Maulik Shah

How to Do Sentiment Analysis Using Surveys in Sugar

Blog Post created by Maulik Shah on Jun 12, 2017

Sentiment Analysis (also known as opinion mining) is quite a popular term in the current social media age. The explosion of user generated content on social media platforms is a goldmine for brands to understand the sentiment of their customers and followers.
But is sentiment analysis limited to social media monitoring?
To answer the above question we first need to understand what is sentiment analysis. It is all about determining the emotional tone and attitude of customers or users behind their opinions/view about a particular topic or product or brand. The reason why sentiment analysis is so popular in social media monitoring is because social networks are public forums and it is necessary to understand the overall sentiment of people towards your brand to be able to control or even change the sentiment from negative to positive.
But social media monitoring and sentiment analysis using monitoring tools is, by no means, the only way to do sentiment analysis. You can achieve the same goal, perhaps in a much better way, by conducting sentiment analysis surveys through SugarCRM. In fact, conducting a survey is perhaps the best way to understand the sentiments and attitudes of your customers by directly asking them open ended or opinion based questions, rather than relying on machine learning to mine sentiments from social comments. Machine learning can only interpret words but not the context in which they are spoken.
So how do you conduct sentiment analysis surveys in SugarCRM?
Here is a simple three step process to do this.
Step 1: Figure out a way to conduct surveys in SugarCRM.


Conducting survey within your SugarCRM system has many advantages over general online surveys. To do this you can either use a general web based survey tool and integrate it with SugarCRM using a compatible integration solution or use a survey tool like Survey Rocket, which is built specifically for SugarCRM and doesn’t need any third party integration tool.
Step 2: Prepare a sentiment oriented questionnaire.
If you want to understand your customers’ sentiment towards your products and services you can straight away ask them about whether they are satisfied with your products or not. But such a question will not give you an idea of the degree of satisfaction, and hence, will not help you in getting a sentiment score.
So your questionnaire for sentiment analysis should be different from any satisfaction survey. For example, you could ask a question like:
How satisfied are you with our after sales support?

And provide answer options like:

  • Disappointed
  • Somewhat Satisfied
  • Satisfied
  • Very Satisfied
  • Delighted

 

 

You can assign a numerical value to each option for getting a sentiment score - positive integer for positive emotions and negative integer for negative emotions. You can also ask customer to rate you on a numerical scale, but keeping words as options gets better results as words can be easily associated with emotions and hence you will get more accurate responses.

Step 3: Send out the surveys and analyze the results.
Send out the survey to the right set of people whose attitude you want to understand. And once you have got all the survey results you can add up all the scores from individual questions in your survey and the consolidated score for each customer.
As you would have realized, in the above process, the key is to frame the right set of questions. The more variety you have in your questions and the associated set of answers, the bigger range of emotions you will be able to capture in your survey.

Outcomes