How start-up accelerators work to facilitate successful commercialisation: A critical realist perspective

Author Identifiers

Farzaneh Eslamloo https://orcid.org/0000-0001-7999-6063

Date of Award


Degree Type

Thesis - ECU Access Only

Degree Name

Doctor of Philosophy


School of Business and Law

First Advisor

Denise Gengatharen

Second Advisor

Kerry Brown


Start-up accelerators (SA), a form of innovation-focused program, have spread rapidly across the world in recent years. Many countries now understand that these programs can be highly effective in helping start-ups generate novel products with commercial potential. However, little is known about how SAs lead to successful commercialisation. This study addressed this knowledge gap using case studies of three Western Australian (WA) and six Iranian SAs, looking at the structures, agents, and the causal mechanisms that affected commercialisation. It also examined the influence of the WA and Iranian start-up ecosystem contexts, including the powers and liabilities that determine commercialisation opportunities.

The study used Chen’s taxonomy of program theory (comprising the action model and the change model) as a theoretical framework. Danermark et al.’s methodology of explanatory research, under a critical realism philosophical paradigm, was used to identify the causal relationships and generative mechanisms that determined commercialisation outcomes. In building theories, a modified version of Eastwood et al.’s explanatory theory-building method was used, comprising emergent, construction, and confirmation phases.

In the exploration phase, a realist review of prior research and grey literature was carried out, covering SA programs and the start-up ecosystems of WA and Iran. The review allowed the formulation of a general hypothesised action model of SA programs. This model was then confirmed by analysis of first-round interviews with experts familiar with the studied SAs, based on which the general SA program theory model (comprising the action and change models) was constructed. In addition, explanations of the role of SA programs in successful commercialisation outcomes were derived, using the context–mechanism–outcome (CMO) framework.

In the construction phase, second-round interviews were subjected to analytical resolution to abduct and confirm the most important mechanisms in each case study. In addition, the theorised CMO models were confirmed by considering the contextual differences between SA programs (their particular configurations of structures and agency).

In the confirmation phase, all previous data were reevaluated to retroduct the features of start-up ecosystems in WA and Iran, identifying the powers and liabilities at play in them, and concretising how they affect commercialisation outcomes.

The study showed that context—of SA programs and of the start-up ecosystem—is an important determinant of commercialisation outcomes. Screening, learning, networking, and product–market fit were the most important mechanisms in the case studies; that is, those with the causal powers to bring about commercialisation opportunities. However, the study also revealed factors that impede SA program success, including low follow-on investment, ecosystem immaturity, the ineffective agency of the WA and Iranian governments, and low levels of talent among start-up agents. These elements weakened the power of causal mechanisms to lead to commercialisation opportunities.

The research confirmed that program theory is a suitable theoretical base for the evaluation of SA programs and contributed to knowledge about SA programs. It determined what works, for whom, in what context, and why. It extended the existing body of knowledge by developing the action and change models of the SA programs studied in the context and start-up ecosystems of WA and Iran, which can be applied in research on other SA programs in different contexts.

Access Note

Access to this thesis has been embargoed until 8 July 2024

Access to this thesis is restricted. Please see the Access Note below for access details.