I made my New Year’s resolutions early this year and am trying to actually KEEP them. While I don’t think I’ll do much on my resolution to lose 35 lbs before my daughter’s wedding until AFTER the holidays, I’m keeping another today by posting the promised follow-up to Friday’s post on building understanding from text analytics.
If you didn’t read Friday’s post (shame on you), a quick recap:
- Businesses thrilled with FINALLY having some data coming out of their marketing communications, feel they have to analyze every bit of this massive amount of data
- Machine analysis, which is the only way to build understanding from the massive amounts of text data available on social networks, is a dismal failure because of vocabulary, semantics, and data collection problems.
- The only thing scarier than having NO data is having INCOMPLETE or INACCURATE data.
Building understanding from text analytics
So, lets take a look at another approach to building understanding from text analytics. First, I’ll start with some facts about doing better text analytics:
- You have to start at the end. What answers do you need? What are your goals? Then, figure out how to get those answers. If you need insights to build a better concept or hone in on specific concerns in creating your marketing communications, then use a text analysis designed to answer these questions.
- You don’t need to analyze EVERY utterance to take the temperature of your audience. Take a look at the polls surrounding the presidential election. Pollsters ask about 300 Americans — a very tiny sample of the over 300 million Americans — and accurately predict the winner with a high degree of accuracy. If something is wrong with your product, it’s pretty easy to find on Twitter or Facebook. You don’t need to exhaustively examine every utterance made on these social networks. If people have something to say, they say it pretty loudly. You’ll hear them!
- Using human analysts trained in ethnographic techniques can uncover insights to help build understanding from your text analytics.
I presented 2 examples of using text analytics to build understanding. The first came from an analysis of Disney bulletin boards. By collecting a few thousand pages of text about Disney parks from various digital boards and analyzing it using ethnographic techniques, I was able to provide understanding Disney could use to improve their market performance.
The first thing I learned was what visitors found important factors for staying in the Disney Resorts. They didn’t talk about special amenities, such as early admission to the parks, or that the Resorts were located near the theme parks. They mentioned how important it was that the property be “well themed” and close to the shuttle buses taking you to the parks. Understanding what’s important in making the buyer decision, provides guidance for Disney when creating or modifying Resorts to emphasis these elements as they have the most impact on decisions to stay at the property. Disney can also use these insights in advertising the properties by emphasizing desirable features rather than other features that are less impactful on the decision.
Microsoft talked about building understanding from text analysis surrounding introduction of their Surface computer. Within just a few days, Microsoft had the verdict on the tablets and got engineers working on modification for the next version based on what consumers liked and didn’t like about the tablet.
Samsung hired an agency to build understanding using text data from a variety of sources, including social networks, blogs, and review sites. They were especially interested in competing with the photo capabilities of smartphones. Hence, they listened for discussions of cameras, not their brand.
What they found was a link between cameras and vacations (among other insights). This told them to focus their camera advertising on travel sites. They also used insights to provide a better user experience around Black Friday, where they identified 3 groups of shoppers. They provided solutions specifically designed for these user segments through advertising and on their website. Understandings from text analysis increased Black Friday sales significantly.
So, while it’s nice to have a huge data set and it’s comforting to collapse this data into easily digested figures, such as sentiment, that can be tracked over time, it may not be as meaningful or provide insights able to guide future actions in the way possible by building understanding from text analytics.
Whether you need a complete social media marketing strategy or some consulting to optimize your existing social media marketing, we can fill your digital marketing funnel or create your brand — online and off. Our analysts can build understanding from text analytics and show you how to optimize you communication or create products consumers will love. We can help you do your own social media marketing better or do it for you with our community managers, strategists, and account executives. You can request a FREE introductory meeting or sign up for my email newsletter to learn more about social media marketing.