Adaptive big data maturity model using latent class analysis for small and medium businesses in Thailand

Abstract

Big data analytics (BDA) is widely adopted in large enterprises. However, very few small- and medium-sized enterprises (SMEs) have adopted BDA because they lack the relevant knowledge, which makes BDA development expensive and unsuitable. A big data maturity model (BDMM) is a tool for assessing the stage for using big data in a company, and it acts as a guide for improvement. However, most BDMMs are designed for large enterprises using rule-based scoring, which is static over time. Developing a suitable BDMM for SMEs is a challenging task for professionals in terms of acquiring small-scale expertise owing to the lack of case studies for verifying the maturity level. This study proposes a new BDMM for Thai SMEs and a new methodology for developing a dynamic model using latent class analysis (LCA), which explains the behaviour of each latent class and provides non-rule-based scoring. We define four types of capabilities in SMEs: organizational and attitude factors, information technology, technology, and people readiness. Data are collected from 135 SMEs in Thailand. We introduce a methodology for developing multiple building stages of the BDMM. Further, we experiment with several clusters suitable for SMEs using statistic-based and data visualization approaches. The proposed BDMM is validated via a secondary evaluation of 11 firms, nine months after the initial evaluation. Further, we introduce a web-based application for respondents to obtain their firm’s assessment results. The visualization-based result helps the respondents compare their business with other companies at the same or higher maturity level. In summary, SMEs can use the proposed BDMM to plan for continuous self-improvement and thus optimize their business using BDA to maximize its value to the organization.