For every Twitter Mention of a Twitter page, quintly calculates the time the page needed to respond to it. This article gives you some insights on how quintly calculates the response times of mentions and what data we use for our response times metrics.

Twitter mention vs Twitter question

Whenever a Twitter user mentions your Twitter page in one of his tweets you can see it in the Mentions Table in quintly with it’s respective response time. If a Mention wasn’t answered within the following 7 days we show “n/a” for the response time. If a reply has multiple mentions in it, we won't count it as a reply in our default metrics, however, this is possible in a custom metric for which you can contact our support team.

Screen_Shot_2017-09-04_at_11.52.38.png

Beside the Twitter Mentions quintly also offers you to look at the, what we call, Twitter Questions. In quintly we identify Twitter Questions as Mentions which include at least one question. One main characteristic we look at when classifying Questions is whether the text includes question marks or not. So the Twitter Questions are basically a subset of the Mentions.

If you take a look at the Questions Table in the screenshot below and compare it to the Mentions Table above you can see that the Questions Table holds all Mentions which include at least one question.

Screen_Shot_2017-09-04_at_11.53.02.png

When clients analyze the customer support activities of their pages the response time is one of the most important metrics to look at. Customer support teams mostly focus on Mentions which include questions as questions imply that the user needs help. Because of this we decided to also only use Twitter Questions in the response times metrics you find in quintly. This makes the analysis more accurate as it filters out Mentions which don’t need any replies.

Response times calculation

Technically Mentions and Replies are just normal tweets. This means that you technically don’t have a conversation feed in Twitter. That’s why we say page replies to users questions can also be replies to previous questions of the same users. The following will give you a few examples to understand better how it works.

Example 1:

12th of August 2017 at 5:24pm: @userA mentions @Skype in one of his tweets (Tweet A)

Tweet A: “Why do they keep changing @Skype online? I used to be able to see contacts at a glance, if they were online or away and the time diff!”

15th of August 2017 at 6:49pm: @Skype replies to Tweet A

Tweet A would have a response time of 3 days.

Example 2:

12th of August 2017 at 3pm: @userA mentions @Skype in one of his tweets (Tweet A)

Tweet A: “I can't login, can you help me @Skype?”

@Skype does not directly reply to Tweet A

18th of August 2017 at 2pm: @userA again mentions @Skype in one of his tweets (Tweet B)

Tweet B: “@skype why did you change your settings?”

18th of August 2017 at 3pm: @Skype replies to Tweet B

Tweet A would have a response time of 6 days and Tweet B would have a response time of 1 hour. Although Tweet A doesn’t have a direct reply we count the reply of @Skype to Tweet B also as a reply to Tweet A.

In general you can say that if a user mentioned a page we count the first reply of the mentioned page as the reply for this mention. It doesn’t matter if the mentioned page replied to the same mention or a different mention of the page of the same user. Only if the reply is more than 7 days apart from the initial mention we would count the mention as “not responded”.

Here’s an example for this:

Example 3:

12th of August 2017 at 3pm: @userA mentions @Skype in one of his tweets (Tweet A)

Tweet A: “I can't login, can you help me @Skype?”

@Skype does not directly reply to Tweet A

22nd of August 2017 at 2pm: @userA again mentions @Skype in one of his tweets (Tweet B)

Tweet B: “@skype why did you change your settings?”

18th of August 2017 at 3pm: @Skype replies to Tweet B

Tweet A would be flagged as “not responded” because the reply of @Skype is more than 7 days apart from the created time of Tweet A. Tweet B would have a response time of 1 hour.

Did this answer your question?