Sep 2, 2009
- Sonia Riahi
OIC 2009 Challenge: ThinkSM
About ThinkSM
ThinkSM is a social media consulting firm that specializes in using social media technologies to help companies find, connect and interact with their clients online. An off shoot of ThinkSM, SWIX, is a social media metrics tracking tool currently under development, the purpose of which is to provide analytics for important social media marketing campaign metrics while also assessing levels of community engagement across several social media properties.
Challenge Overview
Between ThinkSM and SWIX, the capability to provide insight and content for social media as well as measuring social media marketing success are in place. A sector that is gaining more traction in social media however is social media listening. The market is currently populated by a small amount of companies offer software as a service that companies can use to research public opinion and sentiment on key issues.
These software options scour the web for mentions of key topics, collecting as much data as the account allows for as requires the user to sift through and grade the context and tone of the mention. Some options also associate a sentiment and authority value to the mentions, but the values must be checked to ensure and marked as acceptable or not to ‘teach’ the software was to mark as positive and negative sentiment. There are also a few companies that are employing data visualization to perform social media listening.
ThinkSM is interested in entering the social media listening sector with a software as a service (SaaS) option that employs data visualization in order to compliment the array of social media capabilities that it currently offers. The ability to meet all social media needs in house is a very appealing concept, so the ability to measure sentiment on social media properties is desired. However, there are many questions that need to be addressed.
- Are there natural language processing tools available that can improve with assessment input (will it learn from its mistakes)?
- What methods are there for associating authority to web properties?
- Why are there so few data visualization options currently used in the market?
- What type(s) of data visualization would best suit social media listening?
- How can social media properties be integrated?
- What are barriers to entering the market?
- What would be the most effective method to garner interest is this product?
Technical Details
Specific social media listening tools employ search tools to collect information based on key words to be processed by the natural language software. There is difficulty in determining methods to retrieve information from social media properties such as youtube or twitter. There are multiple options for collecting the information such as developing solutions using open source webcrawlers such as (crawler.archive.org) . -A description of a proof of concept method or compilation of methods for retrieval of pertinent information from social media properties based on reasoning and substantiated evidence would be desired.
The UI for the data visualization aspect of the social media listening tool is a crucial part of the software. A visual representation of the UI and proof of the ability to design it through examples and reasoning are desired.
Deliverables
ThinkSM is looking for a conceptual model to help grow the actual product from. A good solution should resolve the following issues.
- Is there a specific target market, or should the solution be geared towards all companies, what kind of business model would best suit the product?
- What value would this tool add to a social media marketing campaign?
- What type of campaign would suit this product?
- Is there a good method for leveraging the current client base of ThinkSM and SWIX?
- From the technical side, what methods can be employed for content searching across all social media platforms to provide information for the natural language processing software?
- What kind of UI would suit the ideal prospective user best?
- What would be the best types of listening metrics to track?
A high level proof of concept with substantiation is desired.