Reading the wrong tea-leaves: Why unstructured sentiment analysis is so important

 

blog  - tea leaves.jpgI used to manage a marketing department of 13 people. One of the team members was great at data analysis. She would download all the sales orders for her products and spend hours slicing and dicing the data, trying to gain insight into what was happening.

And yet we had persistent problems in certain areas that she just couldn't figure out.

The mystery was eventually solved when I went to a UK user group meeting for her products. During one of the breaks, a customer asked me if I'd been to our competitor's website. He explained that in the communities on this website, customers were trashing our products and suggesting alternatives.

We immediately went to the communities in question, read through the posts there and fixed the problems (they were relatively simply to fix, having to do with upgrade pricing in certain situations).

 

Merging the “systems of record” with the “systems of engagement”

Looking at the structured data— the "Systems of Record" as Geoffrey Moore <http://www.aiim.org/futurehistory> calls it—doesn't give us the whole picture.

It's when we join the Systems of Record with the "Systems of Engagement"—the unstructured data that you get on communities like Facebook, Tumblr, Google+ and Twitter—that we get the whole story.

This combined view gives us information about "what's trending" and what is hot regarding a product or solution. It tells us what people think of a product versus the competition’s product. It tells us what substitution solutions are being considered. What can blindside project managers is not direct competition, but people going for something completely different. For example, rather than buying one type of laptop or another, consumers may be buying tablets instead. It can also determine major support issues that are hurting a business’ sales.

I heard of another example just the other day. HP consultants were working with a car manufacturer. This car manufacturer believed that its cars were as good as the competition’s, but they just weren't getting the same level of sales. They matched the competitor on price and advertising spending, but to no avail.

 

Determining how and why data trends


We set up an analysis of the social media "chatter" about their cars and found that they had a solid reputation. So solid that "old people" loved them—they recommended them to their friends to such an extent that one of their models now had a reputation as "the car to buy if you are old.” Great. Not. The "younger people" saw this trend and thought, "No way am I buying an old person's car.” Sadly, there are more "young people" buying cars than "old people,” hence the lackluster sales.

This effect showed up quickly in analysis of the "systems of engagement" around the company’s cars, yet there was nothing in the "system of record" that could ever have provided this insight.

 

Embracing this new relationship

The combination of system of record and system of engagement is very powerful. Analysis of this combination tells us all sorts of things we just can't get from looking at the system of record. System of engagement typically leads system of record. For example, the HR department will see dissatisfaction from employees in social media before it starts to see the "system of record" showing that top performers are starting to leave for places that social media tells them are better to work for.

I've been a product manager with HP for 10 years. I believe that analysis of system of engagement and system of record together will be one of the key tools for product managers in the years to come.

We have just released our "Marketing 20/20" vision for the marketing department of 2020. It is part of HP's crowd-sourced vision for the future, Enterprise 20/20. Please visit the Enterprise 20/20 website to share your own point of view.

 

Author : Mike Shaw

 

Comments
tennis divider | ‎11-21-2012 02:24 PM

The net has become a very useful tool not only in marketing but also in decision making and analysis.  Its always good to know the bigger picture.  Look at the pros and the cons to be able to make a better decision.  It was funny though, to hear such a story, that younger people were not buyingthere cars even if it was a great product because older people like them. Lesson learned : You will never know what people think unless you ask them directly. 

mikeshaw747 | ‎12-10-2012 12:49 AM

I have to agree. While I think that the link between unstructured and structured data will help, there is no substitute for actually talking to people. I think that marketing people are deluding themselves if they believe that in the future they will be able to sit in their offices and never "get out there".

 

I believe it was Mikael Gorbachev who said, "to see for yourself is to hear a thousand times". 

Leave a Comment

We encourage you to share your comments on this post. Comments are moderated and will be reviewed
and posted as promptly as possible during regular business hours

To ensure your comment is published, be sure to follow the Community Guidelines.

Be sure to enter a unique name. You can't reuse a name that's already in use.
Be sure to enter a unique email address. You can't reuse an email address that's already in use.
Type the characters you see in the picture above.Type the words you hear.
Search
Showing results for 
Search instead for 
Do you mean 
About the Author
Mike has been with HP for 30 years. Half of that time was in R&D, mainly as an architect. The other 15 years has been spent in product manag...
Featured


Follow Us
The opinions expressed above are the personal opinions of the authors, not of HP. By using this site, you accept the Terms of Use and Rules of Participation.