A leading bank discovered it was losing customers to an aggressive competitor. The method was simple: whenever the competitor saw a customer complaining about the leading bank on Twitter, they invited him or her to change banks.

The tactic worked only too well to be ignored.

The morale of the story: if you want to make the most of the web, you don’t just monitor what people say about you online. You need to take action based on what you learn from those discussions.

Furthermore, you are not limited to monitoring what people say about your company. Everything that you can monitor about your own company, you can also monitor about your competitors.

Therefore, it’s no wonder that most organizations, large and small, are now monitoring social media, and in fact not just social media but the web as a whole. There are literally hundreds of tools and services out there to offer you social media monitoring and analysis.

Accuracy of social media monitoring determines your success

In order for monitoring to be useful, though, the tools that we use must come with advanced analytics.

Imagine the following scenario. You want to measure brand sentiment on the web and categorize brand mentions into positive, negative or neutral. Easy, right?

You go ahead and try one of the free social media monitoring tools, for example  socialmention.com, and stumble upon the first problem – noise. Noise is for example if you’re measuring online mentions for Santander (a bank), but your monitoring tool includes mentions of the Spanish city of Santander in your results.

If your monitoring tool is not able to filter out the irrelevant content, imagine what happens to the reliability of your sentiment analysis. Garbage in, garbage out, as the old saying goes. So you’ll need a good filter that is able to stop the irrelevant stuff from contaminating your sentiment analysis – and your own professional reputation.

Check out the amount of noise that your monitoring tool will capture if you work for – or compete with – for example If (insurance), SAS (airline, software, military), or 3 (telco operator).

Enter text analytics and natural language processing

Having filtered out the noise, your tool needs to figure out which statements are positive, negative or neutral about your organization.

Are the words “high” or “low” positive or negative? There’s no way to know if you don’t know the context of the word. This is why your analytics tool needs some heavy duty text analytics as well as what they call natural language processing. You will want your analytics tool to be able to capture and understand not only individual words, but terms and phrases consisting of several words.

Your tool also needs to understand the structure of your language, or many languages, if you are an international company.

For example, if our bank’s customer writes about “a high interest rate”, we don’t know if he or she is happy or unhappy. Your analytics tool needs to figure out if the person is talking about his home loan or his bank account.

These are just some of the reasons why you’ll want your (social) media monitoring and analytics tool to come from someone who is strong in advanced analytics.

Accuracy of measurement is key, whether it’s about brand sentiment or other social media metrics. Without good accuracy, you can’t take any meaningful business action based on your monitoring.

Our bank under attack revisited

So what happened to our friends in the bank that was losing its customers to an aggressive competitor?

They started using a social media analytics tool that captured and analyzed customer complaints in real time. As a result, they are now able to contact dissatisfied customers immediately and prevent the competitor from stealing their customers.

Posted previously in BI Blogg, Norway’s leading business intelligence website http://biblogg.no/2012/06/15/is-your-competition-eating-your-lunch-in-social-media/