Necessary cookies enable core functionality such as security, network management, and accessibility. You may disable these cookies by changing your browser settings, but it might affect our site functions.
Our in-house developed tools identified previously undetected shortcomings
Our client's company specializes in acoustic design and development to bring premium audio products at a reasonable price to customers worldwide. Our experts were requested to perform a marketing analysis to investigate why the client’s brand performance was lower in German and Danish markets than in other countries.
Our methodology approach included data crawling, natural language processing (NLP), data analysis and finally reporting of the results. Although the crawled reviews are not written in English but in Danish, Finnish, Norwegian, Swedish, and Italian, all the information has been translated using Apiron Technologies’ in-house tools.
When processing data via NLP, two methodologies were applied:
• Sentiment Analysis algorithms
Processed reviews and marked them as positive or negative.
• Meaning Extraction techniques
The main topic and subtopics of each and every customer review was segregated. Based on these results, the most important product technical specifications (features), product price range, and brand awareness were identified per product category and specific client base.
In summary, the following data analysis methodology was applied:
Data mining/ Data crawling
Searched the web and gathered data related to customer reviews and product specifications via Apiron Technologies in-house proprietary software.
Natural Language Processing (NLP)
Processed and analyzed via our machine learning (ML) based meaning extraction tools in order to isolate the product specifications that are at the top of the customers’ interests.
Isolated the top product specifications from the crawled data and based on the top selling products, the respective client’s products were positioned accordingly.
All these tasks have been achieved via an in-house data science and AI platform: Pythia.
Identified product categories based on the e-shop common classifications in order to conduct analysis for each category separately.
According to our findings, in both countries, the main competitors list is extended for a more comprehensive analysis. A feature analysis for competitive products shows product strengths and weaknesses and customer ratings.
Brand awareness analysis also ranks client’s brand versus their competition.
• Prices were relatively higher compared to the competition (except for sports earphones).
• The brand lacked awareness regarding those specific markets.
• In some cases, competitive and great product quality was significantly higher than the competition's.
• Product prices in Denmark seemed reasonable to customers.
Customer insights about smoothie blenders example
An example for blender products quality ratings of the brand and its competition in the Scandinavian market, based on 5.000 customer reviews and over 1800 product SKUs, was filtered down to top 50 SKUs for the category. For the analysis, quality ratings were crawled and a sentiment analysis was run to improve existing products for the brand.
Although the crawled reviews were in non-English languages, namely Danish, Finnish, Norwegian, Swedish, and Italian, all the information was translated using in-house tools.
The Research and Development department can improve an existing product or launch a new product, based on customers' reviews such as changing the bowl material from plastic to glass to meet customer expectations.
Product feature analysis for blenders example
The following example illustrates how a brand's marketing department can isolate the top product specifications from the crawled data and based on the top selling products per data source, position the respective blender products accordingly.
Major features were extracted using NLP and were ranked according to importance as shown in the bar plot.
Weak product features were revealed by analysing customers' positive and negative comments.
Information about major features were ranked according to significance and customer satisfaction was used as input for various brand departments to align marketing campaigns with features, kick off promotional campaigns or design new products to meet customer demand and market opportunities.