Value creation from analytics with limited data: a case study on the retailing of durable consumer goods

ABSTRACT

Companies are pinning high hopes on competitive advantages through data analytics. So far, value gains through analytics have been demonstrated for IT-heavy and data-rich business areas. Yet, research has paid little attention to value creation through data analytics in the plethora of companies with limited data (i.e. having transactions in the hundreds and attributes in the tens). Building on the literature of big data value creation and the resource-based view, we carried out an in-depth analytics case study with a retailer of renewable energy systems. Firms in this business area operate with expensive but few sales, so their available data are notoriously limited. Our findings demonstrate that data analytics capabilities and value creation mechanisms (democratise, contextualise, experiment with data, and execute data insights) are also effective in situations with limited data. Practice and research should therefore put not only emphasis on the volume and the variety of data but also on contextual factors related to managers (e.g. clear strategy, vision, leadership) and all employees (e.g. openness for agile working mode, data awareness).

KEYWORDS:

• data analytics
• information systems (IS) value creation
• resource theory
• value creation mechanisms
• machine learning(ML)
• retail
• renewable energy systems (RES)

Disclosure statement

No potential conflict of interest was reported by the author(s).

Correction Statement

This article has been republished with minor changes. These changes do not impact the academic content of the article.

Notes

1. On the one hand, the scientific discourse on big data is closely related to big data analytics (e.g. A Abbasi et al., 2016; George et al., 2014; Woerner & Wixom, 2015), sometimes the two concepts are even used interchangeably (Constantiou & Kallinikos, 2015; McAfee & Brynjolfsson, 2012). In any case, analytics is ‘a part of processing the data and one of the potential first steps in trying to realize value from big data’ (Günther et al., 2017, p. 191). On the other hand, the discourse on the use of analytics in business needs to talk about big data as a core resource for which new, more efficient analytical methods are needed (H Chen et al., 2012; Gillon et al., 2014).

2. Two conditions explain that limited data is sufficient for analytics in practice: (i) data should allow to be applied in appropriate algorithms (technical requirement) and that (ii) reasonable results can be expected from this (business requirement).

3. Tremblay et al. (2021) define theories in flux ‘as evidence-based inferences that emerge from analyzing large amounts of data or big data, often gathered from business processes and in partnership with practitioners. … A [theory in flux] generally takes shape when a pattern of a phenomenon emerges from the analysis of data.’

4. A sales lead refers to a potential customer for whom address or contact details are known.

5. For example, EnergySage (https://www.energysage.com/) and NREL (https://pvwatts.nrel.gov/) for the U.S., or Eturnity (https://eturnity.ch/en/home-eng/) for the European market.

6. Features are predictor variables for the ML models.

7. RF was used with mtry of 5 and 500 trees. SVM was used with a gaussian radial base kernel, sigma of 0.009 and cost of 1.

8. The numbers for the heat pump case with a selection of the top 20% are as follows: Sensitivity DV1: 25.5%, Precision DV2: 12.7%, Sensitivity DV2: 34.8%, Precision DV2: 48.9%.

9. Kuhn M (2020) caret: Classification and Regression Training. https://CRAN.R-project.org/package=caret

Additional information

Funding

We gratefully acknowledge this joint funding by the European Union, the Swiss Federal Office of Energy (grant number SI/501521-01), and the German Federal Ministry for Economic Affairs and Energy (grant number: 03050010).