Optimising Tool Selection for Emerging Businesses - A Sentiment Analysis-Based Approach
Selecting suitable software is challenging for small and medium-sized enterprises (SMEs) and start-ups. Many companies lack the resources and expertise to make informed decisions when selecting software tools in a complex market environment. This was also the case for BORTOLI AG, a Startup company that encountered significant problems selecting and implementing software tools. The wrong choice of software tools had financial implications and adverse effects on employee motivation and productivity. Against this background, this bachelor thesis aimed to develop a data-based approach to optimise software selection for SMEs and start-ups. An objective picture of user satisfaction and specific needs was to be drawn by applying sentiment analysis to user reviews and feedback.
The work combined web scraping, data cleaning, and sentiment analysis. User reviews were systematically collected, cleaned, and analysed to obtain an objective picture of satisfaction. The SentiWS lexicon, the SYUZHET package with the NRC emotion lexicon, and a custom lexicon were used for the sentiment analysis.
The analysis revealed that specific software tools were rated highly in their respective categories. Pipedrive for Customer Relationship Management (CRM), Miro for Collaboration, Movable Ink for Content Creation, Storyblok for Content Management Systems (CMS), Haufe HR Services for HR Tools, Mailchimp for Marketing Tools, Trello for Project Management, Swat.io for Social Media and Hotjar for Web Analytics were recommended. These recommendations are based on detailed sentiment analyses and the identified user needs.
The research aimed to answer the question: "Can sentiment analysis of user reviews and feedback enhance tool selection for emerging companies?". The findings indicated that sentiment analysis can indeed enhance tool selection by offering valuable insights into user satisfaction and specific needs. This, in turn, enables more informed and targeted decisions when choosing software tools. The next steps would include testing the recommended software, adapting the scraping code to possible changes, and expanding the custom lexicon. These measures are intended to optimise the selection process's efficiency and accuracy. Companies can improve their software decisions through this data-driven approach, thereby increasing efficiency and satisfaction within the organisation.