Great bagel, but Lousy Coffee...

Leveraging Actionable Insight from Social Conversations

Jane pulled out of the drive-through after picking up her breakfast on the run. After sipping her coffee, she almost spit it out, then not even a minute later, she tweeted about it, “Seriously @Coffee, great bagel, but lousy coffee :( I shoulda gone elsewhere”.

Customers say the darndest things on social media! But, do you get what they say? Do you understand what they mean? When there are thousands of things being said about your brand, product, and services on multiple social channels, how do you handle them? To make things more challenging, the same customers expect you to engage with them in real-time, with relevant responses every time. As a customer engagement leader responsible for social channels, how do you hurdle this challenge? Manual methods to decipher every tweet or post are impractical. Separating noise from relevant conversations that impact your brand could make or break your engagement team. This is why you need advanced technologies such as text analytics and sentiment analysis to power your engagement platform. Let’s look at how text analytics, a foundational technology, helps you succeed in your social engagement mission.

Text data, unlike ‘structured data’ present in traditional databases, is highly nuanced and complex to analyze. Social conversations add even more complexity. Text Analytics uses a a mix of natural language processing (NLP), information extraction, data mining, and artificial intelligence. Among many use cases for it, customer engagement-related use cases, especially in sales, marketing, customer service and online experience, have emerged as most beneficial to large enterprises. The right analytics technology can transform your customer engagement strategy for social channels through a unified four-step process: Listening, Processing, Analyzing, and Engaging.

  • Listening: Real-time listening to relevant conversations through pre-built connectors to social channels such as Twitter and Facebook, and other social sources is the first step. Built-in smart filtering technology can eliminate spam, duplicates, and other irrelevant content. Significant noise can be eliminated, reducing costs and wasted time.
     
  • Processing: Linguistic processing using NLP techniques prepare the text for deeper extraction of information. Key facts such as topics, people, products, places, and companies can be extracted to add to the metadata. For example, take Jane’s tweet. Text analytics can extract the brands mentioned (@coffee is made up in this case). If locations, people or any named entities were mentioned, it would extract them too. Sentiment scoring determines the tonality and emotion expressed on specific features. In the example, bagel as a feature has positive tonality, and coffee as a feature has negative tonality. The technology associates sentiment words “great” and “lousy” to the right feature words. This type of granular scoring at the feature level delivers superior insight compared to determining positive, negative, or neutral sentiment at the overall tweet level. When combined with metadata such as location, the engagement team could then take action on what could be wrong at that specific retail location. Also, since the customer said they’ll switch brands, you’ll want to prevent customer attrition and protect your brand image. Auto-categorization to taxonomy relevant for your business enables even more in-depth analysis, previously not be possible with traditional tools. New generation text analytics technology achieves all this by leveraging machine learning models that learn iteratively over time from data and user inputs. 
     
  • Analyzing: Advanced processing achieved in the previous step renders the output for complex queries. Dashboards and reports with interactive charts aid in discovering hidden insights so the enterprise can act. Auto-categorization can help in routing social conversations to the right team increasing response speed.
     
  • Engaging: A single unified platform that seamlessly combines the previous steps with enterprise engagement capabilities delivers the most value. By directly linking customer insights with your goals, messaging, campaigns, and content on a single platform, engagement teams will be better equipped to meet and exceed your customers’ rising expectations.


Speedy, direct and meaningful engagement is crucial to restoring customer loyalty. The right step is to immediately engage with Jane on Twitter offering a quick response, sincere apology, complementary product offer, and commitment to fix the issue. Insight-based social engagement is a key differentiator in any customer relationship strategy.

You can join us at PegaWORLD, June 8 -11 in Washington, D.C., to see for yourself how Pega can help with your social engagement initiatives. Check out CPM Booth #2040 while you’re there.