This chapter draws on the literature to develop our theoretical definition of a scale-up. A critical discussion of previous literature helps distinguish between aspects that are essential or unnecessary for a definition of a scale-up. We then put forward the following five suggestions for a theoretical definition of a scale-up: (1) Scale-ups are in all sectors, not just IT; (2) Scale-ups have relatively high levels of spending on marketing and sales; (3) Scale-ups have low marginal costs of production; (4) Scale-ups are young, but not too young; (5) In defining scale-ups, we start from the set of HGFs. The chapter ends with our own theoretical definition of a scale-up.

5.1 Firm Growth or Scaling: Reinventing the Wheel?

There seems to be a fashion among scholars to give new names for already-established phenomena. This might be an effort to make their contributions seem more original, while at the same time releasing them from the duty to cite previous related work.Footnote 1 For example, “knowledge acquisition” sounds different from ‘learning’ although it would be difficult to explain how these two concepts differ. It seems that many scholars use the word “scaling” when they could simply refer to “growth”. There is already a huge literature on firm growth. However, a 2014 paper on “scale adjustment” in Strategic Management Journal starts with these words:

While much is understood about the general pattern of industry dynamics, a critical element underlying these dynamics, the rate of expansion of individual firms, has been largely overlooked.Footnote 2

If, however, we call it “firm growth” rather than “the rate of expansion of individual firms” it is much harder to claim that the topic has been overlooked. That is why, four years earlier, McKelvie and Wiklund (2010, p. 261) wrote that “Firm growth constitutes one of the central topics of entrepreneurship research”, and refer to the “massive empirical research” that had already focused on firm growth.

One the one hand, research on scaling might be an attempt at reinventing the wheel, using a new term (scaling) to describe an older familiar concept (HGFs), jettisoning the duty to cite previous work, and avoiding previous inconvenient results (e.g. difficulties in predicting HGFs, lack of persistence in HGFs over time, and the surprising ability of random walk models in providing statistical explanations of firm growth patterns). This motivation of ‘reinventing the wheel’ seems particularly salient when scale-ups are defined precisely according to the HGF definition, which will annoy both HGF researchers (because of reinventing the wheel) and scale-up researchers (because a scale-up is not really the same thing as an HGF).

On the other hand, scaling is not just reinventing the wheel, if scaling is considered a distinct type of growth from just rapid growth. This chapter emphasizes that scaling up is indeed distinct from just ‘plain old growth’.

We seek to provide a clear definition of scaling up, and encourage future scholars to use the word “growth” unless they specifically have in mind scaling. Scaling and growth are not synonyms, and we hope that after this book, the two terms will be used less often as synonyms. Furthermore, a scale-up is not a synonym for an HGF, neither at an empirical level nor at a theoretical level. Instead, the core concept of scaling up refers not to growth in general, but to a unique mode of growth.

While a precise and standardized definition of scale-up has not yet been agreed upon by the broader community, we discuss the core idea of a scale-up, and its essential characteristics, before pragmatically discussing (in the next chapter) how a scale-up could be empirically defined with the available data.

5.2 Previous Definitions: A Critical Discussion

Many definitions of scale-ups can be found in the previous literature, as the study of scale-ups has evolved and gained sophistication. Our goal here is not to “name and shame”, but to highlight the different definitions that have been used, explain their strengths and shortcomings, and to encourage some steps towards a standardization of a definition.

Coutu (2014) defines scale-ups in line with the Eurostat-OECD (2007) definition of a HGF, i.e., with average annualized sales growth greater than 20% over a three-year period, and with ten or more employees at the beginning of the observation period. This definition has been adopted by a number of other scholars as well (e.g., DeSantola and Gulati, 2017; Duruflé et al., 2017; Gulati and DeSantola, 2016; Belitski et al., 2023; Denney et al., 2023). Scale-ups are indistinguishable from HGFs in this definition. Some scholars consider that HGFs and scale-ups are synonyms (e.g. Zeng et al., 2023).Footnote 3 However, applying various terms to study a single phenomenon leads to confusion, impedes interdisciplinarity, and retards the development and accumulation of theoretical and empirical insights. On this point, we agree with Coviello (2019, p. 5), who writes with bold in and block capitals: “Scaling is NOT just about high growth.”

Other definitions of scale-ups are difficult to measure and evaluate. Coviello (2019, p. 15) argues that scaling is more a reflection of a stage of development where a firm delivers a proven concept to a wider audience and includes standardization/automation for efficiency gains, has a diverse management team, has high absorptive capacity, moves into international markets and is relatively asset-light. We agree with some, but not all, of these suggestions. In our view, a definition of scale-ups does not necessarily require a diverse management team.Footnote 4 While such diversity (i.e. employee diversity, diversity of the top management team (TMT)) may be desirable (Hoffman and Yeh, 2018), it does not seem to be a critical deciding factor that allows us to discriminate between scale-ups and non-scale-ups.

Also, in contrast to Coviello (2019), we consider that scale-ups do not necessarily have to have growth that is international (as opposed to being domestic). There is a lot of interest in scaling from IB (International Business) scholars (e.g. Mihailova, 2023; Mithani, 2023; Monaghan et al., 2020; Reuber et al., 2021; Tippmann et al., 2023). Scaling lends itself to internationalization well (e.g. Table 1 in Reuber et al., 2021). Internationalization is indeed compatible with a scaling approach to growth. However, it is not a strict requirement: there are some firms that can scale up without internationalizing. Indeed, scale-ups may thrive for a long time in countries with large domestic markets such as USA or China.Footnote 5 Stallkamp et al. (2022) investigated a sample of digital firms that achieved an IPO and observed that “among the digital firms in our sample, almost 30% had no foreign sales when they filed for IPO” (p. 103).

A further divergence with Coviello (2019) is that, in our view, it seems unnecessary to refer to absorptive capacity, first because absorptive capacity is difficult to measure, and second because there may be some scale-ups that can grow fast without necessarily having above-average levels of absorptive capacity. These factors (diversity, internationalization, absorptive capacity) may be in place among scaling firms but provide undue characteristics of a definition.

Similarly, Shepherd and Patzelt (2022) draw upon Sutton and Rao (2016) to conceptualize ‘organizational scaling’ as “spreading excellence within an organization as it grows” (Shepherd and Patzelt, 2022, p. 1). This seems to us to be an inspiring although impractical definition. Spreading excellence is an assumption that is not only unhelpful for empirical work (since excellence is rather vague and unmeasurable), but also unnecessary for theoretical work. A scale-up presumably has excellent routines and practices to have such a rapidly growing user base (otherwise it would not grow), but even if it has only a rapidly growing customer base but not excellent routines (e.g. clumsily exploiting a valuable government monopoly) then it could still be classified as a scale-up, largely on the basis of its realized growth.

There are thus distinct challenges in pinpointing what a scale-up is, and how scale-ups should be operationalized and measured using available data. It is also challenging to argue that the characteristics attributed to scale-ups do not apply to HGFs, furthering the definitional blurred lines between scale-ups and HGFs.

Some authors suggest that scaling can be either fast or slow (e.g. Büge and Ozcan, 2021; Stallkamp et al., 2022). The idea of ‘slow scaling’ could be potentially confusing because in our definition scale-ups are a subset of high-growth firms. Büge and Ozcan (2021) suggest that Facebook’s Libra would have done better to slow down its scaling in the face of high regulatory complexity and risk. In our approach, a firm engaging in ‘slow scaling’ (e.g. reducing its growth rate from 250 to 25% per year) would still be a scale-up if its average annual growth rate is above the growth threshold.

Another recent definition of a scale-up is from Jansen et al. (2023, p. 6): “We define scale-ups as those HGFs up to 10 years old that have grown to at least 50 employees or more by the tenth year of existence or at the year of measurement, whichever is less.”Footnote 6 This differs from the definition that we will later propose. What we like about this definition is that scale-ups are a subset of HGFs. Where we disagree with this definition, however, is that considerations of scaling up as a growth stage (as discussed in Chap. 3.2) are not considered, and that the structural changes due to scaling are not mentioned (e.g. low marginal costs, ramping up of marketing expenditure).

Some definitions of scale-ups mention that eligible firms should be innovative (e.g. Vandresse et al., 2023, pp. 40, 46). While our expectation is that scale-ups are somehow innovative, their innovative nature may not be visible in standard indicators such as R&D investment and patents. Indeed, it is hard to believe that an HGF would not be innovative in one way or another. At a minimum, all HGFs (and a fortiori all scale-ups) are innovative in the sense of having the new-to-the-firm ‘organizational innovations’ (i.e. structural transformations) which are an inevitable outcome of growth.

Further discussion about defining scale-ups can be found in Bohan et al. (2024) in their ‘Journal of Business Venturing’ article entitled “What is scaling?”. They define scaling in terms of sustained exponential growth. This definition seems unhelpful for several reasons. First, exponential growth is actually the baseline understanding of growth in the academic and policy literature, and is nothing exceptional. For example, growing 10% in each year will inevitably mean that the growth amount is increasing exponentially, while the growth rate is constant (e.g. Tornqvist et al., 1985). Hence, claiming that scaling is exponential growth is somewhat trivial. Second, there is nothing in this definition about the minimum growth threshold to qualify as a scale-up. Growing 0.5% in each year for a decade should hardly qualify a firm as a scale-up. Third, there are growth setbacks, such that growth may be negative in year 2 but positive overall for years 1–3, hence problematic for fitting an exponential curve. Fourth, there is insufficient consideration of scaling as a growth stage: the ramping up of production after finding product/market fit in the context of a stages-of-growth model.

In sum, the literature is missing a shared understanding of how scale-ups should be defined. The need for a shared understanding and standardized definition of a scale-up was recently expressed by Jansen et al. (2023, p. 7):

Various definitions and key attributes of scaling have emerged during the last couple of years, yet contributions have been developed rather independently. This has led to a scattered collection of narrowly defined studies across different domains with limited synergies in conceptual development.

To clarify this definitional quagmire, we seek a data-friendly definition that is simple, direct and captures the efficiency arguments of scaling in the literature while, at its core, takes into consideration the logic of economies of scale and the potential for having a rapidly expanding business model that has certain characteristics such as high fixed costs, low marginal costs, and efficiency in production and replication.

5.3 Five Suggestions for How a Scale-up Should Be Defined

Scaling-up is a stage in the life-cycle model of firms, and there have always been major practical problems with attempts to empirically operationalize these stages-of-growth models (for reviews, see Coad, 2009; Levie and Lichtenstein, 2010). Furthermore, even if the theoretical constructs in the stages-of-growth depiction of the scale-up construct correspond to real-world phenomena, they might not be measured in databases that are available to most scholars. Ideal data would include variables such as marketing expenditures, and perhaps also hard-to-obtain data such as marginal costs as well as business model reconfiguration (Osterwalder and Pigneur, 2010). Ideally, we would have large-sample representative data, to accurately gauge the proportions of firms that are scale-ups.

That said, clearly not all firms that suddenly ramp up marketing expenditure are scale-ups. For example, a firm producing a stable product that ramps up marketing in time for Christmas (an annual event) or a one-off event (such as the Olympics or FIFA World Cup), might temporarily display the symptoms of scaling-up (i.e. high marketing spend), but would not necessarily correspond to the concept of a scale-up. Hence, not all firms that look like scale-ups are “true” scale-ups. This is not the case for HGFs though—any firm satisfying the growth requirements would qualify for the HGF label.

Based on these insights we propose five suggestions for the definition of a scale-up:

Suggestion 1

Scale-ups are in all sectors, not just IT.

IT intensive sectors are particularly amenable to the scaling up of growing new firms, because of characteristics linked to digitalization. Digital resources such as data, software, and AI are essentially scale free, such that firms’ marginal costs remain low for very large production quantities (Adner et al., 2019). As a result, studies of scaling have often focused on the IT sector (e.g. Giustiziero et al., 2023), or the fintech sector (Jansen et al., 2023, page 10).

Giustiziero et al. (2023, p. 1392) motivate their theoretical focus on digital firms:

“We use the term “digital firms” (contrasted with “industrial firms”) to describe firms that participate heavily in the digital economy by either using a significant share of digital resources (e.g., software, algorithms, data) and/or by selling a significant share of digital products and services (e.g., platforms, software, media). We posit that digital firms tend to have more scalable resource bundles due to significant economies of scale in their productive resources and due to markets with low distribution costs and strong network effects”

That said, scale-ups can also be found in other sectors that do not focus specifically on IT, such as the case of Copenhagen Seafood A/S (Nielsen and Lund, 2018) and IKEA in furniture (Jonsson and Foss, 2011).Footnote 7 This is because digitalization and IT have features of a general purpose technology (GPT) that enhances the productivity of operations across many sectors.

It is difficult, and in our view unfruitful, to try to delimit which sectors include bundles of scalable versus non-scalable productive resources. Therefore, we infer whether it is possible for firms in a sector to scale based on whether we observe growth behavior of firms that is indicative of scaling (without attempting to observe particular aspects of the scalability of their resource bases). It is perhaps unlikely that a firm in the ready-mixed concrete industry can scale as rapidly as a firm in the software industry, but we do not rule this out.

Suggestion 2

Scale-ups have relatively high levels of spending on marketing and sales (Blank, 2013).

Here we could consider the scale-up stage to commence if there is a sudden step-change in marketing costs (in terms of total amount as well as in terms of the share of total revenues) that happens a few years after entry and that continues for a few years, and that occurs before or at the same time as the firm achieves rapid growth. When looking at the distribution of marketing spend among HGFs, it would be reassuring if there was a peak in the time series, indicating a step change or discontinuity to distinguish between scale-up HGFs and non-scale-up HGFs, such that scale-up HGFs would have a visibly separate cluster of characteristics compared to non-scale-up HGFs. If there were no such bimodality in the HGFs’ distribution of marketing spend (i.e. non-scale-up HGFs with low marketing spend, and scale-up HGFs clustered around a second mode corresponding to high marketing spend), such a lack of bimodality would reinforce intuitions that scale-up HGFs are only different from non-scale-up HGFs by degree, and not by kind (see above, Chap. 4.6).

That said, the gang of scale-ups comprises many exceptions. Some scale-ups may grow through viral marketing, which need not be expensive. One example would be Hotmail, which grew fast in terms of its user base thanks to its strategy of automatically attaching a short advertising message at the bottom of emails sent by Hotmail users to others.Footnote 8 Another example would be LinkedIn, which expanded its network at viral speed by introducing an innovative feature that allowed users to import their address book contacts into LinkedIn (Hoffman and Yeh, 2018). Hence, an indicator of scale-ups should not necessarily exclude firms that are not starting to spend more on marketing.

Suggestion 3

Scale-ups have low marginal costs of production.

Low marginal cost of production is a key feature of scaling up (De Ridder, 2023), and it explains why scaling up is strongly linked to profitable growth (Nielsen and Lund, 2018). This idea of low marginal costs of production leads to two major practical problems. First, as competition policy and industrial organization researchers are painfully aware, it is hardFootnote 9 to proxy for “marginal costs of production” with available data. Second, it is too simplistic to distinguish between fixed costs and marginal costs because of feedbacks between them. For example, a minor improvement in a software’s functionality might not be worthwhile when there are ten customers but would be worthwhile when there are 10,000 customers. Hence, as the customer base grows, the product improves in a way that corresponds to fixed costs, but these fixed costs only become worthwhile once the firm has crossed a certain size threshold. Before crossing such a size threshold, scale-up firms would be recommended to do “things that don’t scale” (Graham, 2013; Hoffman and Yeh, 2018). Hence, they are “fixed costs” (paid one time only for all customers) but they depend on the size of the customer base (if there are enough customers, the fixed cost is worth paying). The total “fixed costs” thus are not so fixed because they vary with the size of the customer base.

Suggestion 4

Scale-ups are young, but not too young.

Scale-ups need to spend time iterating, revising and reconfiguring their business model before they reach the scale-up stage. There is presumably no such thing as a “born scale-up” because this neglects the time spent in the two search stages in Fig. 3.1, unless of course the first two search stages occurred during the ‘gestation’ period before the firm’s official birth. The search stage may have occurred in ‘stealth mode’ before the firm was officially registered in a business register, or perhaps while working for a previous employer. It thus may appear to be a “born scale-up” even if a lot of preparatory work was done before the firm’s first official business day. In other words, the firm-specific preparatory groundwork before scaling-up may be brought to the nascent firm as an initial endowment at the time of birth, drawing on work done during the gestation stage, rather than being categorized as post-entry activity.

Suggestion 5

In defining scale-ups, we start from the set of HGFs.

Scale-ups are assumed to be HGFs, while not all HGFs are scale-ups (Coviello, 2019). Scale-ups are thus a subset of HGFs that satisfy the ideas above, and in particular, Suggestion 2 (ramping up of marketing investment before the high-growth episode) and Suggestion 4 (scale-ups are not newborns but not old either).

5.4 Discussion of Well-Known Cases

5.4.1 Is Uber a Scale-up?

Is Uber a scale-up? Uber has been notorious in making losses in the years leading up to its IPO in May 2019 (O’Reilly, 2019), with an EBITDA more negative than -$2billion in each year from 2016–2021.Footnote 10 It would thus not qualify as a scale-up, if scaling up is defined in terms of profitable growth. However, it did rapidly scale up its operations on the basis of a business plan that had been refined and redesigned to be rolled out at scale. So, on that basis, we could consider that, ideally, an indicator of scale-up used in empirical analysis would highlight that Uber is a scale-up. Other authors have considered that Uber is a scale-up (e.g. Hoffman and Yeh, 2018; Pfotenhauer et al., 2022). This implies that a profitable final state should not be a necessary part of the definition of scale-up.

A desirable property of an empirical definition of scale-up is that firms that are generally considered to be prototypical “scale-ups” would not be classified by the empirical definition as non-scale-ups. For example, if we restrict the set of scale-ups to profitable firms only, then we would have to exclude Uber which was unprofitable for a long time (even after its IPO). Kuratko et al. (2020, p. 112) actually state that the majority of blitz-scaling companies are NOT profitable.

5.4.2 Is Apple a Scale-up?

Sometimes large mature multiproduct firms are described as engaging in scaling. For example, Apple is presented as a serial blitz-scaler, due to scaling waves enabled by products such as the iPod, iTunes, the iPhone, and the iPad (Hoffman and Yeh, 2018, p. 13). Microsoft has been described as a blitz-scaler on account of its cloud computing service Azure (Kuratko et al., 2020, p. 111).

At the time of the iPhone’s success, was Apple a scale-up? Apple was already large and had many products at that stage. Theorizing about scale-ups often considers that it is a unique growth stage that happens shortly after entry (e.g. Hellmann and Thiele, 2022). Some consider that scale-ups are specialized (Giustiziero et al., 2023) rather than multiproduct firms.Footnote 11 Most would agree that, to the extent that the iPhone operated as a standalone product, then the iPhone could be considered as a scale-up—at the product-level of analysis. At the firm level, it would also be possible that Apple could be classified as a scale-up, if it satisfies (at the firm-level) the empirical requirements presented in the next section, such as 20% annual growth for a three-year period.

5.4.3 Was Obama’s 2012 Campaign a Scale-up?

Hoffman and Yeh (2018) consider that Barack Obama’s 2012 presidential candidate campaign can be considered to be an instance of blitz-scaling. Obama’s campaign grew from zero to 700 employees in a single year, as well as having orders-of-magnitude more volunteers. The campaign leveraged technology to achieve viral growth and powerful distribution through existing networks. Non-profit-oriented social movements such as Obama’s campaign can be extremely effective ways of amplifying new ideas and messages to have large scale impacts on society (Chliova and Ringov, 2017; List, 2022). According to our definition, however, Obama’s 2012 campaign would not be counted as a scale-up, because it would not satisfy the empirical requirements of being a business firm that survives for at least three years.

5.5 Our Theoretical Definition of a Scale-up

We conclude this chapter by stating our theoretical definition of a scale-up:

Theoretical Definition of scale-up

scale-ups are a subset of high-growth firms that have a specific style of growth that requires ramping up production of a new product (or service). This product has finished the processes of refinement, the business model need not pivot but has become clear, and no further iteration stages are required to find a good fit with the market. All that remains is scaling up. Crucial elements of this definition are that scaling-up is a stage in a stylized life-course model of firm evolution. Scaling-up changes the proportions of a firm (in terms of raising fixed costs, reducing marginal costs, having an above-average capital intensity in terms of tangible capital and intangible capital and also with regards to IT and perhaps also robots and automated systems). Scaling-up connects favorable supply-side conditions (initial product development costs are near their end; low marginal costs of production) to favorable demand conditions (marked perhaps by increasing returns due to network externalities). Scaling-up is often observed alongside a burst of marketing effort to accompany what is essentially a pent-up supply-side push towards a hungry market.