Semantic Analysis

Article author
Anastasia Rybalko
  • Updated

Overview

Semantic Analysis is a paid add-on. You can request a trial right from the page or send a message directly to your account manager. 

The Semantic Analysis feature allows you to analyze app reviews and quickly detect issues that lead to user dissatisfaction, lower app ratings, and uninstalls. The Semantic Analysis toolkit is designed to provide a helicopter view of customer experience and detect the app’s potential shortcomings before they become critical. Machine-learning algorithms at the core of the feature analyze the meaning of all reviews, no matter how many of them an app received or what ratings they have.

The Semantic Analysis feature consists of 4 tools:

  1. KPIs
  2. Semantic Charts
  3. Demographic Analysis
  4. Sentiment Analysis

Getting Started With Semantic Analysis

πŸ’‘To start tracking an app's sentiment, add it to your Favorite Apps. The stats will be gathered and processed within the next 24 hours.

To get started with the Semantic Analysis toolkit:

  1. Open the left-hand navigation panel, and select "Semantic Analysis".
  2. Select the apps you want to analyze from the dropdown at the top of the view.
  3. Select the analysis timeframe.
  4. Choose a language, or select "All languages".
  5. Optional: Click on "Add filters" to select a Device Country or App Version.
    πŸ’‘ Note: To select a Device Country for a Google Play app, the app should have an active Reply to Reviews Integration.

In the next sections of this article, you'll learn more about each tool on the Semantic Analysis page.

Available Languages

Semantic Analysis supports 20 languages. There are 3 language groups that are available as add-ons to your plan: 

  1. English and Russian.
  2. Portuguese, German, Italian, Spanish, French, Dutch.
  3. Arabic, Indian, Chinese, Japanese, Korean, Bengali, Hindi, Thai, Vietnamese, Turkish, Urdu, and Persian. 

To add a new language pack, reach out to your account manager or contact our Support Team.

KPI Section

The KPI section is the perfect starting point for understanding your users' sentiment and what mood your app evokes. By default, this section shows KPIs for all semantic categories. To switch to a more granular view, select one of the tabs at the top of the section:

  • Bugs
  • User Feedback
  • Monetization
  • Report a Concern

πŸ’‘To learn more about tags and categories, see our article on Semantic Tags.

Available KPI Metrics for All Categories:

  1. Reviews – how many of the app's reviews were processed by the Semantic Analysis engine in total, and how many were processed per day.
  2. Sentiment Score – the app's overall user satisfaction score. The percentage is calculated based on the correlation between positive and negative reviews. The higher the score, the more positive the user sentiment is.
  3. Avg Review Rating – the average star rating for all semantically analyzed reviews.
  4. Bug reports – reviews tagged with the "Bugs" category.
  5. Reviews to Report – reviews tagged with the "Report a Concern" category.

Semantic_Analysis__Anastasia_Rybalko__AppFollow_2021-07-06_at_4.24.29_PM.jpg

 

Available KPI Metrics for Specific Categories:

  1. Reviews – how many of the app's reviews were processed by the Semantic Analysis engine in total, and how many were processed per day.
  2. Sentiment Score – the app's overall user satisfaction score. The percentage is calculated based on the correlation between positive and negative reviews. The higher the score, the more positive the user sentiment is.
  3. Avg Review Rating – the average star rating for all semantically analyzed reviews.
  4. Category Share – the share of reviews from this category out of all analyzed reviews.
  5. Top Semantic Tag – the top Semantic Tag from the selected category based on the number of reviews with this tag.

Semantic Charts

The Semantic Chart shows you the share of reviews with different categories or tags. You can adjust the view by switching between tabs at the top of the page.

To view the Semantic Category Chart, select the "All" tab at the top of the Semantic Analysis page. 

To view the Semantic Tag Chart, select "Bugs", "Monetization", "User Feedback", or "Report a Concern". To learn more about the available tags and categories, head over to our Semantic Tags article.

On the pie chart, you'll see the number of reviews with specific tags or with tags from different categories. Hover your cursor over a section to see the share of reviews with that attribute. 

If you have tags from more than 5 categories, we'll display the top 5 categories in this chart. 

In the table, you'll see columns with the following information about a tag or category:

  • Tag or category name
  • Number of reviews with the selected tag or from the selected category
  • Sentiment Score
  • Average rating
  • Reply rate
  • Tag share
  • Average reply time
  • Average Reply Effect

To edit the table to add or remove columns, click on the "Manage Table" dropdown menu in the top-right corner of the section.

When you select a category, for example, Monetization, pay attention to the different metrics provided in the KPI Section and the number of reviews by Semantic Tag in the Semantic Charts.


The KPI Section shows the number of reviews with at least one tag from the selected category, while the Semantic Charts provide a breakdown by each tag. One review can have more than one tag, so the total in the "Reviews" column in the Semantic Charts is different from the "Reviews" in the KPI Section.

Demographic Analysis

Have you ever wondered what your clients from Australia are writing to you about? Or maybe you want to understand the reason behind a low Sentiment Score from Spanish-speaking users? The Demographic Analysis section has you covered: use this view to dive into Sentiment Scores, top tags, and categories by language, country, or region.

Use the tabs at the top of the Map View to choose the attribute you want to analyze: Language, Country, or Region. The Map View shows the top 5 locations or languages. Hover your cursor over the map to see the selected location's:

  • Number of reviews
  • Sentiment score
  • Top semantic category

On the Map View, languages are shown for those countries where they are the official language.

In the table to the right of the map, you'll see some additional information. If you have All Categories selected at the top of the page, you'll see the Sentiment Score for the location or language during the selected timeframe. If you selected a specific category, you'll see information about the top tag from that category.

Scroll down to the Countries Summary section to dive into the details.

If you're looking at All Categories, you'll see additional stats about the app's reviews, average review ratings, and the share of reviews about Bugs and User Feedback.

If you're looking at a specific category, you'll see the top tag and its share in the selected category.

Sentiment Analysis

The Sentiment Analysis chart shows you how the overall sentiment and mood of reviews have changed during the selected timeframe. On the right side of the chart, you can evaluate how the current total sentiment compares to the score from the previous period.

Report an Incorrect Semantic Tag

If you notice that an incorrect Semantic Tag was assigned to a review, you can report the Semantic Tag as incorrect. To do this:

  1. Open the Reply to Reviews page in AppFollow.
  2. Click on the Semantic Tag on the Review card.
  3. Select "Report as incorrect". The tag will be removed from the review, and you'll see only relevant reviews when you apply a filter for this Semantic Tag. 

Whenever you report an incorrect tag, we use this information to improve and update our Semantics model to make it more accurate.

The machine-learning model is updated on a quarterly basis. The model is based on recalls and precision:

  1. How many reviews that should have the tag are tagged?
  2. How many of those reviews that have been tagged are tagged correctly?

 

Was this article helpful?

0 out of 0 found this helpful