Archive for the ‘conversition’ Category

Cell + Survey + SMR: A Social Media Mashup #MRIA2011 #MRA_AC #MRX

Tuesday, June 7th, 2011

I must start with a kind thank you to everyone who chose to come to the Conversition session, despite all of the other great sessions that were available at both the Washington MRA and Kelowna MRIA annual conferences.

As well, thank you so much to everyone who shared kind words with us afterward. It’s always nice to hear that people learned something and felt their time was wisely spent.

As requested by many of our attendees, below you will a longer version of the presentation. Please do feel free to get in touch with us if you have any questions at all. Enjoy!

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.

There is no question but the research validity question

Tuesday, May 17th, 2011

As a budding field, social media research is a magnet for questions from both experts and novices alike. People are curious about the processes and methodologies used to accomplish the various aspects of the research. Some of the more common questions we field on a regular basis are as follows:

  • What sentiment analysis system do you use?
  • How do you carry out the text analysis process?
  • What is your method for identifying and eliminating spam?

In fact, each of these questions is one and the same. They have nothing to do with sentiment, text analysis, or spam.  They have nothing to do with processes or methods or systems. In fact, they have everything to do with validity.

Validity refers to truth. Is the sentiment scored accurately? Is the text analyzed accurately? Is the spam identified accurately? Is the entire process valid? Among all the pieces of the puzzle, this is the one question that must be answered.

Unfortunately, there is no single method that automatically identifies a sentiment analysis, text analysis, or spam detection system as being the most valid one. You simply have to evaluate a large, representative sample of data and determine the answer for yourself.  Are your results valid?

.

Also see
The Sharks and Icebergs of Social Media Measurement
5 Ways to Fool An Automated Sentiment Analysis System
A Formula for Perfect Sentiment Analysis

A Review of Thor by TweetFeel

Friday, May 6th, 2011

It’s Friday, Friday, Friday, Friday today which means you’re deciding what to do this weekend. Why not check out Thor, the action movie with Natalie Portman. But first, make sure it’s a movie you’ll like and a fast, free, and fun way to do that is by checking it out on TweetFeel first.

Just type the name of the movie in the search box and watch as Twitter opinions appear, get scored, and contribute to the overall movie rating.  While you’re there, you might as well check out how Obama and the iPad are doing too! Happy Friday!

May The Fourth Be With You

Wednesday, May 4th, 2011

.
.
.
.

Conversition Presents Mobile Mixed-Mode at MRA in Washington

Wednesday, April 27th, 2011

Mobile Mixed-Mode Research Workshop

Date: 06/06/2011
Time: 9:00 AM – 12:00 PM
Room: Palladian Ballroom

Technology has forever changed the way people live their lives and how businesses do business. Gone are the days of conducting a single mode, landline telephone survey to reach a representative sample. Using multiple data collection methods creates many challenges for marketing research. It also provides new opportunities for reaching and understanding different populations. With the advent of new technologies, researchers now have a wide range of options available to them. But what works and what doesn’t? How do we implement those new and still widely debated methods and provide reliable results?
.
In this three hour immersion workshop, attendees will first get detailed explanations of four separate case studies that provide practical applications of online panels, cell phones, smart phones, SMS texting, IVR, social media, mobile web surveys, and tablets in actual mixed mode research studies.
.
Attendees will then be challenged to put together an effective research plan using a variety of methods to collect opinions and sentiment from various “hard to reach” populations in less than ideal locations. This “hands on” session will provide participants with the knowledge of how to implement different data collection methods while understanding how each method augments and complements the others.
.

Conversition at the MRIA In Kelowna BC

Thursday, April 21st, 2011

I have to admit – I’m pretty excited! Kelowna BC looks to be a gorgeous site for this year’s annual MRIA conference. Add to that a ton of great speakers and it doesn’t get much better. We were lucky enough to be selected to speak at the conference and look forward to sharing what we’ve learned since we presented at last year’s conference. We hope to see you on Monday May 30, at 2:30 in the Chilcotin room!
.

Side by Side by Side: The Survey, Cell, SMR Study You’ve Been Waiting For

In 2010, we saw sample social media research results, and we learned about the pros, cons, ethical implications of the method. Now, it’s about time we saw some real examples of how SMR works in cooperation with other methodologies. This presentation will showcase a cooperative study using results from three different methodologies, online panel surveys (fielded by global online sampling and data collection company, Research Now), cell phone surveys, and social media research. We will see where the results are similar and where they differ. We will discover how the strengths of one method build on the weaknesses of another method. We will discover data fusion at its finest.

Chief Research Officer Conversition Strategies

Annie Pettit, PhD is the Chief Research Officer of Conversition Strategies, a boutique market research company that specializes in social media research.

She has more than 15 years of experience as an online market researcher and specializes in survey research, data quality, and social media market research. Annie is a member of the CASRO, MRA, and ESOMAR social media research committees. She was previously the VP of Online Panel Analytics at Ipsos where she created their proprietary survey panel data quality system.

Annie tweets at @LoveStats and maintains the LoveStats marketing research blog where she occasionally showcases her attempts at being a better baker and gardener.

Recherche sur les réseaux sociaux pour les personne #MRX

Tuesday, April 19th, 2011

.
.
Related links

Nintendo 3DS: Pilotwing vs Dead or Alive

Wednesday, April 13th, 2011

If I could make this post 3D I would, but until that day, just imagine that the charts you see below are bursting out of the screen and poking you in the eyes. Unless you’re wearing glasses. Then you’re safe.

Consumers have had some time now to try out the 3D system and decide whether they like it. Many people have already experienced the headaches, sickness, and vomiting that has accompanied this new form of brain candy. And if you’ve decided the headaches are worth it, or you’ve already trained the headaches away, you’re ready to pick out a new game.

Based on thousands of verbatims from across the internet – blogs, microblogs, forums, video sites, photo sites, and much more – that mentioned the Nintendo 3DS system, we coded and scored verbatims that mentioned these specific games to determine which one should be first on your shopping list.

In the first chart, you can see just how much chatter takes place about each of the games. Games at the top of the list generate the most chatter and that includes Pilotwing, Street Fighter, and Ghost Recon. In fact, these three games generate more chatter than the remaining 13 games put together.

On the other hand, games at the bottom of the chart generate the least chatter and that includes Dead or Alive, Bust-a-move, and The Sims. The six games as the bottom of the list aren’t that much different from each other in terms of the amount of chatter they generate.

But you know very well that volume of chatter does not mean a game is great. It could, in fact, mean a game is so bad that everyone is having a great time making fun of it. We all know about the Double Rainbow and poor Rebecca Black. So let’s make sure we’ve got it straight.

nintendo 3ds conversition social media measure

One of the tricky things with sentiment as that gross, violent, and scary games have a lot of negative words in them that erroneously lead to negative scores. So we took care of that right away. Those words had no influence on our scoring system.

The chart you see here shows the percentage of comments/status updates/messages that were scored as either negative (red), neutral (blue), or positive (green).

First of all, you can see that within Nintendo 3DS conversations, the game generating the most positive sentiment (55% of the messages were positive) was the Dead or Alive game. It may not generate the most chatter but it sure generates a lot of positive chatter.

On the other hand, Bust-a-move, which also didn’t generate a lot of chatter also failed to circumvent it with sentiment. Only about 25% of those verbatims fell into the positive side of the scale. Fortunately, though, almost none of the verbatims about Bus-a-move are negative. These aren’t bad scores but if you can only choose one game, maybe this isn’t the game for you.

nintendo 3ds conversition social media measure

Now that you have the scoop, you’ll have to decide which games are going to make the trip from the store to your home. Lego Star Wars is on its way to mine!
.
.
Related links

The Sharks and Icebergs of Social Media Measurement

Friday, March 25th, 2011

It’s all very deceiving. Social media measurement is superquick and feeds everyone’s desire for instant gratification. Want to measure something right now, lickety split? Here you go!

Unfortunately, it’s not that simple, nor do you want it be that simple. Behind every “instant analysis” are hundreds and thousands of hours of tedious, painstaking work. Let’s take a brief climb up the social media measurement iceberg and check out where the data quality sharks are hiding.

Taking much of the time, the bulk of the iceberg, and many of the data quality sharks is sentiment analysis. There’s no point in being amazed at how quick your measurements were generated if those numbers have no bearing on the true sentiment of a brand.  The data quality sharks biting at the validity of your data include things like failing to measure  “the bomb” correctly and forgetting to measure “FTW.” If these sharks aren’t addressed, your amazingly quick measurement will be wrong.

The next chunk of the iceberg that rarely gets seen is the work that goes into developing variables. The data quality sharks here include recognizing that “tasting success” and “tastes in music” aren’t discussions of how deliciously yummy something is. Behind the scenes, people must spend icebergs of time making sure those invalid mentions are identified and ignored. Otherwise, your instant measure of whether people are happy with the new flavour of Doritos will be wrong.

Now we’re actually ready to gather YOUR data. As always, you can go online and quickly gather a ton of information about Ford or Sears or Harley, but your data quality sharks try to stuff in mentions of “my mayor, Mr. Rob Ford” and “sear the meat quickly” and “my dog Harley.”  If you don’t take the time to kill those data quality sharks, your instant analysis will, once again, be wrong.

The last chunk of work goes into removing spam and irrelevant data. Of course, you can easily set up an automated system to detect and delete spam but you must always be on the lookout for new and improved spam. This data quality shark sneaks in when you least expect it. No checking means no validity.

And now, we’ve climbed to the top of the iceberg, to the last tiny percent of data work that is visible to the outside world. Enjoy your magical, instant measurement. It wasn’t that instant after-all, was it!

Related links