IBM CXO Tweetchat, Segmentation and Big Data, with Gregory Piatetsky-Shapiro

Here is a summary of IBM #CXO tweetchat on Customer Segmentation in the age of Big Data – are demographic, psychographic, lifestyle, and other buckets made obsolete by personalized predictions and recommendations? The tweetchat was so intense, #cxo briefly became the Top Twitter Trend.

By Gregory Piatetsky, Aug 5, 2013.

IBM Big Data #CXO

July 29 #CXO Tweetchat, hosted by IBM and featuring me as a guest expert was quite an experience. We had prepared 9 questions which focused on the issue –

What customer segmentation is still meaningful in the age of Big Data? Are demographic, psychographic, lifestyle, and other breakdowns made obsolete in the face of Big Data-powered personalized predictions and recommendations?

#cxo was the top Twitter Trend on July 29, 2013The experience was very “firehosish”, with over 900 tweets from about 300 participants. The intensity of tweets was such that our hashtag #cxo actually made the Top 10 Twitter trends, and at the end of the Tweetchat #cxo even briefly became n. 1 Twitter trend, ahead of iPhone 5C, #Bulger, and FBI !

Here are some of the more popular tweets:

Q3: What are key steps to using analytics to develop needs based customer segments?

@kdnuggets: A3: Capture recent and relevant customer behavior; augment with external data; use advanced analytics to develop models #cxo

@kdnuggets: A3: Develop the right measures to differentiate segments and operationalize them in a clear, cost-effective way. #cxo

@IBMbigdata: A3: #Segmentation starts with objective refinement and collection of enterprise-wide biz knowledge & data discovery #cxo

@kdnuggets: A3: HBR blog argues for segmention on “jobs”: customer desires. Nice idea, but can it be done? #cxo

@zacharyjeans: A3 Recently been reading @FTC guidlines (I know, I’m a dork), and see a basic theme: Customer must consent to data collection and use. #cxo

Q6: Privacy/legal/ethical issues – can segments be based on lifestyle, age, pregnancy, race? Known vs Inferred ?

@kdnuggets: A6: When Target predictive analytics identified a young woman pregnancy before her father knew, it was creepy #cxo

@kdnuggets: A6: However, unless prohibited, companies will use analytics on all data, because they work and increase ROI #cxo

@kdnuggets: .@AmyL_Bishop @cloudspark demographics < psychographics, but psychographics < actual user behavior #cxo

Q8: Does the trend toward customer experience personalization (customer as a “segment” of one) spell the death knell of traditional customer segmentation (i.e., target customers as groups)?

@kdnuggets: A8: The trend is towards a segment of 1, and even less than one – a customer may have many personas #cxo

@marmarlade: #cxo @IBMbigdata A8 “From one conversation with a million, to a million conversations with one.”

See a Storify of the July 29 Tweetchat at