‘Free form’ text analysis can strike fear into even the most customer-battered, nail-biting researcher. But you can learn how to get actionable information from free form text without doing the charleston with an orange octopus.
And; have fun understanding customers with this easy expert technique.
Learning While Killing Lovers
There’s a lot of lovers on Twitter and I don’t mean the dodgy bio picture: zoomed in, up close and fleshy.
Working with Irish Twitter bio profiles, the free form text where you, and your Twitter prospects wax lyrically about yourselves, it seemed there were a lot of ‘lovers’. And I was killing them.
Killing Lovers Today. Cleaning Up Marketing Lists For #PaidSocial Coffee Lovers, Food Lovers, Dogs, Chocolate, Thou Shalt Leave The List
— Jane E Morgan @JEM 9 (@Jane_E_Morgan) May 26, 2015
But how did this emerge? And what story could these lovers tell?
Emerging: that’s a characteristic of good research. You have a question about your customers. But another different foggy shape floats into view. You sense “something”, maybe.
How to join the dots. How to add depth and colour. How to enrich your understand. How to see your customers clearly.
How do you get a better handle on free form text information without tango’ing the orange monster?
Fighting The Fear Of Free Form Text
Free form data is a sticky, entangled octopus that strikes fear into the soul of market researchers.
Professional researchers know that free form text is a fantastic invisibility cloak. It starts off looking verbose AND boring. It leaves you feeling limp, breathless and mushy.
Clean, clear multiple choice surveys are simple. But nothing emerges from multiple choice surveys. They are only suitable for confirming your understanding; daintly placing your customers into neat little boxes.
Free form text holds the key to many riches. It’s the sticky and muddy, rough and tumble of real customer understanding.
And understanding your customers is the crown jewel of riches, the crown jewel of marketing, the crown jewel of reaching customers.
So sidestep the charleston with that orange monster. Instead use this easy and fun method to work your free form text. (And help me balance the bad karma of killing lovers.)
But before we reveal, write down what YOU think Irish Twitter people love. (No cheating now.)
Got it? Let’s go.
Lifting The Fog On Free Form Text Analysis
I started with a reasonably large data set; 8000+ Twitter bios.
Keep the original context of your data set in mind. For example, geographic searches in Twitter are not absolute. You might say “Cork”. Art might say “County Cork”. And Muriel calls it “the rebel county“.
You will also want to keep a copy of the original data set unedited for reference.
- In Excel, use the function to create a new copy -> Home / Format / Move Or Copy.
- In Microsoft Word do a ‘file’, ‘save as’.
- Google Drive allows you to ‘Make a copy’.
Now you have a file that shall give forth it’s secrets. The secret sauce is tag cloud generators.
Word clouds are a great way of working with, and clearly presenting large volumes of short text. The more frequently a word appears, the larger that word in the tag cloud. Tag cloud generators available for free online include:
Choose a tag cloud generator. I use Wordle. Copy and paste your free form text into the tag cloud generator.
The foggy concept was big and bold immediately.
Step 2. Combine similar terms In this case ‘love’ and ‘lovers’ using the “replace all” function. Wordle provides an option to automatically remove common words such as ‘of’ or “an” speeding up the process. This is what the second version looked like.
The next step was to replace ‘lover’.
Your jewels will start to appear as you refine the word cloud by:
- using ‘replace’
- pasting into the tag cloud generator
Typically I might repeat this step maybe 5 times removing the largest word each time. If you’re looking for just the largest diamonds, TagCrowd allows you to limit the number of words shown.
Dealing With Surprises
‘Drinker’ appeared in the word cloud. Using the original data set revealed non-alcoholic drinks, like “tea drinker”, as well as wine. Use steps 1, 2 and 3 to better understand surprises.
You might decide you want to see the ‘tea drinkers’ compared to ‘coffee drinkers’. To present compound words use a hyphen. Turns “tea drinkers” into “tea-drinkers”.
- Use the find function for: ” drinkers ” (note the space before and after the word)
- And replace with “-drinkers ” (hyphen before and space afterwards)
Presenting Free Form Data
You’ll might do a little work removing noise and distractions. A lot of people said they were a mother/father/student. These were smaller than the main ideas but large enough to distract. (Killing family relations now!)
‘Daughter’ still appears in the final version because it didn’t distract from or conflicting with the overall story I wanted to tell. (The mammies and daddies may be disapointed that ‘son’ did not appear.)
Most tag cloud generators allow you to customize how the tag cloud appears changing font, font colors (colours) and more. You can easily end up with a tag cloud aligned to your brand.
Irish Twitter Love
As I worked the data the lovers were clear to see. Many of us are moved by music; food for the soul. Which brings us to food; coffee, tea, wine, and chocolate.
— Brian G (@brjg1969) January 12, 2015
So here’s a celebration of our passions.
Did you correctly identify the greatest love of Irish Twitter users? What jewels of customer understanding will you find hidden in your free form data?
My Twitter bio used to say: “Marketing Love Is: Creative Ideas and Data Analytics”.
And what kind of lover are you? Put it in your Twitter bio. Make the world a warmer place.