In this article you will find out…
- How text analysis is used today
- How we make it actionable
- What opportunities are available
Text analysis is a field that is developing very quickly and has been doing so for the last few years. But how can you use text analysis technology yourself to actually achieve real results in everyday life? How far has text analysis technology come and what is the next step in maturity?
Text analysis today
Instead of building a text analysis service ourselves or using a single provider, we have opened up the use of text analysis from all common suppliers. This means that you can choose, for example Gavagai who are good at text analysis in Swedish and let you build your own models, Google which have ready-made models for the care area in English, IBM's breadth with different languages or Microsoft's Azure- services.
Making text feedback actionable
Quicksearch's main goal is to help its customers act based on the customer feedback they receive. Without doing something practical with the feedback that comes in, it is not really worth anything or at best a curiosity. We have had this approach also when it comes to text analysis, which is our latest addition.
Based on text feedback, we identify which customers show satisfaction, loyalty or dissatisfaction. This becomes both a segmentable overview where you can follow trends in what works and what doesn't work, but we also flag the individual customers who say they have problems with a certain product, function or process and also have tools to be able to actively help them with solving their customers' problems.
- Customer service teams or KAM receive information about which customers have been flagged for needing action. Who receives the information is tied to the type of problem the customer has expressed.
- Product owners can track which issues are important to their products and how these issues affect the product positively or negatively.
- The company as a whole can follow trends among its customers in order to actively work to increase customer loyalty, reduce churn or increase repurchases.
- Managers and management gain deeper insight into which issues are working or not working in the business and what needs to be done to increase employee engagement.
In this way, it's very similar to what Quicksearch always does; but what new possibilities come with text analysis?
- Through text analysis, you can quantify even feedback that was not explicitly asked for. With surveys, it is very easy to follow what has been asked about and also capture things that were missed to ask about through text comments, but it can be difficult to quantify these. Both in which subjects it is concerned or whether it is positive or negative.
- It is easy to find the answers from the customers or employees who bring up certain topics in order to get a better overall picture of what they bring up. Are there customers with certain products, pricing or only certain segments or only dissatisfied customers who raise certain questions.
We can make all feedback actionable and create the same insights as we can based on surveys without the feedback coming via a survey.
A hammer or a spade?
All new technology is easily rushed before there are clear cases where it does not create the right benefit. You don't know whether the new tool is a hammer or a spade and try it on a little bit of everything.
We have seen a number of different ways of using text analysis with questionable results. It presents data, but does not lead to any insight or action. Sometimes it becomes a bit of a gimmick. But those cases are quite easy to assess if you ask yourself the question: How can I achieve something with the help of this?
As always, the result is no better than what you put in and it's easy to be too generic. They usually say that content is king, but here context is king.
- Text analysis of general texts or comments provides general insights. Whereas feedback compiled in a particular context provides deeper insight into the context. Are there answers to specific questions, for example, the insight comes in a clear context and questions that elicit deeper answers. Here, surveys or chat dialogues are a very good way to gather deeper answers that lead to better insight.
- Text analysis does not get better than the text analysis model. Behind each text analysis is a model that has been trained to understand the text content. In addition, the more specific the model, the better assessment can be made. There is both the possibility of text analysis that is general for Swedish, but a model that is trained on the context and can understand differences in words and context. We expect that there will be ready-made models for individual industries and also company-specific trained models.
Quicksearch has therefore chosen a unique approach
With the goal of being both first and best at the same time.
- It is unclear which of the major suppliers will be the first with the Nordic languages.
- The future will determine which of them will have the best text analysis.
- There are suppliers today that allow you to use your own subject-specific models that allow high precision in any language.
Quicksearch supports all of these and you can switch between them.
In addition, we have the right model to collect feedback with the right context that makes the feedback insightful and actionable.
It's just a matter of getting started. What are you waiting for?
