How we shifted social media listening into overdrive for an automaker

Thursday, January 10th, 2013

If you’ve tried social media listening research in the past, you’ve been limited to basic types of analyses such as frequencies and mean sentiment scores. This case study demonstrates how advanced statistical techniques can be applied to social media data to learn more about perceptions of vehicle brands and features. We will focus on regression and factor analysis to better understand how consumers group brands of cars and features of cars. We will also show how these groups differ based on the gender of the social media commenter. You need not have advanced understanding of factor analysis to learn from this webinar, just advanced curiosity about analyzing social media data.

You can see the slides here, or even better watch the full 30 minute presentation here.

Read what our viewers said!

@lovestatsThanks for the great webinar. What’s the biggest obstacle to using SM research facing the industry right now? #RNwebinars #mrx

— Z. W. Bennett (@ZWBennett) January 9, 2013

@lovestats Nice webinar. Your initial stepwise slide with 10 jillion variables made my hair fall out, but I recovered. #mrx

— Mike Donatello (@mikedonatello) January 9, 2013

Great information @lovestats!! I enjoyed learning more about Conversition and how it works with social media #RNwebinars #MRX

— Grace Hughston (@GraceHughston) January 9, 2013

An exceptionally clear & understandable description of the several analysis. Kudos Annie @lovestats #RNwebinars#mr

— Direct Resource (@direct_resource) January 9, 2013

Pantene’s Big Change #MRX

Monday, February 13th, 2012

By Ben Johnson

Is social media research easy to do? Must you be a category expert to get significant insights out of it? Well, I know nothing about hair care products so what better way to demonstrate the power of social media research.

Let’s look at a few leading brands to disentangle brand specific trends from industry wide trends.  We have gathered tens of thousands of verbatims from all over the internet and cleaned, scored and categorized opinions about Head & Shoulders, Herbal Essences, and Pantene. The data is provided using a 5 point scale going back to September 2010.

Two questions spring to mind.  First, why was the sentiment for Pantene so much lower than its competitors  in the fall of 2010 and second, what caused the sentiment to rise so quickly through to May of 2011?

Perhaps because of the low sentiment, in the spring of 2010 Pantene completed a multi-million dollar investment in market research and R&D, and rolled out a large brand shift.  They reduced their product line by about a third, removed their 2-in-1 lines, brought out new formulas and redesigned their branding around 4 main types of hair. Despite their efforts, Pantene seemed unable to stop their customer base from being eroded by competition from new in-store and value brands as consumers sought to tighten their belts in the midst of economic downturn.  The effort was not the success the company had envisioned.

So why the rebound?  Let’s understand the composition of the shift.

We should note that it is not so much the case that Pantene’s initiative upset customers as it failed to resonate with them.  This can be seen by a change in the distribution of sentiment from one predominantly blasé to a more positive composition.  So what drove this change?  The factors which drive an individual’s decisions are too numerous to provide a definitive all-encompassing answer.  Unemployment fell, weather shifted, and myriad other statistics changed, but that does not mean our analysis is without traction.  If one takes the time to look, one can find useful and salient points.  Such as this:

Here we see the sentiment for coupons in conjunction with Pantene, Herbal Essences and Head & Shoulders.  The graph speaks for itself but, if you are still not convinced, the correlation between coupon sentiment for Pantene and overall sentiment is 0.89 .  In all their concerns over new competition and their attempts to fine-tune their product, they didn’t see the forest for the trees.  The people at Pantene were too concerned with what a person wanted in a shampoo and not enough with what consumers were looking for in general in these tough times, a good deal.  But hey, what do I know?  I’m a social media research expert, not a hair care expert.

Social Media Sentiment: H8ers and

Monday, August 15th, 2011

H8ers! <3ers! Isn’t social media just full of people who have radical opinions? It’s been a while since we first shared information about the distribution of opinions/sentiment in social media so we thought it was about time we conducted our little experiment again.

For six different sets of data, we gathered hundreds of thousands of sentiment scores and prepared frequency distributions of the results. As you can see below, some brands have more positive (A, B) sentiment while others have more negative (C, E, F) sentiment. You can also see that some brands have more flat (E) or peeked (C) distributions, or longer tails (A, B). No matter which particular feature of a brand’s chart interests you, it is clear that all of the distributions are reasonably normal, they are generally bell shaped.

So is social media full of haters and lovers? Most definitely not. Most social media data consists of lots of moderate like and dislike, plus a healthy representation of haters and lovers. Now the only puzzle is determining which of these charts reflects the sentiment of 1) autism, 2) Lady Gaga, 3) Obama, 4) Steve Jobs, 5) Toyota, and 6) Walmart.

.
.
Related links
Are Only Crazy People Commenting About Brands in Social Media?
Article in the Vue: Words I’ll Live to Regret
Cell + Survey + SMR: A Social Media Mashup #MRIA2011 #MRA_AC #MRX
There is no question but the research validity question

.

Upcoming ESOMAR 3D presentation: Tell me what you want, what you really, really want

Tuesday, July 26th, 2011

3D DIGITAL DIMENSIONS 2011
(ONLINE + SOCIAL MEDIA + MOBILE) RESEARCH
MIAMI / 26 – 28 OCTOBER

.
TELL ME WHAT YOU WANT, WHAT YOU REALLY, REALLY WANT
CREATING DESIRED RESULTS FROM SOCIAL MEDIA RESEARCH

Annie Pettit, Conversition and Research Now, Canada
.
This presentation will teach you how to generate the social media research results you desire regardless of what the true results are. I will demonstrate how to gather social media data from the internet using inappropriate sampling methods, and how to select the wrong pieces of data and code it incorrectly. The topics of sampling, weighting, data quality, sentiment analysis, and text analysis will be highlighted so that you can understand the full range of options for mistreating data. The ultimate goal will be to create set of data that reflects our predispositions towards a topic as opposed to reality.

Attendees are required to come prepared with a sense of humour  (i.e., I will be speaking in jest!)

Social Media Sentiment: H8ers and <3ers

Friday, July 8th, 2011

H8ers! <3ers! Isn’t social media just full of people who have radical opinions? It’s been a while since we first shared information about the distribution of opinions/sentiment in social media so we thought it was about time we conducted our little experiment again.

For six different sets of data, we gathered hundreds of thousands of sentiment scores and prepared frequency distributions of the results. As you can see below, some brands have more positive (A, B) sentiment while others have more negative (C, E, F) sentiment. You can also see that some brands have more flat (E) or peeked (C) distributions, or longer tails (A, B). No matter which particular feature of a brand’s chart interests you, it is clear that all of the distributions are reasonably normal, they are generally bell shaped.

So is social media full of haters and lovers? Most definitely not. Most social media data consists of lots of moderate like and dislike, plus a healthy representation of haters and lovers. Now the only puzzle is determining which of these charts reflects the sentiment of 1) autism, 2) Lady Gaga, 3) Obama, 4) Steve Jobs, 5) Toyota, and 6) Walmart.

.
.
Related links
Are Only Crazy People Commenting About Brands in Social Media?
Article in the Vue: Words I’ll Live to Regret
Cell + Survey + SMR: A Social Media Mashup #MRIA2011 #MRA_AC #MRX
There is no question but the research validity question

.

Dear Brand, you are never to talk to me even if I talk to you first: The Social Media Puzzle

Tuesday, July 5th, 2011

Ok, so in the last couple of posts, we agreed that not everyone feels comfortable about brands responding to comments they make online. Sometimes, people want to make a comment and then just be left alone. But when is it acceptable for brands to respond?

We conducted an online survey (based on the e-Rewards survey panel) to determine how people feel about companies responding to them about a comment they made in the social media space. We surveyed a census representative sample of 1000 Americans and 1000 Canadians to find out in which situations they thought it was appropriate for a company to respond to them about comments they’ve written in social media.

First of all, it’s interesting to see that there are no situations where there is 100% (or even 90% or 80%) agreement that it is appropriate for a brand to respond. In fact, even when people write a comment on a brand’s website, only about 54% of people think it’s acceptable for the brand to respond to them. And, it doesn’t matter if they casually mention a brand or tweet directly to them, about a quarter of people indicated “it is never appropriate for a company to respond to you about comments you have written in social media.” Wow!

So who are these people who think it is never appropriate for a company to respond to them? Perhaps it is some odd, demographically unique group of people? No. It’s very slightly more men, slightly fewer younger people, slightly more uneducated people, and slightly more people without children. Just remember, when you generalize to say that most people are ok when a brand responds to their comments in social media, you’re ignoring the preferences of a lot of people. 26% of people.

Commenting in social media means you want a response, right? Wrong!

Thursday, June 30th, 2011

If you talk to people about social media engagement, you’ll hear a common thread – people want to be responded to when they make a comment about a brand or company in social media. People like it, they expect it, it should happen. This always makes me wonder about people who make comments online but don’t necessarily want to chat with the company. Do such people exist?

We conducted an online survey (based on the e-Rewards survey panel) to determine how people feel about companies responding to them about a comment they made in the social media space. From a census representative sample of 1000 Americans and 1000 Canadians, we identified 152 people who said a company had responded to them when they made a comment online.

Given a multiple choice question listing a number of positive, negative, and neutral feelings, we asked survey responders to select as many items as appropriate to describe how they felt (which means these numbers will add up to more than 100%).

About 41% of people said that they liked being responded to and about 40% of people said they appreciated being responded to. That is a nice, healthy, positive number. However, about 10% of people were annoyed and about 10% felt that they were being stalked. Would you be comfortable annoying 1 out of every 10 people you talked to?

Now, if about 41% of people liked or appreciated it when the company responded to them, that left a lot of people who did NOT necessarily like or appreciate the response. Perhaps they didn’t care one way or the other, or they actually disliked it. Either way, they did not feel the need to indicate that they liked the response. What are the demographics of the group of people who didn’t like the interaction?


.
So now let’s think again about the generally agreed upon idea that people want to be responded to when they write something about a brand online. Is that really true? I don’t think so.

.
.
Related links
Are Only Crazy People Commenting About Brands in Social Media?
There is no question but the research validity question
Cell + Survey + SMR: A Social Media Mashup

.

6CFBPK6DVC9T

Are Only Crazy People Commenting About Brands in Social Media?

Monday, June 27th, 2011

We know who uses social media to comment about brands. They are crazy people who have nothing better to do with their time.  Really? Is this still true?

We recently conducted an online survey (based on the e-Rewards survey panel) to determine who participates in social media. We talked to a census representative sample of 1000 Americans and 1000 Canadians and asked people about their social media usage.

Specifically, we asked them “Over the last month or so, have you written any comments or questions about a brand or company in social media?” I’m sure a lot of survey gurus will have fun pointing out alternate wordings of that question which would have generated far more accurate and precise results, but let’s take the answers for what they are: An indication of who is using social media to talk about brands.

22%. Let that number sink in.
.
22% of people have knowingly written something about a brand in the social media space. Add to that even more people who don’t realize they occasionally chat about brands and we’re working with a lot of people. It’s not just men, it’s not just educated people, it’s not just people with kids, it’s not just employed people. The only demographic that skews a bit away from average are older folks, but even a good percentage of them share their voices online.

So what kinds of people comment about brands in the social media space? All kinds of people.
.
.
Related links
Social media monitoring vs social media research: Can you see the difference?
The Conversition Hierarchy of Social Media Insight
Observational Research – The Original Research Method

Article in the Vue: Words I’ll Live to Regret

Friday, June 10th, 2011

This article was originally published in the Vue, June 2011.  When people answer research surveys, they don’t sign their names or write their email address. This degree of anonymity gives them the power to share their true feelings with as little pressure as possible. But what happens in the social media space? Read along to find out. It can be pretty scary!
.

.
.
Related links
Social media monitoring vs social media research: Can you see the difference?
The Conversition Hierarchy of Social Media Insight
RW Connect: Privacy and Ethics in Social Media Research #MRX
ESOMAR Launches Consultation on Social Media Research Guidelines

Write here, write now: Research Live Quotes Conversition #MRX

Friday, June 3rd, 2011

Paul Golden recently published an article in research live about the controversy between automated and manual sentiment analysis. An excerpt is below while the full article can be read here.

Those who bemoan the impact of email, text messaging and Facebook on our ability to write tend to forget one thing: that the people who have grown up with these tools write far more than any generation before them. They might not have much respect for the conventional rules of writing, but they’re making the language their own, and adapting it to different formats and audiences.

In the process they’re producing vast amounts of data, which marketers are keen to make sense of. That means understanding the meaning and sentiment of text – and numerous tech companies are queuing up to help researchers do this, each with grander claims than the last about their tools’ ability to accurately interpret and categorise millions of online comments. These text analytics tools are adept at picking up brand references from the massive volumes of online communication generated every day. But opinion is widely divided on their ability to determine whether a comment is positive, negative or neither.

….

“these approaches have flaws, but then so do people, says Annie Pettit, head of research at Conversition. “The important thing to remember is that we don’t care if individual messages are coded correctly. We care that the aggregate coding across large numbers of messages validates well. We aren’t trying to prove whether I hate or love Pop-Tarts, we are trying to prove whether a million people hate or love Pop-Tarts. A reading that is 70% accurate is actually a very accurate reading.

If humans can score hundreds of messages and validate at 85%, and computers can score millions of messages and validate at 70%, then computers are a valuable option”

Read the rest of the article here.