It is strange that some brands spend upwards of several hundred thousand dollars on market research and customer feedback, but don’t listen to the average customer on social media who is writing about his tastes, preferences and aversions day in and day out.
The average response rate for a customer feedback form is less than 5% (across industries), while over 15% of the customers who purchased online share their experience and opinion online.
Your customers and prospects are constantly exchanging feedback on social media. They are talking about their likes, dislikes, aspirations, and hang-ups. They ask their friends for opinions on social media when they are considering your product. If they are unhappy with your product, they post a negative review and announce to the entire world. If they are delighted, they refer you to their friends with a glowing review! There is no doubt that social media plays an important part in truly understanding your customers.
However, here is the catch – social platforms are designed for humans. They encourage informal conversations which may not always be grammatically correct. Human communication is also highly context dependent. Understanding a simple phrase like, “He saw the boy with the telescope,” is tricky. Did he see the boy through the telescope or carrying the telescope?
Coupled with this is our inability to keep up with volumes of data from social media. Reading through thousands of pages of comments is not only inefficient but also ineffective.
This is where text analytics comes into play. Text analytics strives to derive meaning from the written word. Armed with the capability to handle tons of data, it is one of the most useful tools in a brand’s toolkit.
Major advantages of using text analytics
- A real-time, objective and consistent analysis: Unlike people, computers do not suffer from race, gender or cultural biases. They remain faithful to the rules, and those rules can be refined as and when required. They are always consistent, unlike people – who make mistakes from time to time.
- A more accurate measurement of reach: Estimating how many people have been exposed to your brand or just have read about your brand, has simply become impossible. Customers sometimes use short forms for the brand names, and sometimes pronouns such as “it”, “they”, etc. Humans excel at deciphering this kind of language, and once this intelligence is transferred to machines, calculating reach becomes much more accurate.
- Spotting trouble: Brands can avert major catastrophes if they can spot issues well ahead. If they can proactively resolve issues, it saves a lot of time, effort and PR firefighting.
- Trend analysis and predictive modeling: The more you track public opinion, over time the more likely you are to recognize trends. Attitudes about your brand or products may change seasonally. Response to your offering may be affected by many factors, including weather, sports seasons, holidays or fiscal calendar.
Monitoring sentiment over time not only provides a map of the popularity of your brand over time but can offer clues as to what to expect for the next interval. Brands can now plot promotional activities against a historic map of ongoing consumer response. This lets brands logically create compelling social media campaigns and better understand the result.
Categorisation using text analytics
Text analytics works primarily on keyword strategies. From the entire chunk of text data, the software gleans specific keywords and then figures out if the customer is a happy one or an unhappy one.
Overlaying NLP (natural language processing) and machine learning on top of text analytics takes the analysis to the next level. The software can now figure out what stage of the journey the customer is in – Is he considering to buy the product? Is he comparing it with a competitor’s product? Is he facing some difficulty in reaching out to the company? Are there some features he would have preferred? Is he displeased with the call center? Is he going to leave a bad review everywhere?
This insight is extremely valuable to the brand. It helps the frontline staff tailor its communication to the customer, and address his concerns turning a situation with a potential bad review into another with a satisfied customer.