This webinar is hosted and promoted by Crimson Hexagon: Social Media Monitoring and Analysis

The Crimson Hexagon platform aim is to help you to achieve cross-functional social intelligence for business.

Zach Hofer-Shall (Forrester Research) presented the nuts and bolts of current Social Intelligence

Social media includes everything: marketing, news; commentary; interaction of mainstream media to social media; peer interaction. The underlying factor is that information is shared (social). The difference between mainstream media and social media is now irrelevant. The underlying reason is the inherent social nature of the internet. Social media analysis differs from traditional market analysis by a squishiness. Squishy in the sense that it has no benchmarks, combines qualitative and quantitative analysis, and is constantly changing.

7 levels of social media users: creators, conversationalists, critics, collectors, joiners, spectators, inactives. The future will be made up of everyone at least at spectator level.

Social Intelligence def: Technology, services and data to inform customer development. Using the data is spreading into all arenas of companies.

Social Intel means:  Listen; Learn; Act to achieve cross-functional analysis. Only Listening and Learning can be outsourced.

Functions of social intelligence: Tracking, support, listening, process with textin mining, human analysis, and natural language processing. The final step of insight delivery is organized, interpretive.  Currently, clients are varied in their satisfaction levels with social media analysis. The automated sentiment analysis is lowest on satisfaction level, and human interpretation and service scores the highest. There is a gap between human and machine interpretation of data.  

The challenges of social media data are: Not all social media data is valuable or relevant. Picking out the valuable data from the messy, huge, constantly growing pile:. Social media is fragmented due to  many venues (Twitter, blogs, Fb); filled with spam/porn/garbage; global with multiple languages (including nontext); heavily sarcastic and emotional. Ultimately it’s unstructured text (no controlled SurveyMonkey tools here).

Listening means: Using platforms to organize in collections

Listening consists of: triage by day-to-day data management (can outsource); Data analyst interprets the data to identify risk and opportunity (can outsource); and Process Lead develops goals and conducts action while managing the team (in-house).

Current Platforms used to listen: social dashboards(web based ans self sufficient), multichannel analytics providers; litening service partners that combine tech and proessional services (fastest growing group).

The Social Intel process for your company: 1st Step: Determine what your capabilities are to out-source or focus on in-house listening. 2nd Step: find tech assistance. 3rd step: determine your leader/coordinator/manager/delegator

Melyssa Plunkett-Gomez (VP)

Melyssa presents the Crimson Hexagon tools applied to two specific cases. She finds that the application of social intel to business strategies that eventually achieves ROI is a challenges. She points out that the traditional tools don’t work. Keyword buzz monitoring, sentiment monitoring use of keywords and the NLP semantic search don’t work in social media because the traditional tools don’t understand context. No comupter can replace human judgment because the predefined rules cannot keep up with the dynamic contextual environment.  Human judgment controls context nuance by training the algorithm. The computer collects the massive amount of data. The computer role is to address scale, accuracy, and repeatability. The human teaches the algorithm by identification of the theme of the conversation, manually codes random samples for placement into buckets.  

She argues that sentiment is only a small part of the actual picture. This platform aims to uncover the drivers of positive and negative sentiment witilizing Gary Kings algorithm. Her two case studies highlight deep dive visualization and color tools that categorize themes,,and show levels of sentiment in relation to percentage of comments. 

Overall no one can guarantee to have a finger on everything. The database grows exponentially every day. Achieving comprehensive content means the provider must be constantly updating the sources, and find the difference between quantity and quality of data.

Other events are archived and available at: