Artificial Intelligence | News | Insights | AiThority
[bsfp-cryptocurrency style=”widget-18″ align=”marquee” columns=”6″ coins=”selected” coins-count=”6″ coins-selected=”BTC,ETH,XRP,LTC,EOS,ADA,XLM,NEO,LTC,EOS,XEM,DASH,USDT,BNB,QTUM,XVG,ONT,ZEC,STEEM” currency=”USD” title=”Cryptocurrency Widget” show_title=”0″ icon=”” scheme=”light” bs-show-desktop=”1″ bs-show-tablet=”1″ bs-show-phone=”1″ custom-css-class=”” custom-id=”” css=”.vc_custom_1523079266073{margin-bottom: 0px !important;padding-top: 0px !important;padding-bottom: 0px !important;}”]

Accessing the Full Power of Text Data as Part of Your Insights Framework

Most companies have huge amounts of text data and, for the most part, it is giving them little to no information. Free, or unstructured, text data is rampant across business ecosystems, containing vital insights from things like social media chatter, product reviews and customer feedback. Mining and analyzing these hundreds of millions of data points can be difficult, to say the least. It’s messy. It’s huge. And, traditional processes can’t handle it.

In a world that demands speed above all else, manual processes to analyze text data are painstakingly slow, with humans only able to process about 3-5 entries (an entry being, for example, an open-end survey response or product review) per minute at best. They bring their own bias into the process as well. There are some hybrid approaches such as rule-based systems, where a person creates rules and tells the system what to find. While AI then applies these rules, it is still slow because of manual intervention.

Top AI ML Stories: Role of AI in Helping B2B companies that are Missing Out on Buyer Intent Data

Plus, bias can remain because of the rule creation process. There are a few other approaches, such as supervised categorization and topical modeling, that can be flexible and scalable, but the manual pieces of the process can cause downstream issues.

In order to take full advantage of this rich data, we must thoughtfully implement AI, machine learning and natural language processing algorithms to help us make sense of it to its fullest potential. Of course, as I have long maintained, there must be a deft balance between the use of technology and human skills, and that is no different when it comes to text data processing and analysis. AI can build themes as it ingests the data, organize it and then, with the right system, an expert can apply context and human understanding to the organized data – quickly and efficiently. When businesses find a way to truly utilize text data and turn it into hard facts, plus make connections with other data sources like sales or survey data, they can start to access true audience understanding.

Text data example: Customer reviews

Let’s take a closer look at one vital source of text data – customer reviews. Review data can help companies fully understand how their audiences make decisions and how these individuals feel about their products. Secondary research has shown us how important product review data is, with about nine out of 10 people going online to get information for a purchase decision, and two out of three telling us that product reviews are their number one decision influencer.

Related Posts
1 of 832

Recommended: Meet DailyTalk: The Latest Conversational Text-to-Speech Dataset based on FastSpeech Framework

Think about how this might play out in your own life.

Let’s say you’re looking at AirBnB for example: you find the perfect place, the photos look great, the location is prime, and the price is fair. But, if the reviews are negative, or there are no reviews at all, you will likely move on quickly. In fact, any time you are making a major purchase decision, the likelihood is that you will probably at least glance at other people’s reviews. This makes the kind of data held in product review text incredibly valuable to companies (or AirBnB hosts for that matter).

Recently, Alyona Medelyan from Thematic and I took a closer look at how to best access and use this type of data for real insights. We presented a case study about how we used AI-driven sentiment analysis to help the world’s largest soft drink manufacturer tap into the massive amount of data it had from customer reviews.

The company’s insights team went from manually coding using a system that could only handle a limited amount of data rows to using well-designed AI text analysis to tap into hundreds of thousands of rows of data, representing consumer comments about its products, coming from all around the Internet – Walmart, Amazon, Google and more.

Upon accessing this data and using AI-driven text analysis that fed into an advanced data integration, analysis and reporting platform, the insights team was able to immediately investigate the relationship between data points (i.e. both the respondent/review and comment/sentence level) avoiding traditional, manual-heavy data preparation steps. They could look at multiple themes and sentiments per person, easily, and bring much greater depth to the text data than could be accomplished through traditional methods. The findings were then combined with other data streams, and provided invaluable insights for business decision-making, communications and even product development.

This is just one example of the power of text analytics for consumer insights. The importance of knowing what people are saying about your company or your products is enormous, but it can be difficult to unlock. The insights held within text data can help you understand the nuances inside your consumer audiences, and if it is done right, it can be fast, efficient, and unbiased. This data can advise things like allocating spending to where it is most needed to directly address customer needs and concerns, which can improve overall KPIs and metrics that are being tracked by the company. Understanding the “why” behind the “what” will provide richer context and boost the overall value of your consumer insights program.

[To share your insights with us, please write to sghosh@martechseries.com]

Comments are closed.