IBM #CXO Tweetchat: Sentiment Analysis and the Customer Experience

Good content is both timely and lasting, correct? I hope that my responses, in a May 16, 2013 IBM #CXO tweetchat, were both. Decide for yourself; I’m posting below my conversation-starter responses to eight questions. I prepared them in advance — tweetchat lesson #1 — in order to keep the pace up, although the ensuing exchange was of course unscripted. For instance, view the full set of tweets prompted by the question, “How can organizations mine unsolicited social feedback to drive customer experience strategy?” by following this link.

#CXO stands for Customer Experience Optimization, also readable of course as Chief {Executive | Marketing | Information} Officer. Thank you to the IBM Big Data folks for inviting me to participate. Here’s what they asked, and the start of what I had to say, in tweet-size chunks of 140 characters of fewer, on the topic of Sentiment Analysis and the Customer Experience:

1.    Is it best to categorize customers by demographics, brand, product type or sentiment? Why?

A1: Q: How to categorize customers? A: By categories useful in better reaching & serving them. [1/4]

A1: You have to know what product types a given customer (or demographic) is interested in, or you’re wasting effort & spamming. [2/4]

A1: And you’ll want to understand what brands (i.e., image, positioning, channels) people respond to. [3/4]

A1: And then you’ll use sentiment analysis, along with social network analysis, behavior models & profiles, to craft brand initiatives. [4/4]

2.    How can organizations mine unsolicited social feedback to drive customer experience strategy?

A2: The process of mining unsolicited feedback starts w/listening, but you have to filter, prioritize & make sense of what you hear. [1/3]

A2:  Hardest part is linking feedback to business outcomes, for predictive purposes, within your customer-experience strategy. [2/2]

3.    Is the customer voice uncovered from social sentiment more accurate than say surveys? Why or why not?

A3: There are big differences between social & surveys. With surveys you know who respondents are. Also you’re framing the questions. [1/3]

A3: By contrast with social, your customers (& prospects & the market) are expressing themselves more freely and you get to eavesdrop. [2/3]

A3: So my answer is, you need Voice of the Customer sentiment, from both social & surveys, to get a complete, accurate picture. [3/3]

4.    How do you effectively map and connect sentiment analysis with other customer data?

A4: Whether you’re mining sentiment at a topic, message, or feature/aspect level, obviously you need to be able to associate… [1/2]

A4: … to attribute sentiment by customer segment or to individual customers, by associating via a common customer ID. [2/2]

5.    How can businesses use sentiment analysis to impact customer interactions?

A5: #1 train machines to detect sentiment & #2 figure out how to react in different situations. There are apps that do it!

A5: Then track, measure, improve, and extend.

6.    What are primary hurdles in sentiment analysis and how can companies overcome?

A6: The primary sentiment analysis hurdles I see involve finding the right tools, aligned with biz needs, that provide usable insights.

A6: What do you all think? What sentiment analysis adoption/usability/reliability hurdles have you experienced, or do you see?

7.    What are best practices for analyzing large amounts of data to understand customer sentiment?

A7: We’re only now discovering best sentiment analysis best practices. I say, work backwards — [1/3]

A7: What are your biz goals? What data & analyses will produce the insights that guide you in meeting goals? [2/3]

A7: The key, you have to find the right data, via filtering, and

8.    Can businesses really monetize sentiment analysis? Why or why not?

A8: Monetize? Only if you’re a solution provider or data provider? Otherwise ROI is indirect… [1/3]

A8: Boost customer satisfaction and loyalty; create promoters & advocates; increase product quality & lower costs… [2/3]

A8: These are our #CXO sentiment analysis goals. [3/3]

Leave a Reply