I see businesses struggle with how to benefit from sentiment analysis. Many reviewers and analysts have covered this topic : see this post from Jason Falls. What are the major problems with automated sentiment analysis? accuracy, reliability and variation in the results from different tools (discussions abound about what is the percentage of accuracy for a given tool) and the presence of noise. My goal in this post is to provide you ways to maximize what you can get from today’s sentiment analysis technology.
Here’s my quick guide on how to effectively use automated sentiment analysis in social media.
- Spend time to research the keywords which are best fit for your business objectives and optimize your search based on the result set that you get.
- Use sentiment analysis to detect change in sentiments over a reasonable duration, say 3-months or longer. Shortcuts don’t work well.
- Competitive Analysis – compare your results against those of your competitors.
- Don’t miss the forest for the trees – sentiment analysis is more valuable in the aggregate. Explore the drill-down capabilities but don’t get lost on individual messages.
- When using a tool, look for the capability to re-classify the sentiment, which should train the tool.
- Don’t underestimate neutral comments – there are plenty of business insights and business opportunities there.
- Source is important: Twitter users tend to complain more. Facebook fan pages tend to have more rich media associated with content.
- Understand the limitations of software-based automation: irony and sarcasm are not well understood by software. Content with mixed opinions are hard to catch.
- Context is key for improving accuracy of sentiment analysis. Optimize for your industry or vertical.
- Combine user and content ranking with sentiments to prioritize your results.
I believe that one of the reason for the disappointment with sentiment analysis is the unrealistic expectations from what the science of today can deliver in terms of understanding human sentiments. With the explosion of user-generated content, Sentiment analysis has become one of the active areas of research in Computer Science and Linguistics (for example, follow this link for a recent paper on sentiment analysis).
I plan to write more detailed posts on the topic of sentiment analysis including the various approaches and options beyond fully automated analysis – call it semi-automated sentiment analysis where the tool does more than 90% of the work in automated fashion. I will also share case studies of how MutualMind has helped its clients with sentiment analysis.
I hope this was helpful – please suggest some additional points and share your experience.
Sentiment without action is the ruin of the soul.
— Edward Abbey
References:
I came across an interesting article which talks about the next phase of the battle between the two giants of social networks and search. Both want to dominate how information is captured and shared. Robert Andrews makes the assertion that both Google and Facebook are working to attain similar goals, but using a very different approach.
Facebook and Google are each coming at the same social search prospect from opposite ends of the continuum—one a search site looking at learning relationships, the other a social network trying to upskill in search.
Andrew provides interesting perspective on Google’s long term plan.
Slowly, Google has been learning our social graph. One outcome has been the Social Search feature debuted in October – but this still feels experimental, rather than a final destination, and is only available to U.S. users.
For all the pounding it’s taken, Google Buzz is just one more expression of the graph Google has assembled – the input/output part. To me, Buzz was only ever intended as an inducement to populating that graph. The reasoning: if people are invited to aggregate all their lifestream outputs and to update all their social profiles from one place, they might see value in plugging their social identities in to Google Profiles (after all, how many Google users have ever bothered to create a Google Account?).
The more a search engine knows about you (and the people you know), the more relevant its results might be (and, perhaps, the higher its CPMs), because—the theory goes—we tend to trust personal recommendations.
The article ends with an interesting line: Google is no more a social network than Facebook is a search engine—perhaps even less so. But, while neither has yet cracked this nut, the game is long and the rewards ample.
What are your views on the big war? Which side are you on?
We are pleased to announce the release of MutualMind 1.3.2. Here are some of the highlights from this release:
Near Real-time Sentiments
As of this release, sentiments and sentiment stats are updated in near real-time. You’ll be able to get up to the minute sentiment analysis for all the social content we capture for your campaigns.
Hour-by-Hour Activity Charts
We have also added hour-by-hour activity visualizations for most of the data we capture (whenever applicable).
The following data can now be visualized in this way:
- Sentiments
- Facebook page activity
- Twitter engagement, mentions and retweets
Sentiment Reclassification
We’ve added the ability to manually reclassify sentiments tagged by our system. Just click the star next to any content item to reclassify it. You can now help us train and improve our sentiment engine!
All Time Campaign Stats
We’ve added a concise summary of a campaign’s all time stats to the main insights dashboard. You can now see how well your campaign has performed since the day you started it.
Date Range Shortcuts
We’ve added convenient date range shortcuts to the insights date range bar. You can now quickly switch between last week’s and today’s data.
And Much More
We’ve also some minor changes and squashed some bugs along the way. We look forward to your feedback.
You have probably seen the headlines about the massive numbers of ‘Likes’ served by Facebook. As a nod to the power of the simple Like button, other major social networks such as LinkedIn, Google, Blogs (Wordpress) and Communities have also started offering ‘Like’ functionality. With the social widgets, it is easy for any website to add the like buttons to their site. It has been reported that Facebook’s Open Graph and Like Button are going mobile as well. For Twitter, you have the more powerful retweet button which is closer to the share functionality. Although one wonders why Twitter has not made better use of the Favorite feature.
What is the driver for this massive adoption of ‘Like’? It goes back to the concept of ’single tap engagement’. Most of the users on social media would love to interact in quick, simple ways – and Like is the embodiment of the simplicity. Here’s a video from FriendFeed founder Paul Buchheit who talks about why they introduced like feature.
As a simple to understand metric which tells you whether your content got approval or not, Like will become an increasingly important measure for marketers and will impact everyting from SEO to influence.
When you combine the engagement information with demographics, psychographics, location and other contextual data, it leads to new opportunities for businesses. This recent article from Wall Street Journal talks about the social-context ads which are powered by Likes from your network.
Facebook Inc. is trying to rev up its advertising business with a little help from your friends.
The so-called social-context ads, which Facebook started rolling out over a year ago, are based on data it collects on the likes and friends of its users. The ads appear on the right side of a user’s homepage, with an image and headline from the advertiser. With the ads are the names of any of the user’s friends who have clicked on a button indicating they like the brand or ad. The user is also offered a chance to indicate he likes the ad.
“Marketers have always known that the best way to sell something is to get your friends to sell it,” says Sheryl Sandberg, Facebook’s chief operating officer. “That is what people do all day on Facebook. We enable effective word-of-mouth advertising at scale for the first time.”
Its hard to argue about the scale but there are some other views on this topic. First, as everyone knows Facebook has been criticized for its privacy policies. As these ads become common, more people will notice them. On the left is one example from my Twitter timeline today. Next time you ‘like’ a brand’s page, think about this.
I found this line from an article on PC World interesting:
Facebook’s “Like” button program told Pandora my musical preferences even though I didn’t ask it to. It was both cool and creepy.
There are some interesting cases as well. For example, you have to like a Facebook fan page in order to post a complaint on the wall of that page. So it means that if you are upset about a brand, you first ‘like’ them and then post a negative comment. I have done that.
What about you? Have you made good use of the Like feature?
I was at a social media marketing conference in San Francisco earlier this week. It is always good to meet and talk with with a diverse group of practitioners of social media. The conference was packed with attendees and sessions ran all day and provided views and information about a wide range of social media topics. One of the key topic was engagement and evangelism for brands. What is it, how do you plan for it and what are the best practices and tools. Here are a few points.
Trends
- @Britopian shared the observation that more brands are listening, few are engaging (see the presentation below)
- New demographics: Augie Ray shared Forrester’s segmentation of online influence: mass influencers and social broadcasters
- More ways to share content – widgets, social footer bars eg Meebo bar
Key Points by Thought Leaders
- Creating Evangelists via social media: what is the difference between loyalist, advocate and evangelist.
- Authenticity in social media is over-rated. Be honest.
- Value of a FB fan is zero unless you act on that relationship and add value to it
- Just 16% of users account for 80% of online brand advocacy
- @AaronStrout emphasized that brands need to give consumers reasons to share
- @AaronStrout mentioned that there’s a scientific process to creating an advocacy program
- Identify your brand supporters and pay attention to them
- How should brands build communities? @ChrisHeuer had a simple advice: brands should participate in 3rd party communities before creating their own.
Overall, great discussions and lots of take aways for the MutualMind team to build upon!
For more details see the conference blog post.
Why is it so hard for the social media “experts” to agree on a definition of engagement? With any major change, such range of opinion is natural. The challenge is to create broad understanding and then drive adoption which will lead to accepted standards.
The Forrester report clearly admits that it is very difficult for organizations to gather all metrics. For social media analysis tools, it is important to articulate the assumptions for calculating engagement. Most of the times, a few select metrics will be used to provide a view of engagement on social media. I will elaborate further on this in my next post.
Here are 3 point of views on Engagement.
- From 2007 – 2009, Forrester presented a series of reports on engagement in which a big picture view of engagement (not limited to social media alone) was presented.
- In 2009 Altimeter Group created Engagement DB report which “ranks the world’s most valuable brands based on how they leverage social media to interact with customers.”
- In 2010, Engage, a book by Brian Solis was published. In this book, Brian talks in details about his views on engagement and invents the mantra of “Engage or Die”.
Let’s start with Forrester’s definition:
Engagement is the level of involvement, interaction, intimacy, and influence an individual has with a brand over time.
Four components make up the engagement framework — the four I’s:
· Involvement — the presence of a person at the various brand touchpoints. Metrics include Web site visitors, time spent per page, physical store visits, impressions from mass media advertising, etc. Data sources include Web Analytics, store traffic reports, etc.
· Interaction — the actions people take while present at those touchpoints. Metrics include click-throughs, online transactions, in-store purchases, uploaded photos or videos, etc. Data sources include eCommerce platforms, POS systems, social media platforms, etc.
· Intimacy — the affection or aversion a person holds for a brand. Metrics include sentiment measurement in blog posts, blog comments, discussion forums, customer service call sentiment etc. Data sources include brand monitoring services, survey responses, customer service call centers, etc.
· Influence — the likelihood a person is to advocate on behalf of the brand. Metrics include brand awareness, loyalty, affinity, repurchases, Net Promoter, satisfaction ratings, forwarded content, etc. Data sources include market research services, brand monitoring, customer service call centers, surveys, etc.
Brian Solis writes in a blog post about the importance of engagement and how it ties to the consumer’s decision process. And this is exactly why engagement as a concept is elusive: it spans across various business processes, some of which are hard to capture and quantify.
One of the most sought after answers in Social Media is whether or not engagement in social networks such as Twitter or Facebook directly correlates to customer acquisition, retention, and advocacy. Before we can earn customers however, we have to recognize that at any given time, there are also prospects. And, prospects require information and confidence in order to make decisions, in your favor of course. The answer to our question lies in social engagement.
Prospects are not only searching for guidance, comparisons, and experiences through Google, they are also becoming increasingly social in every step of a decision making process. If brands do not identify the various stages of choice and resolution and also the networks where they socialize and explore, opportunities will be missed.
If we’re not part of the decision making cycle, we are absent from decisions.
In the next post on engagement, I will delve deeper into some of the interactions metrics which are very good indicators of engagement.
In DFW, we are lucky to have a dedicated group of social media professionals and enthusiasts: The Social Media Club of Dallas. Last week SMC Dallas had invited Eric Qualman, author of Socialnomics and producer of popular videos on social media, for a talk with the group. Eric did an interesting presentation (with cartoons and music) and shared his experience, including the reason why he created the videos (CEOs were not getting the message even after 1 hour of talk) and his views on how social media will evolve.
The topic of social media ROI keeps coming up in most discussions. This video from Qualman is a good reference to keep and share with others. At MutualMind we take a simple approach for ROI.
- Include measurement in your plans before you start a social media initiative.
- Measure quantitative results (that’s what we specialize in) for Reach, Buzz and Engagement.
- Understand and recognize the value of qualitative benefits (brand reputation, crisis avoidance etc).
We are pleased to announce the release of MutualMind 1.3.0. Here are some of the highlights from this release:
Facebook and Twitter Authentication Changes
Due to changes with both the Facebook and Twitter APIs, we had to re-implement how our application interacts with both of these services. As a result, the credentials for Facebook and Twitter accounts you had previously added will no longer work.
You will need to authorize your accounts one more time in order for the application to continue collecting and analyzing your social media accounts. We apologize for the inconvenience this may cause you. We highly encourage that you re-authorize your accounts as soon as you can to prevent any loss of data.
Go to “Admin” and then click “Brand Facebook Accounts” or “Brand Twitter Accounts” to get started.
More Detailed Insights
We’ve added a bunch of new views for various metrics like buzz, engagement and more.
Custom Date Ranges
All the insights views now support custom date ranges. Just select a start and end date and you’ll be able to visualize your data in interesting ways.
More Facebook Stats
As an added bonus to the changes with the Facebook API, we are now able to bring you even more detailed Facebook stats and insights with more are on the way in upcoming releases.
And Much More
On top of the major changes mentioned above, we’ve also made a ton of minor changes and bug fixes all around. We’re planning some really cool features for the next release. Stay tuned. As always, we look forward to and really appreciate your feedback!
We added support for Facebook fan pages to our MutualMind platform a few weeks ago. With this addition, you get all the rich monitoring, analytics/insights, response management features you’ve grown to know and love, but for your Facebook pages!
Setting up Facebook support in MM is really easy. Just follow this guide to get started.
Step 1: Add a Facebook Account
Click “Admin” from the top menu then click “Brand Facebook Accounts”.
Once on this page, click “Add Brand Facebook Account”. You will be transferred to Facebook for authorization.
Once you grant the application the requested permissions to manage your Facebook resources, you will be transferred back and your account will have been added.
Step 2: Select Fan Pages
After your account has been added, you can click the link labeled “Manage Pages” beside your account to select the pages you wish to mainly interact with from the application.
Check the ones you are interested in and then click the “Update Selections” button. The pages you select here will limit the page lists in the following features:
- Facebook stream browser
- Facebook component set up in the campaign editor
Notice the button labeled “Refresh Page List” at the top. If you happen to ‘like’ any new Facebook pages, you can come back here and click that to keep the page list in MM in sync with your actual list.
Step 3: Add a Page to Your Campaign
Now that you’ve added your account and selected your pages, you can now go to the campaign manager by clicking “Admin” and then “Campaigns”.
Click on the campaign you want to modify. Once in the campaign editor, click on the link labeled “Add Facebook Page” for either the “Brand Intelligence” or the “Competitive Intelligence” column.
Select one or more pages from the drop down list and click “Done”.
Voila. You have successfully set up Facebook page monitoring for your campaign!
eMarketer posted an article today with the title Social Media No. 1 Emerging Channel for Lead Gen. It is a common practice among business-to-business salespeople to use LinkedIn (and more recently, Twitter) for prospecting. No doubt that social media presents huge potential for lead generation. However the rate of success and the value derived from social media for lead generation is not yet clear enough for most people. I believe that the success of using social media for lead generation clearly depends on the process, discipline and quality of tools that you use for managing and organizing social media information.
Let’s look at the market and trends. Forrester reports that B2B interactive marketing spending to hit $4.8 billion in 2014, over double an estimated $2.3 billion in 2009. Forrester analyst Michael Greene mentions that:
B2B marketers can’t ignore social media. Consumer-focused marketers have been relatively fast out of the gate in adopting social marketing, but I believe that it’s B2B marketers who will ultimately gain the most from social technologies. B2B marketers have always understood the need to develop deeper relationships with customers, and while traditional sales and client service operations aren’t going to disappear, social media has emerged as key platform for informing decision makers during the sales process and engaging and supporting existing clients.
Lead generation vendor landscape is fragmented with social media as one part of the bigger picture. The diagram from Forrester Research aptly describes the situation.
Here is more from the e-Marketer post:
According to virtual events provider Unisfair, social media is the top emerging channel for lead gen among technology marketers surveyed in May 2010.
Lead generation was marketers’ first priority, with 66% saying it was their greatest concern for 2010, compared with just 17% who chose brand awareness.
When online marketing firm R2integrated surveyed US marketing professionals in April 2010 about why they had a social media strategy, the No.1 response was to increase lead generation.
Unqualified leads are a major problem, with one-third of Unisfair respondents saying they normally rejected 16% to 30% of their leads as unqualified.
The most important types of information for determining whether a lead is valuable include interest and demographic information—which might make social media an even more useful prospecting channel, since profiles can include exactly the type of data salespeople are looking for. In addition, social media can provide a venue for nurturing leads that are not yet ready to buy.
What about you: do you use social media for lead generation? What are some of the lessons that you have learned from your efforts to generate leads from social media?