The Net-Savvy Executive

Sentiment analysis is generating blog headlines again. After reading about the non-response bias of automated sentiment analysis, and that it has no place in social media monitoring, I decided to run a sentiment analysis on sentiment analysis (apparently I like that phrase). I have an account on Biz360 Community from my current project (look for it next week), so I tossed it a quick query and found that the recent buzz was mostly… positive? Hmm.

Here's the thing. Sentiment is not the golden metric. Virtually every social media analysis platform can show you a pie chart on sentiment. At best, it's a first glance—which way is up. Unless you go deeper into the data, all you're looking at is a mood ring.

	Color	Brand Mood
	Green	Happy
	Red	Sad
	Grey	Confused

At the very least, you need to compare sentiment across brands and over time. Yay, it went up! Aw, it went down.

Oh, look, a mood ring. Maybe there's a secret decoder ring somewhere around here that we can wear next to it.

Set aside the methodology question
The automated sentiment debate continues, but I want to focus on what to do with sentiment data once you have it. On scoring methodology, remember that it's not a simple question of human vs. computer (though this Attensity post explains more of the automation than most people have probably seen). Most of the social media analysis (SMA) platforms I've just reviewed allow users to edit sentiment scores, so when you find a post with the wrong sentiment score, you change it. About half of the automated sentiment processors learn from users' changes, too.

But today's topic is what to do with the sentiment data you have.

Trends, segments, and causes
Sentiment, by itself, is a mood ring—a happiness indicator. It's nice to see the happy color, but there's not much information there. If you dig into the sentiment data, though, it starts to contribute to useful analysis.

Take the trend chart. Direction is interesting, but what about slope? Sudden changes are especially interesting. Any spike—not just in sentiment, but in volume or anything else—is the chart's way of saying "look over here." A spike on a chart is a big ol' why, waiting to be asked.

Sentiment really gets interesting when you combine it with other measurements. Most SMA platforms use sentiment scores as a filter for segmenting the data. What are the prominent and emerging topics within negative-sentiment content (and again with positive)? How does sentiment compare within a topic, across different media types or specific sources? Is a topic emerging from a source that writes negatively about you, or is it a friendly source?

Crafting queries and combining filters could be a whole series of posts, or maybe a book. That's why insight isn't automated: what you do next depends on what you find. If you're looking at the mood ring and wondering what it means, you haven't even started.

Join me at the Sentiment Analysis Symposium (New York, 13 April), where I'll talk about how to make an informed purchase decision in social media analysis.

It's been 10 days since the first AnalyticsCamp. 11 days since the last-minute preparations and wondering if the weather would force us to cancel, about eight weeks since we committed to the date. All in all, we had a good day, including the predictably painful decisions about which sessions to attend—and which to miss. For me, it was the BarCamp where almost every session was appealing, and I needed many more hours to spend with the people there.

"We" means "we"
I'm not using the royal we when I talk about AnalyticsCamp; a small group of volunteers put in significant work to make the event happen, and a few organizations provided crucial support. Without their contributions, AnalyticsCamp would still be stuck in the idea phase.

• Tong Vudhikosit (@tong_Orn) took the lead at UNC Kenan-Flagler Business School, bringing together multiple student clubs and coordinating with the administration throughout the project. For those of you in hiring positions, she's about to finish her MBA and is interested in marketing, social media and analytics as a career. She's already demonstrated her ability to take on a project and follow through to make it happen.

• I-kong Fu (@ikongsgf), Brian McDonald (@bmcd67), and Varsha Chawla (@VarshaChawla) took on various pieces of the project. They already have jobs. :-)

• The Kenan-Flagler facility is outstanding, and we got nothing but support from everyone at the school. If you can arrange to host your event at a top business school with a beautiful, new facility, I recommend it. Really, when was the last time you went to a conference that had installed power outlets at every seat and solid wi-fi in every room?

• We had great food and drinks, provided by our sponsors. With no t-shirts or banners, the sponsors didn't get a lot of visible recognition, so I want to thank SAS and Capstrat again for their support.

Blogging AnalyticsCamp
I could tell from the tweets that I was missing some good sessions, and all I could do was hope that people would blog them. Fortunately, some did. I like the positive reviews, of course, but I'm particularly happy with the reactions from folks who weren't sure about the unconference format going in (even some of the organizers). Get the right people together to talk about an interesting topic, and it's really all you need.

• Why Social Needs Search, Ellie Johnson (pre-event)
• All about AnalyticsCamp, I-kong Fu (pre-event)
• First Analytics Camp Is A Hit, Martin Smith
• Social Media, Analytics and Story, Gregg Morris
• Analytics Camp, or I Cannot Fathom Numbers, Jay Dolan
• Designing Dashboards Successfully, Angela Hall
• Interact With Twitter Friends Using Just Your E-Mail, Damond Nollan
• One time at AnalyticsCamp..., Jim Hazen
• Learn How To Increase Your Readers, Followers, Fans, and Buyers, Damond Nollan
• My first unconference, David B. Thomas
• Content Analytics 'All Abouts,' Angela Hall
• The Tao of Bar Camp, Jack Perez
• B2B Marketing Analytics, Jep Castelein
• Real Time Analytics Take Aways, Angela Hall
• Analytics Camp NC, Andy Ratcliffe
• Encourage the (New) Conference Speakers, Andy Ratcliffe

Lessons
For a first effort, I think AnalyticsCamp went well, but we did learn some lessons. The agenda ended up too front-loaded, probably because we advertised the availability of too many spaces. Next time, we'll have someone to help keep the schedule balanced. We assumed a 20% no-show rate in our planning, but we ended up with more like 40% no-shows. The weather was a factor, but we turned away people who could have been there, and we had a lot of extra food at lunch. We'll allow more registrations next year.

What's next?
Before the day was over, people started asking about next year. We do want to do it again next year, but planning won't start until late this year. There was some discussion of starting a regular analytics meetup, but nothing to announce yet.

I'm working with a few folks in the Washington area to put together AnalyticsCamp DC. I've also heard from some folks who are interested in having similar events in Dallas, Milwaukee (or Chicago), and Boston. The thing about BarCamps is, nobody owns them, so if you want to have one in your city, go for it. Just let me know, and I'll do what I can to help you promote it.

Hey, I finally get to go to London! I'm speaking at Monitoring Social Media Bootcamp* on 31 March, along with Marshall Sponder, Philip Sheldrake, and Katy Howell. Luke Brynley-Jones is putting together a full day of how-to sessions on monitoring and measuring social media, and I'm happy to be a part of it.

The sessions are all about practical, let's-get-to-work topics:

1. Getting Started with Social Media Monitoring
   
2. How to Choose The Right Social Media Monitoring Tool
   
3. How to Build Your Own Social Media Monitoring Service
   
4. How to Monitor Sentiment and Benefit from The Insight this Provides
   
5. How to Identify influencers and Build Valuable Relationships with them
   
6. How to Monitor and Engage with Customers in Real-time
   
7. How to Measure the Success of your Social Media Marketing Campaigns

A couple of years ago, I realized that my business network was stronger in London than where I live (probably not the case now). Now, I finally get an excuse to meet more of you in person. Wednesday, 31 March, Bootcamp.

*haircut not included.

Consulting Days in London
While in town for MSMBC, I'm going to have time for other meetings, too. First priority goes to paid consulting days, though I hope to connect more casually, too (tweetup?). I'm offering full- and half-day appointments, where the content will be tailored to your situation—perhaps a market update on SMA software platforms or a planning session on listening strategies. For details, send an email to nathan@net-savvy.com.

Only a few hours left in the year, and I haven't fulfilled the apparent obligation of bloggers to post some sort of year-end list. Best of, worst of, predictions... you've seen 'em all by now. I have my predictions, most of which are embedded in my business plan, but the prediction that is most on my mind is this: 2010 is looking big from here. So, before the ball drops and the fireworks go up, here's a look back at the most-viewed posts in 2009:

1. Monitoring Social Media Before You Have a Budget - May 2008

2. The Sentiment on Sentiment Analysis - September 2009

3. Defining social media relations - November 2006

4. Visual text analysis - April 2007

5. Corporate social media specialists - September 2007

6. Sorting out social media measurement - July 2007

7. Human vs. machine analysis - April 2007

8. Social Media Analysis for Workgroups - August 2008

9. Guide to Social Media Analysis - June 2007

10. Companies Downplay Online Reputation Risk - March 2009

It's striking how many of these are old posts—only two of these are from this year. I suppose I should look through the archives to see what else is hiding there! To be fair, though, the most-visited page is the front page, which always has the most recent posts, and RSS subscribers have about doubled since last New Year's Eve. So somebody's seeing the new stuff. :-)

I'll keep focusing on quality over quantity, but I'm not about to stop writing here. Any topics you want to get into?

While the archive looks back, I'm definitely looking forward. '10 looks big from here. I hope it's big for you, too.

Here's a crazy idea: if an event you want to attend doesn't exist, organize it yourself. Which brings us to AnalyticsCamp (@AnalyticsCamp), a free day of networking and learning for folks interested in any kind of analytics (web, email, social media, marketing. big enterprise BI, you name it). The plan is to attract a mix of different analytics specialties and stir.

It's a Barcamp-style unconference, and anyone can pitch a session (we have some ideas), but we're also planning to seed the agenda with some serious experts to make sure we have solid content. Sessions will include technical, business and career topics, from beginner to advanced levels, so everyone is sure to learn something.

AnalyticsCamp grew out of a panel discussion on measuring social media at a recent Web Analytics Wednesday. It's intended to be primarily a local event, building community among interesting people working in analytics silos in the Triangle. But if you'd like to join us from out of town, you're welcome. You could lead a session while you're here, too. ;-)

Details
AnalyticsCamp will be held on Saturday, February 6 at UNC Kenan-Flagler Business School in Chapel Hill, NC (directions). Our hosts for the day are UNC's Marketing and Business Technology Clubs. Details are posted on the AnalyticsCamp wiki, and registration is open (it's free, but please register to help us plan appropriately).

UPDATE: Here's my wrap-up post from the day.

Posting has been light here lately, but I have a good excuse, really. I've just gone through 30 demos in a comparison of 23 social media analysis platforms, from Alterian to Whitevector and including all the usual suspects. Now, I'm in the testing and writing phase of a comparison report that will be available in a few weeks. I'm getting hands-on with a lot of software, so blogging, tweeting, and even lunch are taking a back seat to the project.

The focus this time is on software that's designed to support social media capabilities in a multi-user, multi-project environment (which describes most companies and agencies). While asking the usual questions about features and coverage, I'm noticing interesting trends:

• Tools are aligned with the 5 modes of listening; not all platforms try to do it all.

• At the high end, companies are getting serious about social media analysis as enterprise software, adding features that address IT interests like user administration, security and system integration.

I'm test driving software options you probably haven't considered yet. At this point in the project, I'm fairly certain you don't want to duplicate the effort, but I think you'll be interested in what I'm learning. Meanwhile...

Nose. Grindstone.

In my copious spare time, as they say, I'm also working on an analytics unconference—mostly because it will be an event I want to attend. We're not quite ready to announce the details, but the wiki gives the general idea.

Without Internet access, there are no Internet media—no "social media." Unless you've worked in network infrastructure, you probably don't think about it, but a lot of work and investment go into providing Internet access, whether you connect at work, at home, or at the corner café. As you may have heard, Internet access is not equally available around the world.

A new infographic from Appfrica International, Infostate of Africa, looks at the current state of Internet access in Africa. It's full of interesting information, and demand for the graphic has led to its availability as a poster (wouldn't that be nice as a series covering the world?).

Looking at the map reminded me of a point Hans Rosling (@Hansro) makes in his AIDs talk: Africa is not a country, and characterizing regions oversimplifies the answers to meaningful questions. That observation applies equally to the question of Internet availability.

I thought both the map and Rosling's talks (all of them) were worth sharing with you. We now return to your regular workday, which is already in progress.

Is it a problem of overpromising/underdelivering, or are people developing these unrealistic expectations on their own? Either way, I'm seeing more examples of people who seem surprised that software doesn't do all of the work in social media analysis. I really don't think this is controversial: regardless of your choice of tool, there's a necessary human contribution to the process.

This notion that the software isn't good enough because it requires a person to do something with it seems to be picking up speed. The first post that really kicked off a conversation was probably Asi Sharabi's. I saw a couple this morning, including one from Mark Schaefer that focuses on the graphics:

I’ve been spending time studying the trends in social media monitoring and have been impressed with the rapid progress. But there is still a lot of noise like this chart that really tells us nothing. The fact is, the most meaningful keyword and sentiment analysis is all still being done MANUALLY.

I'm not arguing that anybody's tool is perfect (the steady stream of updates strongly implies that the vendors don't think that, either). This is still a new category, and the software will evolve. So observations about which pieces work well—and which pieces need improvement—provide a valuable contribution. But we're not going to see a product that (a) analyzes the world, (b) develops meaningful insights, and (c) delivers it in a tidy, executive-ready package.

Building the social media spreadsheet
Think about spreadsheet software. When you first open a new spreadsheet, the software gives you a blank page. In the right hands, the software is a power tool for running financials, forecasting results, analyzing historical data... I've seen some impressive examples, but the most powerful spreadsheet software is useless without someone who knows how to use it. Which, if you think about it, is true of most software.

Social media analysis tools are software; they do some of the work, but to get the most out of them, you need someone who knows how to use them. The more you expect from your tool, the more the user needs to know. Anything that's fully automated either isn't doing much, or it has a lot of human effort baked into it.

Regardless of the tools used, at some point people take over. It may be earlier in the process (manual content analysis) or later (analysis and reporting), but eventually, a person takes what the computer produces and does something with it. All of those agencies that sell services based on the same SMA platforms presumably think this is where they add value.

If you want it done for you...
There is an answer for the company that wants the insights without putting in the effort, of course. Have someone else do it. You can't bypass the requirement for a human analyst, but you don't have to do it yourself. When you're shopping for social media analysis, just be sure to include analyst services among your requirements. That eliminates some of the best-known software companies, but it opens the door to an entirely different set of service providers.

If you want to make spreadsheets, you buy Excel. If you want financial projections, you roll up your sleeves or hire a financial analyst. It's up to you to decide whether that analyst will be an employee or work for an external service provider.

Where's the disconnect? Are unrealistic customer expectations coming from vendor hype, or is it just hope that things will be easy?

I try not to be too obnoxious with my iPhone, but it's hard to avoid being impressed with how it's changed my expectations about mobile computing. I've been wondering when social media analysis apps would start showing up, and now I've found one to play with: iCrossing's new Say What?

Compared to the tools I usually look at, Say What? (iTunes link) is pretty basic. It runs searches across Twitter, Digg, forums and blogs, returning a sampling of the results from each. It's not much, but it might be enough to get a quick sense of what's going on with a topic (especially if it's currently controversial).

Why are people talking about that?
Say What? is best used for getting a clue about a current topic—especially if it's controversial or newsworthy. Looking at results, four per screen, you're looking for someone to provide a hint about what's going on.

Looking up (Rush) Limbaugh, I immediately found mentions of his interest in buying into an NFL franchise. Ford returned items about the latest product recall. But when the company isn't making news, the results are far less interesting—people are apparently having breakfast at Dunkin' Donuts.

From Search to Analysis
I've been looking into web-based collaboration and project management tools for my own company, and as soon as I realized that some of these tools have iPhone apps, that became a requirement. We have so many web-based tools for monitoring and analyzing social media; who's going to be the first to offer a simple dashboard that delivers clients' data to smart phones?

Think of your favorite model or metrics for measuring social media activity. Flip through Olivier Blanchard's presentation on social media ROI. Now, with that in your head, read Tom Davenport's 2007 book, Competing on Analytics. How far do you get before realizing that the enterprise analytics crowd is asking some of the same questions as the social media crowd, but looking for answers in different data? What if the two groups met?

When we asked CIOs to identify their visionary plans for enhancing their enterprises' competitiveness, business intelligence and analytics was the top answer, selected by 83 percent of our sample... "Facts drive decisions," said an Insurance CIO. "Plans for imbedded analytics need to enable data capture at the customer touch point."
— IBM's 2009 Global CIO Study (via KDnuggets)

What would happen if you were to analyze social media data alongside operational data to look for insights in the interaction between what people do online and what they do with your company? You could measure the ROI of marketing in social media, but that's a defensive move (protecting your job/budget). Beyond learning what works and what doesn't, what would you learn by looking at the data together?

Are you doing this now? I'm looking for companies to interview for my research.