Rapid Deployment Solution DI6 - Social Media

[Former Member]

Hi Gurus

I’ve been viewing the Social Media Load RDS video and am puzzled by several aspects. The core of the problem is that it steps through a number of operations while not describing what’s going on.

At about 50 seconds we add “Hobbit” as a search term.

A 1 Minute 40 we assign that term to a Twitter search query. I guess these two mean we will now search for “Hobbit” in any twitter feed.

We then run DS to load the social media data

At 5:20 we run a report to create sentiments from the loaded social media data

At 5:50 we run a report to create contact from the same data

At 7 minutes we view contact engagement and see four interests. One of these is “The Hobbit – Movie”.

• Here is my first question. The word we added was “Hobbit” and this will occur in a number of phrases. We never added “- Movie” so why is “The Hobbit – Movie” our main interest.
• Next where did “Concur”, “Hybris” and “Database” come from? Were they part of a Twitter feed or are they from some other interaction load?

A 7:15 we assign the tag “bmw” to the Interest “BMW”. Where did the list of interests come from?

At 7:45 we view the top 5 interests and now “BMW” has been added to the previous four.  We added “bmw” through the front end whereas we added “Hobbit” through the back end. Does this mean tags and other non-tag strings are set up through front-end and back-end respectively?

Thanks

Cliff

[former_member183196 · Accepted solution]

Hi Cliff,

Great questions.

question 1:  why is "the hobbit - movie" shows up.

It is the description of the interest.

Quesiton 2:  new interests such as concur, hybris and database

the rest of interest are coming from different interaction load but interests are pre-configured

Question 3:  Where did the list of interests come from?  The interests are created in the IMG (SAPGUI-->SPRO)

Questions 4: tags vs interests.

Interests are configured in the backend (SPRO).  Tags are identifed by HANA text analysis and show up on the sentiment engagement screen.  The end user(typically marketer or marketing admin) has the authorization to assigns the tags from the unstructured text to interests.  Once done, if simliar tags were found in the future, the interactions will be assigned automatically with interests.

In short, interests are configured backend and are more static.  Tags are identified on the fly from the unstructured data (HANA text analysis driven)  and can be assigned to interests.

Hope this helps

Thanks

-Bharath

[Former Member · Accepted solution]

We have built this integration into our demo system, ready to demo it to prospective clients.

The question I will surely get is whether a Marketing end user can quickly setup a new search term (or Interest as defined by SAP). Imagine the scenario where the organisation needs to quickly react to a PR disaster by listening to what is being said on the matter.

Having to access SPRO to setup a new search term is not practical enough. Isn't there a way to view and amend the list via a widget on the HTML5 UI?