‘Text’ and ‘data’ are words that you rarely hear together. As a phrase, ‘text data’ is almost an oxymoron, combining the artistic license of language with the scientific rigour of data collection. This means that researchers, agencies and businesses can sometimes be blind to the opportunities, or even existence, of it.
But text data – also called qualitative data, unstructured data or, in the simplest of terms, words – can reveal the ‘why’ behind it all. With the right text analysis tool you can build contextual understandings of your customers and target audiences’ experiences and expectations.
In short, text data has the potential to be an insight gold mine. And chances are, you could be sitting on one right now…
Common sources of text data are hiding in plain sight
Words are everywhere, meaning that many businesses are sitting on a wide range of text data types. But some of the most common places to find text datasets are in surveys, social media, online reviews, customer service transcripts, and marketing content.
Let’s walk through each of these five text data examples in a bit more detail.
1. The hidden gem of open-end surveys
Surveys are the world’s entry point into market research. Many businesses use customer feedback questionnaires and market research surveys as a matter of course – and often, these are dominated by a series of multiple-choice questions, perhaps with a token one or two open-end questions thrown in.
While these open-ends are sometimes left unanalysed, they can yield thousands upon thousands of valuable words. These open-end responses are one of the most insight-rich types of text data going. It is text data gold, just waiting to be mined.
2. The giant text data pool that is social media
Social media is a reality of day-to-day life – for many, it’s the lens through which lives are conducted and viewed. So it’s quite the shift in perspective to see social media instead as the world’s largest pool of text data.
On social platforms, people express their thoughts, feelings, and actions through words and emojis in social posts. Twitter, for example, can be a place for public opinion in its most organic form. These social posts are enormously valuable for understanding and resonating with target audiences, and can be exported simply from social listening tools like Brandwatch.
3. Invaluable online reviews
We’ve all used them. Whether in our personal or professional lives, review platforms like TripAdvisor, TrustPilot and Amazon are a regular feature in the way purchase decisions are made. They help us take a quick temperature check on how brands, products, and experiences are rated among their audience.
But reviews can provide us with much more than just a fleeting temperature check. These review platforms harbour thousands of pages of customer experiences and feelings – all expressed in words. Plus, they are also almost always attached to a quantifiable rating (e.g. four stars) that can be used to segment responses and reveal drivers of customer satisfaction (or dissatisfaction…). By collecting the information in online reviews, we get text data, which can give invaluable insight into our own and competitors’ brands.
4. Emotionally charged customer service transcripts
When the boiler breaks or you’re wondering where that parcel is, a customer service agent or chatbot will often be the first port of call. But these seemingly innocuous exchanges can be a great source of text data. The opinions a customer might express to a chat agent or the information that features in the conversation, for example, could reveal priceless (and often emotionally charged) insights into what your customer wants, needs, and feels.
Even if words are spoken orally, they can be transcribed (using transcription software such as rev.com) to become valuable text data. A half-hour phone interview, as an example, can yield over 3,000 words of valuable insight.
5. The treasury of easily accessible marketing content
The internet is a veritable ocean of marketing content. Think website copy, blogs, landing pages, and the near-infinite other types of collateral and content. This is a huge body of easy-to-access, published text data that is just itching to be analysed.
Supposedly, there are more than 600 million blogs on the internet (and growing), with millions of new posts going live every day. At the same time, an average FTSE 100 company webpage contains between 2,500 words. These giant pools of data are worth their weight in gold, and exceptionally useful for competitor analysis and market intel.
Other text data examples to consider…
While we’ve covered five of the most common text data examples, there are many more hiding in plain sight. Think forums (such as Reddit), media coverage, focus groups, interviews, academic research papers, speech transcripts… the list goes on.
But there’s one big question still to be answered…
The golden question: how do you mine text data?
Where there are words, there is text data – and where there is text data, there are insights. But how do we extract these? It all comes down to text mining. This is the process of transforming unstructured text (like the examples above) into tangible insights. These text analysis insights can take the form of key concepts, emerging trends, hidden relationships, customer sentiment, and more.