Voxeo Talks

Social media analysis is a huge growth area right now, as brands start to graduate from the days where merely having a presence was enough, to actively seeking ways to measure and monitor their social media activity. Consumers, similarly, may want an easy way to find and interpret social media data.

I met Lance R. Vick, founder and lead engineer of Tawlk, at the Voxeo Venture Lounge event, and he took me through a demo of his product. I was also fortunate enough to catch his presentation at BarCamp Orlando this month, and I’ve been exploring some applications of the product. Tawlk’s mission statement – “Search and Analyze the Social Web” gets to the core of social’s current measurement dilemma, and it has several features that are unique when compared to other sites in the sentiment analysis space, like SocialMention – Tawlk continuously updates data in real-time and scores it with its unique algorithm, allows multimedia content like photos and videos to be viewed from within its results, and, amazingly, has made its sentiment analysis and data collection techniques open source. The entire platform has been built by a small core team consisting of one full-time and two part-time members.

New to sentiment analysis? Here are a few examples of how to use a tool like this:

Example 1: Simple Brand Monitoring

A business evaluating its social media presence and reputation will want to know a few critical things:

• What people are saying about you, so you can respond if necessary
• Where they’re saying it, so you can properly allocate your time among channels
• How positive/negative the sentiment about your brand is running

Tawlk can address all of these items. Input your brand name and you’ll see a string of mentions scrolling across your screen in real-time (which could be kept open on an extra monitor by a dedicated social media employee, or displayed in a meeting room or other public area to keep all company employees aware of relevant brand chatter).

For example, some recent multimedia mentions of Voxeo:

Each item is labeled by source – all major social sites are included, such as WordPress, Digg, Flickr, YouTube, Picasa, Reddit, Twitter, Imgur, Identica and more – and can be filtered by type, so if you are only interested in posts, links, videos or images referencing your brand you can screen the content. In the left sidebar, the site additionally measures sentiment and keeps a running tally of the ratio of positive to negative sentiment about your brand. If you determine that you receive a significant number of questions requiring the same response, or get many common queries for customer service, this can suggest certain courses of action like implementing autoresponders that provide assistance and links to additional information. (For more information on how Voxeo supports this kind of automation on Twitter, check out our whitepaper.)

Example 2: Lead Generation

There are many people expressing a need – and looking for help – on social media. For example, suppose you offer web design, search-engine optimization or domain name services. A simple search for posts containing the words “new website” on Tawlk yielded several blurbs like the following:

Using additional modifiers, like the words “need” and “recommend” – e.g., “need new website”, “need a web designer”, “recommend a web designer” - could be used to find cases such as blogs where a quality, thoughtful comment might result in a new customer.

Example 3: Stock Market Ideas

In March, a study came out from the University of California suggesting Twitter data could ‘predict’ stock prices by scanning the volume of tweets and what they linked to. The chart on the right shows a comparison between all mentions of the word Apple and Apple Computer’s stock ticker AAPL. Sentiment is similar, while the stock ticker is considerably less popular. However, the amplification is higher for the stock ticker, meaning those posts are more often re-shared. Trending this data over time and in comparison to the stock price would be an interesting exercise. While not designed to provide investment advice, a tool like Tawlk could be used to examine a number of relationships to determine if they had any correlation between the stock’s price and their magnitude, for example:

• Discrepancies in sentiment between the stock ticker (traders) and the company name (consumers)
• Sudden increases in popularity for the stock ticker
• Slow or sudden increases in popularity for the company name
• Sentiment, popularity or amplification of the company name versus sentiment, popularity or amplification of competitors

Swing traders might be interested in analyzing rapid changes in popularity of stock tickers when combined with sentiment, to see if these correlate to price changes, to time future trades accordingly; everyday investors might be more interested in changes in popularity or sentiment for the company names of stocks they hold.

In contrast to enteprise-level players like Voxeo partner RightNow, Tawlk is currently in public beta and seeking funding to build out a range of new free features, relevancy improvements and subscription upgrade options priced at the lower end of the market.

Try it out: http://www.tawlk.com/

Continuing the series of questions coming out of our webinar, “Best Practices in Deploying Communication Applications: Cloud vs On-Premises vs Hybrid“, let’s ask the question:

#6 What kind of real-time analytics and linkage to business intelligence solutions are possible?

When you are evaluating cloud solutions, it’s not enough just to have a cloud platform run your application, you also want to understand how people used the application. You want to know how often people interacted with the app… what channels they used (voice, SMS, IM, web, Twitter, etc.)… how successful they were… etc., etc.

You want to be able to get reports that show you the usage – and dive into the business processes. Where in the call flow of the application did people hang up? Where did they start to press “0″ to get to an operator? Where were you able to successfully upsell them (if that is a goal) and where were you not? What can you do to make your application more successful?

You need to be able to get this data in real-time, too. If you have launched an outbound campaign calling people, you want to be able to see the results in as close to real-time as possible. If you can spot trends where the app isn’t working so well, you can try out fixes to see if you can change the results while the campaign is underway.

You also want to be able to tie these results into the business intelligence systems and “dashboards” that you already have within your business.

Can your cloud do that? If so, go nuts! If not, either find a cloud that can or look at a premise solution that can deliver.

As I mentioned last week, I was part of a webinar panel on Wednesday looking at questions like:

Would you like to learn how to use analytics and business intelligence to create a better experience for your customers? As you move into multi-channel communication across voice, SMS, IM and more, would you like to have your analytics work across all those channels?

While the event itself is now over, you can still register and view the archive of the webinar. I’ve also made the slides available in our SlideShare account… although since I’ve done them in the minimalist style I prefer, your best bet may be to listen to the recording:

You are also welcome to go to view our analytics data sheet or

to take our analytics for a spin!

Would you like to learn how to use analytics and business intelligence to create a better experience for your customers? As you move into multi-channel communication across voice, SMS, IM and more, would you like to have your analytics work across all those channels?

You can learn about all of this and much more on a free webinar sponsored by DestinationCRM on:

WEDNESDAY, JANUARY 26th, 2011
2:00pm US Eastern, 11:00am US Pacific

You can register for free and either join the call live – or listen to the archive later if you can’t attend.

I (Dan York) will be on the panel along with folks from Angel and Coveo. We’ll be covering topics including:

• How to achieve a unified analytics view across multiple self-service channels, including voice, SMS, IM, and social networking
• How to uncover problem areas in your voice application using BI Analytics
• How to leverage Enterprise Search 2.0 solutions in all Customer Support systems to power real-time Voice of the Customer analytics and enhance decision-making

It should be an interesting and educational event.

By the way, a key product we have that provides this analytics and BI capability is our VoiceObjects application lifecycle suite. You can download a free version of VoiceObjects to try it out yourself.

We also have a whitepaper out about VoiceObjects Analyzer Report Samples with this abstract:

VoiceObjects Analyzer is a service analysis environment enabling instant reporting on key business, operational, and application design questions. In addition to enabling users to create custom reports, VoiceObjects Analyzer delivers over 50 out-of-the-box reports providing valuable insight into the usage and acceptance of application services by leveraging the power of common business intelligence frameworks such as SAP BusinessObjects, IBM Cognos, and MicroStrategy. This whitepaper provides 10 report examples that highlight the analytical capabilities of VoiceObjects Analyzer.

Please feel free to contact us if you have questions about how we can help with analytics and integration to leading business intelligence systems. And please do join the webinar next week to learn more!