula: max (1, round(5*round(Native_Score) /Max_Native_Score)). Some of these sites accepts fractional scores. For example, a Booking score of 3 (out of 1–10) would be normalized to 2. While this table doesn't show fractional native scores, the calculation could be done using the same formula.

Scores for Social Networking sources

Social Networking sources do not have satisfaction scores or ratings. To address this issue, Social Feedback leverages Medallia AI SocialStar to assign a satisfaction sentiment — equivalent a star-rating — to social networking sources posts, using a 1 to 5 scale, where 1= Strongly negative and 5= Strongly positive.

To see the overall score of a post or comment, click arrow. The following image shows a Twitter feedback that is Mildly positive.

In cases a rating is incorrectly assigned, you can manually click on the right sentiment from the Responses List report.

When the feedback from these social network posts is aggregated into Satisfaction reports, the rating assigned to each post corresponds to the star-rating each post is assigned: