Startup

A definitional review. Citations follow IEEE style; see References.

Abstract

A startup is, on the most common reading, a young enterprise pursuing an innovative product, service or business model with the intention of growing. The term is heavily used in policy and investment and yet has no settled definition: the literature applies at least four overlapping criteria — age of incorporation, innovation, growth and risk — with no agreement on which are necessary. This article sets out the competing definitions, the shift from newness to innovation, the divergence between legal and scholarly usage, and what the literature does not establish.

I. Definition

No definition commands consensus. The most institutionally weighty formulation is the European Commission’s, which defines a startup as an enterprise, regardless of its legal form, that is in the process of being set up and whose goal is to develop an innovative product or service, or a scalable business model — reproduced and dissected in [2], which immediately observes that it does not clearly state the criteria for a company to qualify. Two features carry weight: regardless of its legal form removes incorporation type from the test, and the disjunction — innovative product or service or scalable business model — makes scalability sufficient but not necessary, separating this definition from the growth-first ones in §IV-D.

Where a working criterion is needed, surveys stipulate one. The European Startup Monitor 2019 opens by stating that the term has no commonly agreed official definition and therefore adopts three researcher-set criteria: younger than ten years, an innovative product, service or business model, and an intention to scale — to grow employees, turnover, or markets served [3].

Both reviewed papers close by proposing their own synthesis. Ehsan: a rapidly growing firm due to its innovation in products/services and processes through the aid of IT/ICT-enabled services, status contingent on the age of incorporation permitted by the firm’s jurisdiction [1]. Pekevski: a newly established business, founded and run by young entrepreneurs, operating less than five years, that applies innovative ideas and new technologies for achieving fast market growth in a short period [2].

II. Origins

The term predates its present meaning. In economic geography before the 1980s, startup was rarely used and described the early stage of any firm’s activity in general terms — an attribute of all firms; only in the 1980s did it come to denote a particular kind of firm or working practice [2].

Academic definitions began from newness of legal existence, and virtually all studies prior to the 2000s used new as the primary discriminator [1], [2]. As reproduced in both reviews: Keeble (1976) took a startup to be the creation of an entirely new enterprise which did not formerly exist as an organization; Gudgin (1978) a firm which began production for the first time; Freeman et al. (1983) framed startups by their liability of newness and smallness; Carter et al. (1996) as a newly born company, without previous history of operations [1]. So defined, new covers every newly created firm in a period except those created by changes in name, ownership, location or legal status, and requires a known start date — conventionally the date of registration as a legal entity, chosen because those records are accessible. The age cut-off follows from this [1], [2].

From the 2000s the emphasis moved to innovation, attributed in both reviews to the increasingly complex requirements of domestic and foreign markets [1], [2]. Ehsan treats this as a change in what mattered rather than a correction: an earlier study of 121 London startups found marketing and financial control, not innovativeness, to be the most significant reported problems [1].

III. What distinguishes it

The concept is defined against two neighbours, differently in each case.

Against the SME. Small and medium enterprises are categorised on annual revenue, employee count and value of fixed assets. Both reviews observe that these quantitative thresholds exist for SMEs while there is a lack of literature outlining what a startup is, and treat the absence of an equivalent operational test as the core problem [1], [2].

Against the newly established company. Pekevski draws the distinction on trading behaviour rather than age: a newly founded company is focused on creating and maintaining a constant, stable revenue stream and receives income from establishment by selling products that already exist on the market; a startup does not generate income at the moment of establishment and introduces a unique new product or service, or enters an uncovered niche [2]. A practitioner handbook states the same boundary in one line: traditional businesses are not startups because they do not provide an innovative solution which can be easily reproduced to serve another region [4].

Ehsan’s ranking is the sharpest statement of the discriminating feature: of the four recurring variables, innovation is the key differentiator, growth and risk/uncertainty are outcome variables of that innovation, and age of incorporation is not a defining property at all but the cut-off point at which the status lapses [1]. Neither review treats technology as definitional: startups are not just technology companies but any companies in the process of being set up [2].

IV. The evidence

A. Four overlapping variables, not one definition

Ehsan identifies four overlapping variables — age of incorporation, innovation, growth, and risk/uncertainty — and finds that individual definitions select among them rather than combining them: the tabulated comparison shows pre-2000s entries marking age alone, contemporary entries age plus innovation, some adding growth, some risk, few all four [1]. Pekevski concurs from a different corpus [2].

B. The age cut-off is real but jurisdictionally arbitrary

Age is the one variable consistently operationalised, and the thresholds do not agree. Ehsan tabulates Italy under 5 years, India under 6, other European countries under 10 [1]. Pekevski’s survey of legislation adds Spain (five, or seven for biotechnology, energy and industry), Latvia and Lithuania (five), Portugal (ten), Italy (60 months), and France, where young innovative company status runs to eight years for social and eleven for tax exemption [2]. The disagreement reaches inside one organisation: the European Startup Network’s definition requires a firm younger than five years [2], while the European Startup Monitor 2019 — whose named research team are Network staff — sets it at less than ten years [3].

C. Innovation carries the definitional weight but is not measured

Ehsan concludes that establishing an age cut-off uncovers a more significant need for researchers to define and measure innovation in more absolute, universal and quantifiable terms, and that there is no metric for what level of high growth can be attributed to a startup’s innovative nature [1]. The variable doing the most discriminating work is the one with the least operational specification. Legislation substitutes proxies: Italy’s innovative startup status requires R&D at 15% or more of turnover or operating costs, a third of staff holding doctorates, or a registered patent or software licence; France’s JEI requires R&D at 15% of charges [2]. These are inputs and credentials, not innovation.

D. Growth is where the definitions genuinely disagree

The definitions reproduced in the reviews divide on whether growth is essential:

• Graham, as reproduced in [2] — a company designed to grow fast; being newly founded does not in itself make a company a startup, nor is technology, venture funding or an exit necessary, but the only essential thing is growth.
• Ries, as reproduced in [1] and [2] — a human institution designed to create new products and services under conditions of extreme uncertainty. Uncertainty, not growth, is the defining condition.
• Blank, as reproduced in [2] — a temporary structure that seeks a repeatable, scalable business model. The object is the search, not yet the growth.
• Krejci et al., as reproduced in [1] and [2] — innovation- and technology-based, with potential for rapid growth and scalability; growth is required, though as potential rather than record.
• The European Commission, as reproduced in [2] — innovation or scalability, so a firm can qualify on innovation without any growth claim.

Pekevski notes that growth is a weaker criterion than it looks, since founders of small businesses also aim to grow and expand [2]: only growth of a particular rate or kind would separate the categories, and no reviewed source specifies the rate.

E. Legal recognition is the exception, not the norm

Pekevski’s central empirical finding is that few jurisdictions define a startup in law. Formal inquiries to government institutions in Slovenia, Croatia, Serbia, Macedonia and Montenegro found no dedicated startup registries in any except Serbia; elsewhere new entities register uniformly as newly established companies, differentiated only by size, ownership structure, legal form, capital and management composition. Within the EU only a small number of states — Spain, Latvia, Estonia, Lithuania, France, Portugal, Italy — hold explicit legislative frameworks. Critically, ecosystem strength and legal definition are decoupled: Sweden, Germany, France and the Netherlands rank highest on the Global Startup Index 2024 for favourable startup conditions, yet Germany and Sweden have no legal framework or definition, relying on individual programmes and ad hoc initiatives [2].

F. Scholars and practitioners define the term differently

Pekevski reports a consistent split: academic definitions emphasise theoretical dimensions and measurable criteria such as innovation capability and growth potential, while entrepreneurs emphasise market dynamics, scalability, agility and the entrepreneurial mindset itself. The practitioner definitions reproduced there run to the frankly non-operational — a startup as a state of mind, a feeling, or the living embodiment of a founder’s dream [2]. A practitioner handbook offers a more usable three-term test — scalable, repeatable, innovative [4].

V. Scope of adoption

The term is in wide administrative and research use despite the definitional gap. Ehsan frames the problem in exactly those terms: promoting entrepreneurship in the form of startups is a policy activity given high priority all over the world, and governments needing to reach these firms must distinguish a startup from a small business, which requires an operational definition [1]. Pekevski argues the same absence affects policy formulation, entrepreneurial support mechanisms and investment decisions [2].

In practice the gap is filled by stipulation. The European Startup Monitor 2019 surveyed 848 usable responses from 31 countries, reduced from 1,353 after manual quality checks, using its own three criteria and its own four development stages — pre-seed/seed, startup, growth, steady [3]; the definition is an instrument of the study rather than a finding of it. The vocabulary also travels far beyond Europe and North America: a narrative review of Sri Lanka’s startup ecosystem in the 2019–2023 crisis synthesises 52 sources while treating startup as a settled category needing no definition [5].

VI. Limitations

The literature does not establish an agreed definition, and both reviews say so. Ehsan states that studies tend to employ different, often ad hoc, approaches to the definition and measurement of key concepts and relationships, and that this inconsistency creates a generalizability problem and weakens their external validity [1] — so any claim resting on comparison across studies inherits that problem. Pekevski’s parallel finding: no internationally or regionally uniform methodology or universally accepted criteria exist [2].

The key variable is unquantified, and the proposed definitions are untested. Innovation is named as the discriminator while simultaneously being named as the thing that still needs defining and measuring [1]; the same holds for the growth rate that would separate a startup from a growing small business [2]. Both reviews close by proposing a definition and neither validates it against a sample; Pekevski’s introduces young entrepreneurs as a criterion [2] with nothing in the corpus to show that founder age discriminates.

Coverage is narrow and source quality is uneven. Pekevski’s legislative analysis is confined to Europe with a Western Balkan focus and rests on secondary and institutional data rather than firm-level evidence [2]; it also dates Ries’s definition to 2006 in the body text and 2011 in the table. The European Startup Monitor 2019 states plainly that it did not aim for full coverage of European startups, that its steady-stage category is small (2.5% of the sample), and that Austrian data came from a partly divergent survey [3]. The Sri Lanka review is a narrative synthesis with no stated search protocol and reports no definitional finding [5]. The practitioner handbook [4] carries no author and no year and cites no literature, so its three-term test evidences practitioner usage only, not a scholarly definition.

References

[1] Z.-A. Ehsan, “Defining a Startup — A Critical Analysis,” SSRN, 2021. [Online]. Available: https://ssrn.com/abstract=3823361

[2] S. Pekevski, “What Is a Startup?,” Management Studies, vol. 13, no. 2, pp. 68–78, 2025, doi: 10.17265/2328-2185/2025.02.002.

[3] European Startup Monitor 2019, 2019. (No named authors in the bibliographic record; the report itself credits a research team drawn from the European Startup Network.)

[4] Basics of Entrepreneurship Handbook, n.d. (No authors and no year in the record.)

[5] A. M. D. D. K. Abeysinghe and S. K. Majumdar, “From Crisis to Capability: A Review of Sri Lanka’s Startup Ecosystem,” 2026, doi: 10.69968/ijisem.2026v5i165-72.