Tag Archives: Facebook

Why You’re Not Learning What You Should by Sean Bruich, Facebook #Eso3D #MRX

Image representing Facebook as depicted in Cru...

Image via CrunchBase

This is a live blog posting from the Esomar 3D conference in Miami. Written, summarized, and posted just minutes after the speaker has finished. Any inaccuracies are my own. Any humorous side-notes are mine as well.

***                                                                                                                 ***

GUEST SPEAKER
Social media research
Why you’re not learning what you should
Sean Bruich, Head of Measurement Research, Facebook, USA

  • How can we use facebook to understand what consumers do, how they interact, how they are influenced?
  • 800 million people are on facebook every month, half of them every DAY [that’s me!]
  • Types of data on facebook 1) Site/mobile/meme usage, 2) interests, locations, chatter revealed 3) polls, posts, comments solicited
  • How to make feedback fun and easy? Surveys from 1930, the first GM survey, look exactly the same as today. And it’s not easier, nicer to respond to now. [Who’s fault is that?]
  • what is a survey? Stop what you’re doing, leave your trusted environment, give away your info to someone you don’t know, answer questions you don’t care about, get nothing in return. [Sigh. This is pretty much true but we ignore it.]
  • The GM survey did result in something the survey taker could benefit from. They got a buyers guide in return.
  • Facebook never asks people to leave the site to take a one question survey. People are already in the mindset of sharing while in facebook. RR are massively higher. Breadth of respondents massively wider. They can get 2000 rep respondents in 5 minutes.
  • External validity for political opinions. They ask identical question once per week. Their data matches pretty much perfectly. correlations of .9 and higher. [Nice chart. the lines overlap too much to see which is which. How’s that for validity.]
  • Predictive validity for movie viewing intention. Correlates over .9
  • Lightweight facebook users respond more strongly to ads on facebook than heavy users.  Ad recall is higher for lightweight users. These people watch less tv, read fewer newspapers. There is just more share of space for these people. They aren’t necessarily paying more attention to the ads.
  • What are the goals of text analysis? People have always had these conversations. We just better abilities to look at it now. 1) Attention on text analysis is misplace. 2) How should we use it?
  • Self selection bias for sure. We only talk about what we want to talk about. Words out of context are meaningless. Sure, people talk about winning, but you need to know it’s in reference to making fun of Charlie Sheen.
  • On Facebook, you can count the number of fans. But you also need to consider people who were exposed via the fans. Friends of friends. What is the influence of what was said. [Now I want to be a fan of Bert the Mini Shar Pei, Sean’s dog. Tessie, this one’s for you!]
  • Twice as likely to take away from an ad if your friends are associated with it. viral reach is essential.
  • We need to be thought leaders not just data analysts. [yeah baby!]
  • Traditional view of influencers: Reach a tiny group and you’ll reach everyone else. Oprah is the classic example. christmas special, the book club, oprah endorsed phil mcgraw, obama.
  • New view of influence: everyone has a platform though it varies in scale. [uh oh! over time now. Poor alphonso!]
  • Facebook influence: influence goes down, across, over, and back again. Nothing can predict who is an influencer, not number of friends, not demographics. There is not a small group of powerful influencers on facebook. It’s just normal people with normal groups of friends.
Advertisements

Quick Poll: Have you read the Terms of Service for Twitter, Facebook, and LinkedIn? #MRX

Nope, not scientifically sound, not randomly sampled. The answers are extremely important however.
.
Please answer honestly. If you have not COMPLETELY read the TOS, please answer no. If you COMPLETELY read the TOS, then answer yes.



Answers, and the reasons for asking these questions, will be posted later.

Here’s what’s wrong with Klout

With so much Klout bashing out there, I thought I would add more fuel to the fire. First of all, Klout is a proprietary tool intended to measure influence. As tweeters, we take pride in our high Klout scores, scoff at its validity when we get low scores, and game our way to the highest score possible.

If you want to game it and get a high schore, here are a few tips:

  • join in several twitter chats particularly those where the people have high Klout (e.g., blogchat)
  • tweet only to people who have high Klout scores
  • tweet on weekdays, weekends, holidays, days, and nights
  • connect your Facebook and LinkedIn profiles.

You’ll be rewarded with high Klout even if everyone thinks you’re a total douchebag.

It’s also a fun thing for number geeks to play with. The numbers go up and down on a daily basis, and we can make pretty charts with them. As a geek who’s used charts as positive reinforcement on myself, this makes perfect sense.

If you really want to see how fun Klout is, run some Klout stats on @MRXblogs. That’s a bot I created which shares blog posts about market research. It replies to no one, retweets no one, and never goes to the bathroom. Its Klout is 46. That’s actually quite good even for regular human tweeters. But would you buy social media research from @Conversition if @MRXblogs tweeted one of their blog posts? I doubt it. Though feel free to prove me wrong.

So what is wrong with Klout? Nothing. Nothing at all. I suspect they have plenty of analysts and statisticians who have already thought of all our complaints and have already worked hard to resolve the ones that can be resolved. Remember – they are attempting to deal with innumerable special cases which simply cannot be solved by changing Z = X + Y to Z = X + Y + W. If it was that simple, they would have already done it.

So enjoy your Klout toy. It’s free and it’s fun. What more do you need.

Neal and Toplansky: Thought Leaders Debate #MRA_AC #MRX

Illustration of Facebook mobile interface

Image via Wikipedia

Session summary of the Marketing Research Association 2011 annual conference. These are my interpretations of the session. They were written during the session and posted immediately afterward. Any inaccuracies and silliness are my own.


Keynote: Thought Leaders Debate
William Neal, SDR Consulting
Marshall Toplansky, WiseWindow

Marshall

  • How far removed are we from people who fund the research? Probably 2 people. They have started putting their internal costs online. Started real-time assessment of competitive market, lead time generation, closure as with Comscore. Have not looked at consumer sentiment data until now. People are now focused on short-term measures, this promotion yesterday, competitor today. Money is coming out of MR because MR is too slow. We like to provide insights but they have no value, no currency beyond marketing ghetto, the people who are playing around with messaging, colors, those people are being more marginalized.

Bill

  • Technology has had an impact and it has reallocated funds. We need to remember what differentiates us. Gobs of data is not the solution. We base our work on science, essential, necessary. Survey method, of course has flaws, criticized for corporate execution. WHAT is going on, WHY is that happening? New technology does not address this. 20 million respondents may say one thing but you must ask is that representative of your target market? Millions of hits are not likely to be highly representative. Large groups of people are not and may never be in social media. (What, and everyone answers survey?) Our role is to be voice of arbitrator in C-Suite. Money comes from CMO and we work for the marketing folks. CMO doesn’t have permanent seat in C-suite. SMR may mislead us and consequences will be significant. What are our roots? Be voice of customer. Work for financial folks, not marketing folks. Sales function vs marketing function, push vs pull. Sales is short term, marketing is long term. SMR is not key, it is one part of it. (Phew, i was starting to get upset!)

Marshall

  • This is not quant vs SMR. It’s about empirically read mass observation, correlate sales numbers, real-time data. We are not in C-suite because we don’t even talk that language. We can’t translate consumer preference into a continuous flow of data. SMR is one technique, not only technique. If we ignore SMR, we’re going back to the middle ages of quill pens.

William

  • What should we measure for corporate well-being and growth? Must pay attention to brand equity. Must manage at C-suite level. False indicators may be sentiment. He has not seen any quant SMR data (HELLO! I’M RIGHT HERE!!!) It must be valid, reliable, and sufficient to bet on to do things differently. (I had to get up and ask a question here – 600 million people are on facebook, 3% of the world answer surveys. What does rep mean to you?)

Marshall

  • People are working on the social media research method right now. All methods get worked on and improved by having the smartest scientists working on it. He mentions more people are tweeting than will read the Alert article. (Dang right! Plus this blog and anyone else who will blog the session.) This is how people express themselves. You’re a stick in the mud if you don’t jump on it.

William

  • He is a strong advocate of mixed mode methods. (He should have come to my session yesterday where I showed how to use SMR, Surveys, and Cell together.) He wants rep samples. (NONE of market research is really rep samples.) We don’t capture all kinds of people, older, lower income, laggards, there are lots of these techno-phobes. We shouldn’t walk away from the technology. (Agreed. Since when did surveys capture all people? I don’t answer my home phone so how is your “rep” RDD?)

Marshall

  • Social media interaction for a small business is great, bulletin board style. Lots of richness. Must use best practices of research methods, do not bias, don’t lead the witness, pull a good sample. (Damn right) Methodology can be completely different.

William

  • Worry about perception that SMR is quantitative. (Ouch! Seriously, come see what we do William! 100% quants, scales, norms, sampling, weighting.)

Marshall

  • Representativeness is crap. This comes out of old media industry. Which demographic segment should I target, we allege we require. It reflects what we can buy in media. Marketers want to go to social segments, not demographic segments. They just want to find them behaviorally. Do we care if segment is 12.3% vs 9.7% of population? No. Let’s go after similarly looking people. People talking about my product is representativeness, not demo rep.

William

  • Brand hand raising is representative, he agrees. Demos as basis for representativeness should have died a long time ago. We don’t lack tools to look at data and model from it. People aren’t aware of these tools. Corporate research depts don’t know tools well enough to know to use them. We have a knowledge problem, not just a money problem. (I agree we have a knowledge problem.)

Marshall

  • Huge believer in multi-mode methods. Traditional methods can never be replaced. We need better ways of gathering data quickly, immediately. MR is just too slow and too expensive.

William

  • Can’t be too dependent on  high levels of qual data. Not sure if it will give us the right decision. Doesn’t believe any data is better than no data. Any data may be wrong. It’s high risk. We need risk reduction which comes from multiple sources.

Marshall

  • People don’t necessarily lie in their profiles but they just don’t tell you. People don’t put correct age in Facebook. (I don’t. I’m usually born in 1930 when I sign up for things online.) Maybe 10% of people give you geography. We need to solve this problem. Will facebook make money by selling profiles with proper privacy? Yes.

William

  • It’s not that people lie, they are just not being candid. Yes, people lie on surveys but that is not bigger problem. Candidness is. We need to do a lot more investigation. Don’t see if SMR will ever have sufficient validity and reliability to be predictive. (Dude, read the literature. People are already proving this.)

Marshall

  • 120 million facebook comments per month – this is the law of large numbers. Strong correlations between what they say and what they actually do in terms of sales is. Outliers are bled out in the volume of data. There are thousands of mommy blogs, people talking about underage children who aren’t allowed to go online. Children are represented online. We don’t need specific lines/channels of people.
  • William – Cascade effect of social media may or may not reflect reality. Facebook is “going with the herd.”
  • Neal – This is how people influence others.
  • William – Cascade changes reality of the percent. 35% becomes 65% because of influences and cascade. It’s not reality.
  • Neal – There are strong correlations between market share of opinion and market share of real world.
  • William – I need the last word. Watch out for herd effects. SMR is just another tool. Don’t overuse it.

When Brands Had Power

In ancient times, more than 20 years ago, brands had undeniable power. They created brand stories and images and slogans and taglines. They designed commercials and billboards and posters. And they had the millions of dollars required to force their messages on captive audiences.

Consumers had little choice but to watch, listen, and believe for there was no alternative. Sure, you could discuss with your friends how much you disagreed with the ads but you certainly didn’t have a million bucks to create and publish your own anti-brand ad.

Those days of one-sided power are gone. Facebook and Twitter and YouTube stole the power from brands and threw it at consumers. Consumers who have no brand experience, no scientific bases, and no expertise have a voice, a very loud and annoying voice. And consumers who know what they’re talking about, people with experience and expertise have a loud voice too, but slightly less annoying.

Brands can write even wittier taglines and even more creative commercials but consumers can fast-forward through those messages and counteract them with their own widely publicized blog rants and YouTube satires.

Quality must come first. Quality can no longer be your tagline. Sorry Ford. Quality must be now be a fact.

Frame this Gorgeous Mercedes Benz ad on your living room wall #MRX

Every television show has a bazillion commercials. Every magazine is 50% ads. The internet is one giant ad. Ask me what brand I just saw advertised and I might be able to tell you if I saw a man or a woman but certainly not the brand name.

Shift to this ad. This is truly one of the most beautiful ads I’ve ever seen. You can fight to get get my attention with celebrities touting your brand or stunning coupons or facebook only ads. This ad had me taken a second and third look and appreciate the beauty. And. I know the brand name. Mercedes. I’ll probably even tell my friends about the ad. I may not buy the car but I’ve engaged and shared.

Pretty.

Will Goodhand: Social Media Research and Digividuals #netgain #mrx

Image representing Twitter as depicted in Crun...

Image via CrunchBase

Will Goodhand, Chief Brainjuicer, Comedien, Brainjuicer
.
What follows are some of my silly musings and key take-aways of the session.
– Going to discuss how social media research works on a very basic level
– “Replace fear of the unknown with curiosity”, fear that clients have (and researchers too!)
– Data is the curse of market research: how do you gather, reduce, present, engage
– What about the unknown unknowns? We know what we don’t know, but what about the things we don’t know that we don’t know?
– DigiViduals – digital likeness of a person, we built it digitally, it has personality traits but it’s not a real person
– Opinion polling is wrong or so far out there it’s useless (wow! wild claim!)
– We need a lighter touch to research because people don’t always have access to their own reasons for doing or saying things
– Even if people aren’t trying to lie, and are being honest with themselves, the act of conducting research changes the results. Hurray! No more questionnaires!
– Data that comes from digividuals is surprisingly intuitive
– You can build personalities of people/brands using the words that they’ve chosen, it’s like building a tv show character
– You can go beyond the basic measures, like brand usage, and consider the sidebar conversations, like which rock bands they listen to
– Twitter is the most human dashboard you’ll ever see, you get words, pictures, music, video
– They created a character Ian using their bot engine. Gave him an emotional profile, gadgets, progressive, energetic, etc. Then the bot searches out tweets that match his profile. (I’ve never heard of this kind of research. I’ll need to check it out!)
– Within 5 posts, they always get an Iron Maiden video. (ha! 6 degrees of Kevin Bacon 🙂
– The collection of posts hitting these terms generates common themes of this fake person, this digividual. A person fond of music, using colors, lovey dovey books.
– You can do this same thing to products. What kind of TV would this digividual want?
– “insight” got used and I think it might actually have been an appropriate use!
– This research is very shareable (and easily understandable by researchers and non), let’s you get to understand your customers, creates concepts to test, brings trends to life
– Really great presentation, nice mix of methodology and results

Related Links
#Netgain5 Keynote Roundup: Last Thoughts
Brian Levine: Neuroscience and Marketing Research
Brian Singh: Insights from the Nenshi Campaign
Monique Morden: Online Communities, MROCs
Ray Poynter – Overview of Online Research Trends
Tom Anderson: Web Analytics
Will Goodhand: Social Media Research and Digividuals

In Honor of Infographics. #MRX

Infographics have become a staple of the internet. Every self-respecting journalist, artist, and blogger is desperate to discover and display a unique and stunning infographic on their own website. And, in honour of the great and powerful infographc, I too have jumped on the bandwagon. I have created this stunning infographic of infographics. Please enjoy and share with all of your friends and infographic specialists.

You are too old to understand social media so don’t even try

Logo for the Addicted to Social Media Blog

Image via Wikipedia

It’s a phrase I’ve heard numerous times over the last year and frankly, I’m quite tired of it. “We need to get some young people in here who know what social media is.”

Since when is social media for young people? (A certain RP can help me attest to this.) Since when are “old” people who have Phds and 30 years of research experience incapable of learning something new? Since when are “old” people ineligible to be knowledgeable about important data sources?

So, to you old farts and young lazy asses, here is my guide to get you back on track and eligible to speak intelligently about social media.

1) Get a facebook page. Friend at least 20 people including friends and family and colleagues, become fans of at least 20 brands like KFC and Old Spice, join at least 20 groups. Read what people are saying on your page, on the fanpages, on the group pages, and write on other people’s pages. Like their messages, follow their senseless links and watch their stupid videos.

2) Get a twitter account. Follow at least 50 people. Actively participate for at least 2 weeks. Write tweets. Read tweets. Reply to tweets. Click and read the links people share.

3) Visit youtube at least 5 times. Watch the videos that make no sense, that seem inappropriate and offensive. Read the video comments. Leave video comments. Reply to video comments.

4) Find a few blogs that interest you. (You are reading a blog now.) Read them. Comment on them. Ask if you can write a guest post for them.

After following these guidelines for two weeks, you will have a fairly good feel for how people communicate online. You won’t know all the details but at least you’ll know what your colleagues are talking about when they say DM and retweet and buzzup.

To really internalize social media, follow the list religiously for two months. (Then try to quit cold turkey.)

If you keep it up as a part of your everyday routine, you will have part of the qualifications to be a social media researcher. (The others including knowing how to design research and use statistics.) Social media is a share and share alike space. So share.

ARF AM5 Day 2: Bees, not buzz but busy!

Coconut macaroons, a type of popular meringue-...

Image via Wikipedia

[tweetmeme source=”lovestats” only_single=false]We started bright and early at 8am again. Today, I claimed my seat in the “reserved for speakers” row right in front. Ok, so maybe I wasn’t speaking on those specific panels but people are scared to take the front row and I like it. How else are you to see peoples faces without squinting and take good videos!

The first few sessions reflected a normal curve with a yeller and a reader on the outlying points with most speakers taking a more moderate approach. My fave talk of the morning was a discussion of whether the ipad is a revolution or an evolution. I’m not sure where I stand as I don’t buy magazines to begin with and certainly wouldn’t spend hundreds of dollars to then spend a few more bucks on a magazine. But then I’m just cheap. I mean frugal. Yeah, that’s it.

Then, I was off to record a 15 minute video summary of my presentation. Somehow, it turned into an hour and a half discussion as the video team and Steve Rappaport had a unique interest in my topic. It does always seem to intrigue people when I explain how social media data can be properly analyzed and turned into research. And then, Joel Rubinson arrived and I got to admire two great minds at work!

Lunch was very nice and surprise surprise, there were macaroons for dessert! Not the coconut macaroons we are used to, but the meringue style you find all over Paris. Needless to say, after I left the table, there were no more left. 🙂

Then came my session. Between myself, Stacey Hall, Heather Milt, and Sean Case of Peanut Labs, we presented a session on using social media research to evaluate television viewing. Stacey started the show by describing just how massive the volume of social media data is. Then, I followed with a discussion of the research itself. And, given that USA had just won a world cup game, it only seemed fitting to begin and end the presentation with a blast from my blackberry vuvuzela app! (On that note, if you are annoyed by the sound of the app, make sure it isn’t your own making the noise when you wander into a quiet restaurant!)

The best compliment received after the presentation came from an elderly gentleman. You could stereotype and say “what does he care, he remembers the 1930 census.” Well, age is no measure of social media research appreciation. He watched the presentation and announced to us he was a convert. 🙂

The day ended with several more panel discussions slotted quickly after each other. Again, there was barely time to breathe before more neat ideas were rammed 2 by 2 into my head. I’m going to have to read my tweets as a refresher.

Finally, the most fun of the day was taping live commentary from the ARF team. There were some refusals but they quickly turned into videos. How could they say no when I won the flip video from them! Watch this space for that video!

As a newbie to this conference, it was definitely an eye opener for me. I heard a lot of new ideas that will float around my brain while I figure out how to apply them to my work. I also heard lots of familiar buzz words, met lots of new people, and listened to lots of tweeple. Now I can’t wait for the next conference. Maybe it will have 20 minute breaks?

Another A! BZZZZZZZZZZZZ

Read these too

  • Radical Stats
  • Ode to a Boxplot
  • Survey Design Tip #2: Short and Sweet
  • Survey Design Tip #1: Responders DON’T CARE
  • Privacy – Do you really think about it?
  • %d bloggers like this: