- November 27, 2020
- Posted by: teX.ai
- Category: Text Analytics
A large e-commerce website specializing in baby and kids’ products was having a terrific quarter, with festive season sales growing robustly. The company’s customer experience team was busy charting out plans to enhance overall customer experience and take advantage of the increased traffic coming in. There was tremendous focus on brand partnerships and catalogue management. Incidentally, over 70% of the company’s traffic came from mobile devices. But what also went unnoticed was that the conversion rate (of completing a purchase) was way higher for the web app.
What prompted the customer service team to analyze this more carefully?
A single negative statement in what was actually a positive customer review with a 5-star rating. It mentioned something about “bad mobile experience, I prefer to use my laptop for this site.” Everything else in the review was positive.
You might want to read: Voice Of Customer Analysis For The World`s Largest Sports Goods Retailer
Thankfully, this e-commerce player was using text analytics to capture feedback from customers, across several touch points including online customer reviews, social media sites and customer service calls.
The point is – in today’s world of data and information overload – it’s possible to miss capturing “important information” made in “passing”. The role of next-generation text analytics is to uncover unique, useful insights from a wide range of text-based inputs. From a customer experience perspective, it is critical to understand and capture insights from both quantitative and qualitative information.
Going Beyond Quantitative Analytics
With text analytics, e-commerce sites can gain insights into their customer experience as well as enhance it tremendously. Text analytics tools such as Indium’s teX.ai can help them:
- Capture trends and patterns from social listening
- Use customer reviews to improve your product and service features
- Upsell or cross-sell based on what your customers are buying
- Create special promotions based on customer preferences to improve customer loyalty
- Reduce return rates by offering the exact product as desired by the customer
- Prevent frauds
- Improve website conversion rate
- Reduce losses due to warranty claims
The possibilities are simply endless. The key is to find the right tool with the right capabilities and the right partner with experience and the desired skill set.
Text Analytics vs Text Analytics
You read that right. There are tools that do text analytics, and then there are tools such as teX.ai from Indium, that do semantic analysis and sentiment analysis for a deeper understanding of online conversations.
The difference between the two can make a huge impact on your business.
Normally, text analytics refers to word meaning and inferring whether it is positive or negative feedback. This is limiting since in human communication, context is everything. Sarcasm, insinuations and innuendoes are part of the semantics that require the tool to go beyond “just the words” to be able to comprehend what the user is actually trying to convey.
teX.ai can be trained to do just that and understand more accurately the trends and patterns through semantic analysis. The tool extracts content from across the web and social media, not just text but also images and videos. This is then categorized as positive or negative based on three criteria, as required by you:
- A simple text analysis is made possible by identifying and classifying a small sample of words with their meanings. This is then submitted to the neural network for creating a model to train the system. Based on this, the model can classify sentences as positive or negative.
- In the second level, the root of the words is identified and they are paired before being classified. This enables capturing the underlying sentiments accurately. This is especially useful in segregating positive and negative views present in the same review. For instance, there could be appreciation for a product but criticism about the service. By capturing each of these sentiments, a clearer picture emerges enabling better decision-making.
- The third type of classification can be used for entities such as a personality or a corporate brand in a site such as Twitter or an e-commerce portal and extracting the views expressed. This gives a broader view where even the performance of competitors can be analyzed and the learnings incorporated into decision-making.
Why This Matters
The depth of training and deployment of text analytics can be a game changer in enhancing the customer experience on your e-commerce website. If you are able to discern their views and incorporate the changes, you will be able to provide them with the kind of experience that will turn them loyal and potentially turn them into ambassadors for your brand! The amplifying effect of social media can be to your advantage if you are able to provide a satisfying or even exceptional experience. The risk of not being able to do so is equally dramatic as even one bad review can have a negative impact.
Some of Indium’s customers have experienced tremendous benefits through teX.ai’s analytical capabilities. One online sports retailer was able to capture customer feedback on social media to improve specific product features.
In another instance, teX.ai enabled automated product categorization for an e-commerce aggregator. Since manual classification of thousands of products can be time consuming and prone to human error, text analytics accelerated the process and provided a more accurate classification based on description as well as images to match the categories better. While this was an advantage for the e-commerce site, no doubt, it also enhanced user experience in locating the right products intuitively.
By capturing user behavior on your site, teX.ai can also recommend the best product mix, promotional campaigns, pricing strategies and so on to reward loyal customers and enhance overall experience.
Multilingual: The feature is applicable for text extraction solution, while reviews could be summarized in English. Product categorization is available for English and Spanish. Having identified the language, it then uses the following steps to process the content:
- Breaking of sentences into parts
- Tagging parts of speech
- Parsing the syntax
- Chaining sentences
Local sentiment: teX.ai finds local sentiment rather than global sentiment. It will not tag entire review as positive or negative. It focuses on each sentence in the review and identifies positive and negative phrases. This ensures we capture all major opinions of a customer.
Multi-Format: Apart from text, teX.ai can even extract tabular data and peripheral (header) data from all formats of content also such as scanned or printed documents, websites, IoT devices, images and videos. It enables pre-processing with different image processing algorithms to improve the quality of images and scanned documents before extraction.
Dashboards and Analytics: It offers customizable, interactive dashboards and parameters to allow document management, monitoring processes and mitigate nuances to improve process efficiency.
Highly Secure: teX.ai is ISO, GDPR and OWASP compliant and incorporates industry best practices such as data encryption and permission-based access.
Platform-Independent: teX.ai can be deployed on cloud, on premise or as a hybrid system.
Quick Deployment: You can start reaping the benefits in just three weeks.
If you are an e-commerce player and want to design more targeted promotional strategies, identify and reward loyal customers, convert more visitors to customers, improve your product mix, product features, or up your service levels, we believe our text analytics product can empower you with the right insights.
To know more about teX.ai, schedule a demo with us