If your app has been experiencing a decline in app downloads for apple devices for the last several weeks, Google and iTunes may be to blame. It appears that there is a bug causing iTunes to display zero current ratings causing Google to extract inaccurate structured data from those app pages. As a result, an inaccurate rich snippet is displayed in the search results.
Currently for many app publishers, their app appears with a 1-star rating when someone searches for the app in Google despite what the actual rating is on the iTunes page. This issue was reported on Google’s product forum October 29. One member claimed they spoke to an advisor over at Apple who responded:
“We’ve actually seen a few cases like this as well from other developers. It appears that if your application for its current version does not have enough ratings to display an average, Google’s search results displays the “null” item as a one star rating.”
However, this surely isn’t the case as this is happening for major apps including Netflix, Snapchat, Pandora, OpenTable UK, Travelocity, Discover, Hulu, and even Amazon.
As you can see, the iTunes result appears with the incorrect 1-star rating whereas the Google Play result appears with the correct rating. If an iPhone user searches for the Amazon app on Google, the Google Play result (with the positive rating) will not appear so the user only sees the iTunes result with the 1-star rating. As a result, it is not likely that the user will download the app. However, if the user clicks into the iTunes result, they will see that the current version of the app actually has an aggregate rating of 3.5 stars based on 390 reviews.
Below are more examples of incorrect ratings appearing in Google Search results:
Netflix
Snapchat
Discover
Pandora
Travelocity
OpenTable UK
As seen in the above examples, this issue is not singling out any particular category. It’s happening across retail apps, travel apps, music apps, video apps, and even social media apps. Since this issue is unique to iOS apps, at first glance it is unknown if this is an issue with how iTunes structures their pages and data, if it’s a problem with how Google is indexing these pages, or a combination of both.
However, when running any of these iTunes URLS through the Google Structured Data Testing Tool, the correct microdata markup is present on the pages and the correct information is populated in the rich snippets (within the tool).
However, when looking at the cached version of these pages there is a common denominator; under “Customer Ratings” where current ratings usually appear is the message:
“We have not received enough ratings to display an average for the current version of this application”
Highlighting that text and clicking inspect element reveals the possible culprit. Examining the schema data reveals the values for “ratingValue”, “worstRating”, and “ratingCount” is zero.
An important factor to point out that many of these pages were cached today, displaying the “not enough ratings” message only to have the live page display current ratings. For example, the iTunes page for the Netflix app was cached today (Nov 20, 2015 02:03:34 GMT) and contains no current ratings but the live, uncached version shows 1,665 current ratings. Surely Netflix didn’t gain that many ratings in less that a day so this indicates that there may be a bug in iTunes that returns current ratings as null.
It’s also worth mentioning that this issue seems to occur more in the “app” results tab than in the “web” results tab.
A possible solution is for Apple to remove the schema markup from the “current” ratings and instead markup the “all version” ratings with the structured data. Another solution is for Google to set a threshold for when they display rich snippets such as setting a condition to only display app ratings if the “ratingCount” is greater than a given number.
As companies are investing more time and money in improving their app and promoting it through search, hopefully Google or Apple can quickly resolve this problem. Since PMG services some of the clients mentioned above, we’ll continue to monitor these incorrect search snippets.