1,721,211 research outputs found
Six pillars of modern entrepreneurial theory and how to use them
Full text linkIn recent years, there has been an explosion of interest in entrepreneurship from both practical entrepreneurs and researchers. While theories are helpful for explaining business-driven activities in a startup, they are also valid in reasoning for the practical activities occurring in the entrepreneurial context. We believe that startups would benefit from the awareness of these entrepreneurial theories and the understanding of how they can be connected to decision-making in both business and engineering perspectives. In particular, we want to focus on theories that are already used by practical entrepreneurs and their advisors. As an example, we have studied the Scandinavian entrepreneurial ecosystem. We selected six groups of theories that might be particularly relevant for the startup population, namely (1) core competence and resource-based view, (2) effectuation, (3) the fulfillment of entrepreneurial opportunities, (4) bricolage, (5) business model innovation, and (6) lean startup. In this chapter, we explain these theories including the ongoing research around them, the connections among these theories, and how they can be applied in a real case study
Software Startup ESSENCE : How Should Software Startups Work?
No full textSoftware startups need to work in a systematic fashion just like mature organizations. However, existing software engineering methods and practices are not aimed at software startups. They do not account for the business aspect of startups and may not be well suited for software startups in general. The Lean Startup Methodology on the other hand contains some useful practices for software startups but is nonetheless impractical, offering little in the way of telling you what to do. Software startups are thus required to tailor their own method. Currently, many software startups simply work ad hoc or use various Agile methods and practices. In terms of Agile methods and practices, little consensus exists between startups. In this chapter, we discuss methods and method tailoring. We give guidelines on how to create your own way of working and recommend a tangible tool for doing so: the Essence Theory of Software Engineering.peerReviewe
Startup Metrics That Tech Entrepreneurs Need to Know
No full textMetrics can be used by firms to make more objective decisions based on data. Software startups in particular are characterized by the uncertain or even chaotic nature of the contexts in which they operate. Using data in the form of metrics can help software startups to make the right decisions amid uncertainty and limited resources. However, whereas conventional business metrics and software metrics have been studied in the past, metrics in the specific context of software startups have not been studied. In this chapter, we present the results of a multivocal literature review to offer you 118 metrics practitioner experts think software startups should measure. These metrics can give you ideas for what your startup should measure.peerReviewe
Gender Bias in the Recruitment Process of IT Startups in the Netherlands
Full text linkIn today’s fast-changing and innovative world, startups must compete amongst themselves and other well-established companies to hire the best talent in order to succeed. Diversity within the recruitment process is typically not a priority, even though it is well known that a diverse team is beneficial for (business) outcomes. Through a multiple case study performed at 5 IT startups based in the Netherlands, we observed that gender bias is introduced from the first moment that the need for an employee has been identified until candidate hiring. This is a direct result from (1) a lack of resources (e.g., time and money), (2) urgency to find the first and best candidate, or (3) the awareness of the startup founders
Growth hacking to retain customers in Nepalese e-commerce companies
Full text linkThis masters thesis investigates the growth hacking strategies used for customer retention in Nepali e-commerce businesses. Growth hacking is an experiment based and data informed marketing approach that uses digital marketing tools and tactics as well as traditional marketing channels to help technology companies show “proof-of-concept” and sustainability before gaining funding. E-commerce enterprises may profit from the use of growth hacking techniques and methodologies, and this research aims to find out what the primary factors are that lead to its acceptance. How organizations may expand by concentrating on customer retention at each step of the customer lifecycle. Despite the fact that acquiring customers is a profitable strategy for organizations, it is not sufficient for growth. Retaining current customers is essential for the long-term success of the organization. In order to expand effectively, certain steps need to be made to keep customers. The purpose of this study was to understand how the ecommerce business companies perceive and practice growth hacking and identify what growth hacking strategies are used by such organizations.
First, it looks at the background literature to study the theories and develop a conceptual model considering the from the theories and models. Then a qualitative study was conducted where 7 semi-structured interviews were conducted through which a thematic analysis was done to develop a model refining the previous model developed through the background literature.
The findings showed that the growth hacking practices are slightly different from that of the literature thus a proper model was created
Exploring Large Language Models for Software Startups: A Design Science Research Study
Full text linkSoftware startuper er sentrale drivkrefter for økonomisk vekst og skapelse av innovative produkter, men de fleste startuper ender med å mislykkes. Dette skyldes i de fleste tilfeller selvpåførte problemer, snarere enn konkurranse. En vanlig årsak til dette er manglende validering av startup-ideen og å finne samsvar mellom produkt og marked eller problem og løsning. Selv om det er utviklet flere startup-rammeverk for å hjelpe startuper med å navigere disse utfordringene, blir de sjelden tatt i bruk i praksis.
For å redusere disse utfordringene, utforsker denne studien hvordan Store Språkmodeller (SSM-er) kan tilpasses for å støtte startuper. Generativ Kunstig Intelligens (KI) og SSM-er viser lovende egenskaper i å assistere selskaper, og modifisering av slike verktøy rettet mot spesifikke bruksområder kan betydelig øke deres nytteverdi for det gjeldende bruksområdet. Ved å følge Design Science Research Metodologien (DSRM), går studien gjennom tre DSRM-sykluser med mål om å tilegne kunnskap om konsepter for praktisk anvendelse av en startup-tilpasset SSM, gjennom iterativ utvikling. Syklus 2 og Syklus 3 gjennomfører begge en runde med brukertesting med representanter fra startuper, samt en kvalitativ og kvantitativ analyse av resultatene.
Det endelige verktøyet har som mål å forbedre støtte til startuper og benytter teknikker som Retrieval-Augmented Generation og nøye utformede systeminstruksjoner for å veilede SSM-ens adferd. Funnene fremhever de spesifikke behovene og forventningene startuper har til en SSM tilpasset deres omgivelser, spesielt behovet for kontekst- og bedriftsrelevante svar. Basert på disse funnene gir studien et sett retningslinjer for å forbedre systeminstruksjoner og implementering av startup-teori. Disse retningslinjene har som mål å ytterligere forbedre nytten av systeminstruksjoner og startup-teori i en chatbot. Verktøyet og funnene legger et grunnlag for videre undersøkelser og utforskning av SSM-er i en startup-kontekst.Software startups are essential drivers for economic growth and the creation of innovative products, however, most startups fail. This high failure rate is in most cases due to self-inflicted issues rather than competition. One common reason for this is not efficiently validating the startup idea and finding the product/market or problem/solution fit. While several startup frameworks have been developed to help startups navigate these challenges, they are rarely adopted in practice.
To mitigate these challenges, this research explores how Large Language Models (LLMs) can be customised to support software startups. Generative Artificial Intelligence (AI) and LLMs show promise in assisting companies, and modifying such tools to specific contexts can significantly increase their usefulness for that context. Following the Design Science Research Methodology (DSRM), the study goes through three DSRM cycles aiming to gain knowledge of concepts for the practical application of a startup-oriented LLM through iterative development. Cycle 2 and Cycle 3 each conduct a round of user testing with startup representatives, accompanied by a qualitative and quantitative analysis of the results.
The final tool aims to enhance startup support and employs techniques such as Retrieval-Augmented Generation (RAG) and carefully crafted system instructions to guide the LLMs behaviour. The findings highlight the specific needs and expectations startups have regarding an LLM tailored to their environment, specifically highlighting the need for context- and company-relevant answers. Based on these insights, the research provides a set of guidelines to improve the system instruction and implementation of startup theory. These guidelines aim to further enhance the usefulness of system instructions and startup theory in a chatbot. The tool and findings lay a foundation for further research and exploration of LLMs in a startup context
AI-Assisted Agile Project Management - On the feasibility of adopting Generative AI for backlog refinement - An exploratory case study
Full text linkEffektiv håndtering av backloger er avgjørende for å sikre at utviklingsteam holder seg i tråd med skiftende krav og forventninger fra interessenter. Men etter hvert som produktbacklogger vokser i omfang og kompleksitet, har de en tendens til å bli rotete med overflødige, utdaterte eller dårlig definerte oppgaver, noe som kompliserer prioritering og beslutningsprosesser. Denne masteroppgaven undersøker hvordan Generativ Kunstig Intelligens (GenAI) kan effektivisere backlog-håndtering ved å automatisere deteksjon og løsning av overflødige saker.
Inspirert av nylige fremskritt innen store språkmodeller og vektorrepresentasjoner, innebærer studien utvikling og evaluering av et GenAI-basert verktøy integrert med Jira. Dette verktøyet identifiserer semantisk like backlog-elementer og flagger dem som potensielle duplikater. Det gir også handlingsrettede forslag til hvordan disse kan slås sammen eller håndteres på andre måter. En Design Science Research-metodikk ble brukt for å implementere artefaktet, inkludert flere iterasjoner av utvikling og brukertesting, med en avsluttende valideringsfase som involverte erfarne prosjektledere og utviklere.
Kvantitative funn viser at AI-assistert backlog-håndtering forbedrer presisjon og tilbakekalling ved deteksjon av overflødige saker sammenlignet med manuelle metoder, samtidig som tiden det tar å fullføre oppgavene reduseres betydelig. Kvalitative tilbakemeldinger tyder på at det intuitive, brukersentrerte grensesnittet fremmer høyere tillit og en sterk vilje til å integrere verktøyet i eksisterende smidige arbeidsprosesser.
Ved empirisk å demonstrere at GenAI kan adressere sentrale utfordringer i backlog-håndtering, bidrar forskningen til den stadig voksende kunnskapsbasen innen AI-drevet prosjektledelse. Innsiktene vil veilede programvareutviklingsteam, prosjektledere og organisasjoner i å ta i bruk intelligent automatisering for å oppnå mer effektiv backlog-håndtering.Effective backlog management is critical for ensuring that development teams remain aligned with evolving requirements and stakeholder expectations. However, as product backlogs consistently grow in scale and complexity, they tend to become cluttered with redundant, outdated, or poorly defined tasks, complicating prioritization and decision-making processes. This Master's thesis investigates how Generative Artificial Intelligence (GenAI) can streamline backlog refinement activities by automating the detection and resolution of redundant issues.
Taking inspiration from recent advancements in large language models and vector embeddings, the study involves development and evaluation of a GenAI-based tool integrated with Jira. This tool identifies semantically similar backlog items and flags them as potential duplicates. It also provides actionable suggestions for merging or otherwise resolving them. A Design Science Research methodology was used to implement the artifact, including multiple iterations of development and user testing, culminating in a validation phase involving experienced project managers and developers.
Quantitative findings show that AI-assisted backlog grooming improves precision and recall in detecting redundant issues compared to manual methods, while significantly reducing the time-to-completion. Qualitative feedback suggests that the intuitive, user-centric interface promotes higher trust and a compelling willingness to integrate it into existing agile workflows.
By empirically demonstrating that GenAI can address key challenges in backlog management, the research contributes to the evolving body of knowledge within AI-driven project management. Its insights will guide software development teams, project managers, and organizations in adopting intelligent automation to achieve more efficient backlog management practices
Enhancing Innovation and Technological Efficiency in Bangladesh SMEs: Investigating the Impact of E-commerce Adoption
Full text linkE-commerce is transforming small businesses in Bangladesh by helping them reach more customers and innovate through technology. This study is being conducted to understand how these businesses adopt e-commerce, overcome challenges, and improve innovation and efficiency with the aim of providing helpful insights for their future use of technology
Exploring Large Language Models for Software Startups: A Design Science Research Study
Full text linkSoftware startuper er sentrale drivkrefter for økonomisk vekst og skapelse av innovative produkter, men de fleste startuper ender med å mislykkes. Dette skyldes i de fleste tilfeller selvpåførte problemer, snarere enn konkurranse. En vanlig årsak til dette er manglende validering av startup-ideen og å finne samsvar mellom produkt og marked eller problem og løsning. Selv om det er utviklet flere startup-rammeverk for å hjelpe startuper med å navigere disse utfordringene, blir de sjelden tatt i bruk i praksis.
For å redusere disse utfordringene, utforsker denne studien hvordan Store Språkmodeller (SSM-er) kan tilpasses for å støtte startuper. Generativ Kunstig Intelligens (KI) og SSM-er viser lovende egenskaper i å assistere selskaper, og modifisering av slike verktøy rettet mot spesifikke bruksområder kan betydelig øke deres nytteverdi for det gjeldende bruksområdet. Ved å følge Design Science Research Metodologien (DSRM), går studien gjennom tre DSRM-sykluser med mål om å tilegne kunnskap om konsepter for praktisk anvendelse av en startup-tilpasset SSM, gjennom iterativ utvikling. Syklus 2 og Syklus 3 gjennomfører begge en runde med brukertesting med representanter fra startuper, samt en kvalitativ og kvantitativ analyse av resultatene.
Det endelige verktøyet har som mål å forbedre støtte til startuper og benytter teknikker som Retrieval-Augmented Generation og nøye utformede systeminstruksjoner for å veilede SSM-ens adferd. Funnene fremhever de spesifikke behovene og forventningene startuper har til en SSM tilpasset deres omgivelser, spesielt behovet for kontekst- og bedriftsrelevante svar. Basert på disse funnene gir studien et sett retningslinjer for å forbedre systeminstruksjoner og implementering av startup-teori. Disse retningslinjene har som mål å ytterligere forbedre nytten av systeminstruksjoner og startup-teori i en chatbot. Verktøyet og funnene legger et grunnlag for videre undersøkelser og utforskning av SSM-er i en startup-kontekst.Software startups are essential drivers for economic growth and the creation of innovative products, however, most startups fail. This high failure rate is in most cases due to self-inflicted issues rather than competition. One common reason for this is not efficiently validating the startup idea and finding the product/market or problem/solution fit. While several startup frameworks have been developed to help startups navigate these challenges, they are rarely adopted in practice.
To mitigate these challenges, this research explores how Large Language Models (LLMs) can be customised to support software startups. Generative Artificial Intelligence (AI) and LLMs show promise in assisting companies, and modifying such tools to specific contexts can significantly increase their usefulness for that context. Following the Design Science Research Methodology (DSRM), the study goes through three DSRM cycles aiming to gain knowledge of concepts for the practical application of a startup-oriented LLM through iterative development. Cycle 2 and Cycle 3 each conduct a round of user testing with startup representatives, accompanied by a qualitative and quantitative analysis of the results.
The final tool aims to enhance startup support and employs techniques such as Retrieval-Augmented Generation (RAG) and carefully crafted system instructions to guide the LLMs behaviour. The findings highlight the specific needs and expectations startups have regarding an LLM tailored to their environment, specifically highlighting the need for context- and company-relevant answers. Based on these insights, the research provides a set of guidelines to improve the system instruction and implementation of startup theory. These guidelines aim to further enhance the usefulness of system instructions and startup theory in a chatbot. The tool and findings lay a foundation for further research and exploration of LLMs in a startup context
Bamberg: Ausländische Mitarbeitende langfristig binden. Wie Unternehmen Migrantinnen und Migranten integrieren : Zwischenergebnisse einer Studie der Universität Bamberg
No full textInterview: 25.06.202
- …
