Work

Breaking the data barrier: Enabling digital commerce capabilities in SMEs under business-model-driven data scarcity

Year: 2026

Type: article

Abstract: Small and medium-sized enterprises (SMEs) rarely adopt the search and recommendation capabilities that large online retailers use. The obstacle, we argue, is a mismatch between SME business models and the data these capabilities assume. Many small retailers require no registration and collect no ratings or reviews; user-level data never accumulate. We call this business-model-driven data scarcity, distinct from transitional cold start. Following design science research, we built and deployed an artifact delivering search and recommendations from first-party behavioural events and catalogue data alone, across heterogeneous platforms. Four Swedish SME retailers ran it; evaluation from July to November 2025 combined performance tests (55,700 requests) with an observational field study of 3,011,283 browser identifiers. Among identifiers with at least two recorded page views before any purchase, with store and week fixed effects and pre-purchase clicks, the odds of conversion were 1.7 times higher with recommendation use and 3.0 with search use; on all identifiers, where single-page visits dominate, they were 4.7 and 11.8. Order values were higher with feature use (seven of eight significant). For 7% to 19% of converting identifiers, an order contained a product clicked in a feature. These are associations, with self-selection. Mean client-side response times for repeated requests were 104–139 ms per deployment; cart-based recommendations, computed at request time, exceeded 200 ms. The main contribution is five design principles, offered as nascent design theory, from designing for persistent data constraints to minimising the retailer’s integration effort. We propose, untested, that integration effort outweighs algorithmic sophistication in SME adoption.

Source: Digital Business

DOI: 10.1016/j.digbus.2026.100191

Authors: Kailash Chowdary Bodduluri, Arianit Kurti, Francis Palma, Ilir Jusufi, Henrik Löwenadler

Institutions: Linnaeus University, University of New Brunswick, Blekinge Institute of Technology

Language: English

Topic: Digital Innovation in Industries

Subfield: Management of Technology and Innovation

Field: Business, Management and Accounting

Domain: Social Sciences

SDG: Industry, innovation and infrastructure

Open Access status: gold

Estimated APC paid: $1,630