The main indicators used to analyze posts of earned media in your Radarly projects are:
- Estimated Reach
- Engagement Score
Impression and Estimated Reach are relevant to compare the share of voice between different platforms and at different period of time. Although the absolute value of the indicators is interesting, it is the order of magnitude and the variation that matter the most.
Impressions of a post
The global audience, the maximum number of users that can see the post at the time of publication
CALCULATION OF IMPRESSIONS
The general formula for social media sources:
Impressions of a post = nb of followers at the moment of posting + (nb of shares of the post* x average number of followers*)
* nb of shares & average number of followers are not available for all social media sources
The general formula for web sources :
Impressions of a post = average monthly nb of visits on the website/nb of new pages created on the website in one month
Estimated Reach of a post
The estimated reach of a post is the estimated number of users exposed to the post at the moment it was published
The estimated reach of a post = impressions of the post x web user's exposure coefficient
Impressions and Estimated Reach
|Source||Impression of a post||Estimated reach of a post||New coef. (jan 2018)|
|Number of followers at the posting moment+ (number of retweets x median number of followers)||impressions x 0.08||0.08|
|Number of fans at the posting moment||impressions x 0.07||0.07|
|Number of followers at the posting moment||impressions x 0.15||0.15|
Number of followers at the posting moment +(number of reposts x median number of followers*)
|impressions x 0.4||0.04|
|N/A for the current version||N/A for the current version||N/A|
|DAILYMOTION||Number of views at the date of publication||Impressions x 0.9||0.9|
|YOUTUBE||Number of views at the date of publication||Impressions x 0.9||0.9|
|WEB (Blog, Website, Media)||Number of visits on the site per month/number of new pages created during 1 month||impression of the post x 0.05||media: 0.95
blog : 0.40
website : 0.05
|VKONTAKTE||Number of fans at the data of publication||impressions x 0.3||0.3|
* the median number of followers on Twitter = 301, the median number of followers on Sina Weibo = 152
Note: the median number of followers and the coefficients to calculate the estimated reach will be reviewed frequently (without retroactivity).
Impressions & Estimated Reach: where to find them?
Listening section - Analytics Overview
- Cumulative impressions of the posts matching the query
- Cumulative estimated-reach of the posts matching the query
- Ability to sort queries by impressions & estimated reach
Listening section - Analytics Details
- For each post, it displays the number of impressions and estimated reach in the list of posts (on the left)
- In the posts page, it displays the number of impressions and estimated reach of the post
- You can sort the list of posts by Impressions and Estimated Reach
- In the tile "Number of posts", it displays the cumulative impressions and estimated reach of the posts matching the query
Listening section - For a single post
- The number of Impressions and Estimated Reach is displayed in the right down corner of a post.
Listening section - Influencers
- For each influencer, the number of Impressions and Estimated Reach is displayed in the influencers list.
Additional information on the methodology- Estimation of the coefficients
« Impression » of posts:
- The impressions of a message represent the total potential audience, the maximum number of people that can be exposed to the post.
- Impressions are estimated differently for different platforms:
- For social media, the impressions come from the author's followers population at the moment of posting (when the data is collected), as well as the number of shares of the publication based on the estimated median of subscribers on the platform.
- For websites, we divide the audience in the month by all the posts captured in the same month. The underline hypothesis: on editorial sites, the audience focuse on the recent publications. We have to make a hypothesis here because without being the owner of the website, it is not possible to have more detailed data.
«The Estimated Reach » of posts:
The Reach of a publication represents the estimated number of users exposed to the publication.
The reach is estimated as follows:
For social media
- Facebook: From the reach provided by Facebook to all customers who have plugged their pages into Radarly Social Performance, we estimate an average reach ratio to the number of fans. Overall, this number is in steady decline. It went from 9% a few months ago to 7%. We reviewed its value in January 2016. For your information, other studies do the same exercise and find fairly comparable results on different panels (e.g: locowise research).
- Twitter & Instagram, the market does not have access to the customers' statistics of reach of a post with the API. Such statistics are available more recently on the web, and they will certainly be available in the future in APIs. To make a sound estimation, we sampled from available Twitters and Instagram statistics shared by our clients.
- Since the information provided is a brief audience number, we have to make an assumption: for a single article, the times of visits and the number of visitors are the same (a visitor rarely reads the same article several times). The challenge of this method is how to distinguish between the exposure rate and the real reading rate. Here is how we do it: for editorial sites such as media, blogs, we use the “bounce rate” to adjust the audience number. The bounce rate excludes the visitors who have not taken enough time to read an article on the web page after the publication, the shorter the visitors stay, the higher the bounce rate.
- The detailed data is not as well listened as that of the website owners. To calculate the estimated reach of websites, we divide the web into three groups: Media, Blog and other websites. The bounce rate of media sites is known to be very low, visitors are the audience(95% of reach), blogs are in an intermediate situation (40% of reach, assumed).
The audience to reach rate is very low (5%) for the rest of the websites, highlighting two behaviors: :
- A high bounce rate: visitors only briefly browse the websites without staying too long on a single page.
- An audience that does not focus solely on new publications (the major difference with editorial sites) but as much on old content, e.g. visitors of an institutional site distribute relatively evenly between old and new contents.
Listening vs Social Performance (Owned vs Earned)
While the goal of the brand is to accurately measure the digital and social performance of its own (owned) publications, the "Social Performance" part of Radarly is particularly suited to this approach.
Missing Statistics (Impression et Estimated Reach)
For your information, certain social media do not permit the capturing of these data in mass (especially Google +, Dailymotion), in this case, Linkfluence chose to not calculate them at the moment.
Note: This methodology will evolve gradually with the increasing transparency of indicators provided by third parties. We remain at your disposal to discuss it in person.
Virality and Engagement: where to find them?
When you have a list of posts in the listening part of your Radarly project (Posts, Clusters, Analytics Details), you have the possibility to categorize publications by virality and engagement score.
How to do that?
By default, the publications are sorted by publication date.
You can change the ranking by clicking on the small arrow (1) or switch between ascending and descending order by clicking on the big arrow (2).
What is the virality score?
Virality quantifies the spread of a post among web users.
How is it calculated?
The calculation of virality is different for each platform:
|Platform||Data used to calculate the virality score|
|Number of retweets|
|The total number of shares (including the share of shares)|
|Number of comments|
|YouTube||Number of web pages, facebook posts and tweets in our database including the URL (or a shorten URL for Twitter) to the video|
|Web||Number of web pages, facebook posts and tweets in our database including the URL (or a shorten URL for Twitter) to the web page|
What does Engagement Score mean?
Engagement score quantifies the public involvment to a post.
How is it calculated?
There are various ways to calculate engagement for each platform:
Number of retweets + number of likes + number of replies to a tweet
Number of shares(including the shares of shares) + number of comments + number of likes (including likes on shared posts)
Amount of likes + amount of comments
Number of likes + number of comments + number of the video being favorited
Number of tweets that include the URL or shortened URL to the webpage in our database + number of shares that include the URL or shortened URL to the webpage on Facebook + number of web pages with a link to this page+ number of comments on the website page
*Sina Weibo = Reposts + Likes + Comments
Note: When you have a list of publications in the listening part of your Radarly project in the publications, clusters, and analytics details views, you have the option to rank publications by engagement score.