Abstract
Early-stage entrepreneurs iteratively revise problem statements (PS) in their pitch decks to persuade audiences of the value of their solution. This study examines how startups seeking funding within an accelerator program revise their PS during training. We analyzed initial and final versions of 66 pitch decks startups produced in Startup Chile in 2017–2018. A quantitative analysis showed that versions changed in their components and media choices, including increased references to existing solutions as well as more statistical evidence. A qualitative content analysis of three specific deck pairs showed that relationships across these edits sometimes impaired the argument coherence of the pitch, specifically the proposed problem–solution fit. We conclude with implications for entrepreneurship education programs.
Introduction
Founders of early-stage technology startups can talk for hours about their startups. But in most pitching situations, they don’t have hours—they only have seconds of minutes. For instance, when giving an elevator pitch, founders only have 30 seconds, in which they must cover “background and contextual knowledge showed in the presentation, project content and venture information, storytelling approach and style, and entrepreneurial flow or ‘algorithm’ of the pitch,” components that Margherita and Verrill (2021, p. 304) break down into 19 separate evaluation items. In an accelerator program such as the one we studied, Startup Chile (SUP), a founder might get a relatively luxurious 3 minutes to pitch, in which they must cover the problem, solution, market, team, and ask. In either case, founders must ruthlessly edit, making wise choices about what they can cut, add, and substitute to make their pitches both concise and effective. Wise editing is critical to both spoken pitches and to pitch decks: short slideshows that support the spoken pitch, but are also circulated independently.
Unfortunately, accelerators and other entrepreneurship education programs do not provide much guidance in editing. In the absence of such concrete guidance, founders often try different tactics in their pitch, gathering feedback from the audience, and adjusting the pitch based on that feedback (e.g., Belinsky & Gogan, 2016). Yet without systematic advice, founders sometimes don’t know how to apply feedback properly. This lack sometimes leads them to approach editing atomistically, without a sense of the argument's coherence. They sometimes focus on adjusting individual gestures and spoken word choices (Spinuzzi et al., 2023, pp. 53–54)—editing choices that threaten the strategic coherence of the startup's argument (Spinuzzi et al., 2016, p. 5). Startups face a tricky balancing act: Their argument must describe a story with a clearly related problem and solution (i.e., it must demonstrate problem–solution fit), told in enough detail to be compelling, while simultaneously being tightly edited enough to fit within constraints and keep their audience engaged.
Founders typically learn how to pitch in an accelerator program: “a fixed-term, cohort-based program, including mentorship and educational components, that culminates in a public pitch event or demo-day” (Cohen & Hochberg, 2014, p. 4). In accelerator programs, training is oriented toward securing seed financing: entrepreneurs are typically taught to craft compelling problem statements (PS), both in applications and presented through pitch decks. This is the case of SUP, a public, equity-free Chilean accelerator that offers early-stage international tech startups 6 months of training and mentoring to validate their products. Being less than 3 years old, having a minimum product, and being in the early stages of market validation, the selected participants create an initial pitch deck at the beginning of the program and present a revised final version at a pitch gala. In this culminating event, startups are expected to showcase a refined pitch that demonstrates the solution is worth funding. As mentioned, SUP pitches are a scant 3 minutes long, so founders must edit them tightly to stay within the program's constraints, but still preserve the argument's coherence.
How do founders balance the competing demands of making a pitch that is both coherent in argument—telling a story with a clear fit between a compelling problem and an equally compelling solution—and concise—tightly edited to fit within time constraints and keep audiences engaged? In this study, we examine how founders in two iterations of SUP (2017–2018) constructed and revised the PS in their pitch decks by investigating the following research questions:
To answer these questions, our analysis focuses on the PS of 66 pitch decks—each including an initial Version A and a final Version B—produced by startups participating in SUP. We undertake a quantitative analysis of our materials as well as a qualitative content analysis of selected materials. Like other investigations of PS (Cabezas et al., 2020; Cabezas & Bateman, in press; Hunter & Pellegrini, 2024; Pellegrini & Hunter, 2024), this investigation does not examine how audiences reacted to the pitches; instead it focuses on changes in argument composition through edits.
Before getting to the study, we review the literature on pitching.
Background
Seeking initial funding largely depends on the entrepreneur's ability to justify the value of their solution. However, at early stages, startup founders often struggle to do so, as they typically operate with a minimal product and an unvalidated market. Thus, this stage is characterized by high levels of uncertainty (Aldrich & Fiol, 1994; Cornelissen & Clarke, 2010; Parhankangas & Ehrlich, 2014). In this context, the pitch plays a central role in making the venture comprehensible and meaningful to external stakeholders. Scholars (e.g., Aldrich & Fiol, 1994; Cornelissen & Clarke, 2010; Lounsbury & Glynn, 2001; Martens et al., 2007) have emphasized the role of “stories” as a resource for constructing perceived value in the early stages of a venture. Stories, as described by Lounsbury and Glynn (2001), function as discursive devices that help reduce the uncertainty surrounding new ventures (cf. Aldrich & Fiol, 1994). As Martens et al. (2007) argue, stories serve three key functions: they help stakeholders understand the firm's identity; they clarify the rationale behind the startup's strategic decisions; and they connect the venture to broader cultural narratives in a way that renders it novel yet credible.
As part of the stories, various dimensions of the entrepreneurial venture can be brought into play. For example, Cabezas and Bateman (in press), in their analysis of early-stage entrepreneurs’ pitch decks, found that founders predominantly focused on specific topics, such as the company, the team, the problem, and the solution. In contrast, other topics—such as funding, clients, competition, and partners—were addressed less frequently. Similarly, Pellegrini and Hunter (2024), also analyzing early-stage pitches from SUP, observed that entrepreneurs often relied on rhetorical strategies grounded in causality. These strategies involved explaining the problem through references to broader societal changes or deficiencies in current solutions (causes) and framing its impact in terms of customer pain points (effects). Taken together, these findings suggest that early-stage entrepreneurs tend to emphasize what they know best—typically the problem and the product—while devoting less attention to business aspects that are more uncertain or still under development, such as competition, customer profiles, or monetization strategies.
To further explore the literature on pitch decks, we review the PS in entrepreneurship communication and startup pitch revision cycles.
Why the Problem Statement? The “Problem” in Entrepreneurship Communication
Early-stage startups must establish a well-defined problem if they are to communicate the value of their solution. The PS is a fundamental category of the startup argument, appearing at every stage of the startup journey. It is required across different accelerators in required applications, initial pitches, and final pitches (e.g., Sabaj et al., 2023; Spinuzzi et al., 2016, 2020) and is a component of the value proposition (VP; e.g., London et al., 2015). As Cabezas and Bateman (in press) have shown, after introducing the company, startup founders typically begin their pitch with a PS, which addresses and articulates this problem in terms that imply both the need for a solution and the salient criteria for that solution. This strategic move prepares the audience to understand the startup's offering as a solution, one that achieves problem–solution fit by demonstrating a thorough understanding of the problem and defining an appropriate solution (Okanović et al., 2021).
The PS continues to be critical to SUP in particular. Although our dataset is from 2016 to 2017, SUP still requires PS in application materials, and its rubrics for applications, intake interviews, evaluations, and pitches explicitly include the PS (CORFO, 2024).
As a strategic element in value creation, the PS has been conceptualized through at least four dimensions: (a) as a form of “pain” experienced by potential users; (b) as something to be “solved” through innovation; (c) as the “foundation” of the VP; and (d) as a “formal narrative unit” with its own structural and rhetorical features.
Pain experienced by potential users. In marketing studies, the notion of “problem” is often linked to the concept of pain, as in the expressions “customer pain points” or “relieving customer pain” (see, e.g., Handfield & Steininger, 2005; Holz et al., 2024; Liu, 2022). Osterwalder et al. (2014) proposed a practical framework that identifies three types of customer pain: (1) pains related to frustrations, costs, or inconveniences; (2) pains linked to barriers that prevent users from getting started or making progress; and (3) pains associated with potential negative outcomes or uncertainties that generate anxiety. Thus, framing the problem in terms of customer pain emerges as a persuasive strategy for legitimizing the relevance and urgency of a proposed solution—especially in the early stages of entrepreneurship, when tangible proof may still be lacking. More recently, entrepreneurship communication research has explored market pain in terms of “aspirin” versus “vitamin” solutions (Spinuzzi et al., 2018).
A problem to be solved through innovation. In this dimension, a problem tends to be part of a problem–solution pair, in which the solution is some sort of innovation, typically the startup's offering. Regarding the PS as a counterpart of a solution, Shmelova (2017) highlights that in the entrepreneurial landscape, the problem and its solution are inextricably intertwined, with the absence of one rendering the other futile. Blank (2007) explains that in startups, the problem is what customers will pay to resolve. Spradlin (2012) adds that a meticulous definition of the PS determines the finding of a solution. Ries (2011) further emphasizes that PS is relevant (among other factors) only when it has an effective and profitable solution. The solution plays a central role, while the problem serves as the grounding or contextualizing element that contributes to gain credibility. This relationship is often characterized as problem–solution fit, which “determines whether you have a problem worth solving” (Maurya, 2012, p. 8). Problem–solution fit is a key way of testing assumptions (Okanović et al., 2021) and is often challenging for inexperienced founders to formulate (Tripathi et al., 2019). Critically, this pairing is often the story of a pitch, although in informal problem-solving, the problem and solution are often discovered together (von Hippel & von Krogh, 2016).
The foundation of the VP. In this dimension, the most prominent approach to PS definition complements other relevant concepts, particularly the research on VP. Companies are considered “value delivery systems” (Lanning & Michaels, 1988), and Osterwalder and Pigneur (2010) state that the VP “solves customer problems and satisfies their needs.” This makes the VP the reason why a customer chooses a product or service that solves the problem over another. In fact, London et al. (2015) characterize the VP as a proposal argument that proposes a solution to an emergent problem. In terms of actor-network theory, the VP stabilizes the results of problematization and interessement, which involve the problem and the market (Spinuzzi et al., 2018). As Spinuzzi et al. (2018, p. 380) put it, referring to a specific student startup competition: “a business must sustain itself via a reliable revenue stream…without identifying a problem and market, these firms could not formulate the value proposition.” Thus, in terms of creating value, startups should pitch their solutions by referring to the market actors, either the competitors or existing solutions (market share) and the specific customers (the market segment).
A formal narrative unit. In the final dimension, a problem refers to a formal unit made up of different components. Focusing on components, Annamalai et al. (2013, p. 858) propose that the PS “should address all six questions: what, how, where, when, why, and who.” In the same line, but in the context of a capstone project, one of the engineering students from Dixon's research argues that: “the problem definition also should include a background/context statement, target specifications (preferably quantified), design constraints, and timeline for deliverables” (Dixon, 2012, p. 7). Among other elements, Kurbanoba (2017) suggests including a demonstration which justifies the need for the solution, a list of the challenges confronted by customers, an insight into the shortcomings of current solutions, and an outline of the market and intended audience. Kurbanoba further suggests that entrepreneurs begin with a single statement, supported by two pain points.
So far, we have outlined the different roles that the PS can play in early-stage entrepreneurial communication. These include: providing a motivating context for the proposed solution (pain); articulating the customer's needs that the solution is meant to address (problem–solution logic); identifying the target audience and the specific value offered to them (VP); and justifying the necessity of the solution through concrete information (formal structure), such as the shortcomings of existing alternatives or specific, preferably quantified, indicators. Regardless of which dimension is emphasized, early-stage startups must craft a well-articulated problem to fulfill key objectives—most notably, securing funding. Yet the PS must also be concise. In what follows, we examine how the “problem statement” is mobilized and revised in actual pitch decks.
Startup Pitch Deck Revision Cycles
Entrepreneurs must edit their pitch decks along several revision cycles in which information is shaped, cut, extended, packaged, or represented in different ways. In this field, Cabezas et al. (2020) provide an example of the editing-oriented studies on pitch. According to these authors, in the revised version, the firms tend to delete technical descriptions, hedge claims, and add VP, among other strategies. In a cluster of related studies (Jakobs & Digmayer, 2020; Spinuzzi et al., 2014, 2016; Spinuzzi, Nelson et al., 2015; Spinuzzi, Thomson et al., 2015), researchers examined iterations of pitch decks and related documents to understand how entrepreneurs responded to mentors’ feedback by selecting evidence from market reports, making arguments more complex, and characterizing markets through quotes. In another set of studies (Hunter & Pellegrini, 2024; Pellegrini & Hunter, 2024), researchers examined PS in SUP, identifying a consistent cause–effect structure in which the problem is expressed as resulting from societal changes or shortcomings in current products or services, resulting in painful effects. Gerding and Vealey (2017) examine a Kickstarter pitch in which civic entrepreneurs pose a “wicked problem”—experienced differently by various civic stakeholders—and offer a “hybrid solution” that can lever this group toward a shared solution. Finally, in an autoethnography, Belinsky and Gogan (2016) examine how Belinsky iterated a pitch over months, “stacking” different “frames” to reach differently positioned audiences by characterizing the problem in ways that reach across audiences. This research suggests that various elements of a pitch can be revised or edited to enhance the communication of value.
Another line of research on pitch decks has emphasized the role of visual elements in slide presentations. This strand of scholarship explores how visual and textual semiotic resources interact in the editing process, understood as a form of meaning-making. From this perspective, several studies have analyzed the semiotic composition of slide decks, although primarily in domains outside entrepreneurship, such as science and tourism. For example, Rowley-Jolivet (2004) compares the semiotic profiles of conference slide decks in medicine, geology, and physics. Her analysis considers linguistic and structural features as well as visual media such as figures, photographs, and charts, revealing that each discipline exhibits a markedly different visual style. Similarly, Hiippala (2013) provides a detailed account of the semiotic resources used in Finnish tourist brochures, describing how various elements—such as paragraphs, headings, illustrations, photographs, captions, and lists—combine to structure the reader's experience and convey meaning.
Crafting the pitch deck in competitive entrepreneurial contexts—particularly the development of the PS—involves iterating, discourse pivoting, and refining how the solution is presented. For instance, during pitch revisions, information such as problem cases, causal explanations, or statistical evidence is often omitted across iterative cycles (Cabezas et al., 2020). Visual elements, including graphs, figures, and images, are also frequently revised. In some cases, additional slides are introduced in response to trainers’ feedback (Spinuzzi et al., 2015). This underscores that creating value through a compelling PS is not solely a matter of verbal content, but rather a “multisemiotic” discourse composed of multiple types of meaning—both verbal and visual— that are deployed through deck slides.
Several scholars have analyzed the features of elements (or media resources) included in pitch deck presentations—for example, images (Feng, 2021; Zhao et al., 2014), text distribution (Garrett, 2015), fonts, tables, animations, and the use of color (Kmalvand, 2015). In the domain of entrepreneurial pitch decks, research has expanded its scope to explore broader units of observation—for instance, the overall functions of presentations (e.g., Daly & Davy, 2016) and the distribution of topics across slides (Cabezas & Bateman, in press). Regarding the latter, Cabezas and Bateman (in press), in their study of SUP pitch decks, found that early-stage entrepreneurs tended to dedicate more slides to specific topics—such as product, problem, client, and market—with the problem being, alongside the product, one of the most prominent topics (or components) and the one that showed the greatest variation in the number of slides. While this line of research is valuable for understanding how entrepreneurs communicate value in early stages, further evidence is needed to shed light on how such communication is crafted through revision. Indeed, since editing decisions—such as reducing the amount of information, removing causal explanations, or replacing text with images to better communicate the problem—often remain undocumented, it is difficult to identify how editing choices affect the overall argument.
Up to this point, we have seen that early-stage entrepreneurs can construct a problem narrative or statement in their pitch decks by articulating various elements (such as the problem, the solution, etc.) and by using multimodal resources (such as graphs, captions, headings, etc.). To understand how trained entrepreneurs revise their PS and to assess whether these revisions present the problem more concisely while retaining a coherent argument about problem–solution fit, we follow the methods below.
Methods
To investigate the research questions, we conducted a case study in which we analyzed changes in pitch deck PS in earlier and later versions of the same pitch decks. We employed content and media analysis to examine how the problem was represented in the initial version and how it was edited in the subsequent version.
This research was declared exempt by the Institutional Review Board of the first author's university. The examples we show are anonymized, although they are considered public information.
Site
All data were obtained from SUP, a public equity-free Chilean accelerator that trained early-stage technology entrepreneurs as they sought funding for their startups.
The data come from pitch decks produced in the 17th and 18th generations of the Seed program (2017–2018). As part of the requirements, participating startups were no older than 3 years, had a functional product, and were in a market validation stage. In total, 153 startups participated, representing diverse sectors and from all over the world. Since all SUP programs are international, all documents produced during participation in the program were in English.
Like other accelerators, SUP participants submitted an application, delivered an initial pitch, participated in various events, received mentoring, and delivered their final pitch during Pitch Day (see, Sabaj et al., 2023, for more context on the SUP program). The most successful startups were then selected and rewarded with additional funds and extensive training.
Data and Data Collection
In this study, we focus on two versions of startups’ pitch decks: Version A, presented at Intro Day at the beginning of the program, and Version B, presented during Pitch Day at the end of the 4.5-month program. These decks followed a template provided by the accelerator.
Since we focused on how PS were edited, we reduced the data through two steps:
We discarded cases in which only one of the versions was available. This first exclusion decision resulted in 96 pairs of pitch decks. During this period, participants go through different intensive training, feedback sessions, and socializing events. We discarded cases in which (a) the PS section was missing or (b) the PS was unrevised.
After applying these exclusion criteria, the final corpus consisted of 66 pairs of slide decks (Table 1).
Number of Pitch Decks Analyzed.
Data Analysis
To analyze the data, we first coded all pitch decks in terms of components or prominent topics (RQ1) as well as media resources such as images, text distribution, fonts, tables, animations, and the use of color (RQ2), yielding a quantitative analysis of these codes. (These codes are discussed in Tables 2 and 3.) Based on this work, we then selected three pairs of decks for a qualitative content analysis (RQ3). These three decks were selected to illustrate three different editing strategies that we detected via the coding analysis.
Final Components Codes.
Final Media and Its Definitions.
Coding for components. We followed a mixed approach that combined top-down and bottom-up strategies, as is common in technical and professional communication research (Lauer et al., 2018). This involved identifying preexisting categories to develop an initial model, which was later adjusted to better reflect the characteristics of the pitch decks. To identify the components included in the PS and those most frequently modified between versions, we began by locating the slides dedicated to the PS, using as a reference the template provided by SUP to guide founders in crafting their pitch decks. This template includes subject/topic units such as company, team, problem, and product, among others. We then analyzed the content and headings of each pair of slide decks to refine our initial findings. Finally, we identified and categorized the components of the PS in both versions of the pitch decks (Table 2), which allowed us to characterize the revision of the PS in detail. Due to the mixture of top-down and bottom-up coding, some codes (e.g., target market) connect clearly with the entrepreneurship literature, while others (e.g., statistical evidence) emerge from our inductive reading of the decks.
After identifying the slides dedicated to the PS, we estimated the proportion of PS-related slides in relation to the total number of slides in both versions of the pitch decks. Since a single component could appear across multiple slides, this count was not exclusive. For example, “context” could be present in slides 2, 3, and 4, while “pain points” might appear in slides 3, 4, and 5. The number of slides was used as a proxy for estimating the amount of information dedicated to the PS.
Coding for media resources. As highlighted in our literature review, each component can be associated with specific types of media resources. Following an inductive strategy, we identified media that were used in the first version and edited in the second version. This task proved challenging, which may partly explain the scarcity of studies examining multimodal revisions in pitch decks. Drawing on previous classifications (Bucher & Niemann, 2012; Hiippala, 2013; Rowley-Jolivet, 2004; Zhao & van Leeuwen, 2014; Zhao et al., 2014), we developed concise definitions for the 120 media types identified in the PS slides. Using reduction techniques based on similarity, frequency, and clarity, we grouped these into nine overarching categories (Table 3).
An important obstacle in analyzing media edits across two versions is the issue of the referential frame, namely, the starting point for edits. In the second version of any element, new media resources could potentially emerge in an unlimited manner, while the removal of elements is constrained by the initial count. Consequently, to overcome this restriction, we analyzed PS editing from a multimodal perspective in two stages. First, we counted all types of media in each component of the PS in Version A, and second, we identified how many times founders reduced or added a media type in Version B (editing strategy).
Qualitative content analysis of selected decks. Our coding helped us to identify changes in frequency of components and media types, but did not help us to examine the relationships among these changes. For instance, examining these codes may tell us that a startup deleted two stock images and inserted a heading and an icon—but it does not tell us whether these deletion and insertion choices were related or how the edits reshaped the pitch deck's argument. To understand how these individual editing choices relate to changing the argument of the PS, we must examine these choices in context.
Thus, to discuss the impact of PS revision, we selected three pairs of pitch decks so that we could examine relations among edits in detail. Specifically, we examined the PS to identify the relationships among editing choices in three pairs of pitch decks. We selected these pitch decks because they represented different patterns of editing choices.
Ensuring Credibility and Trustworthiness of the Data
As Bateman (2022) notes, triangulation is one of the important obstacles when analyzing multimedia data. To approach the steps described above, we bolstered data credibility and trustworthiness through a sequential and iterative process focused on reliability and triangulation, enhancing the proposal by O’Connor and Joffe (2020), namely, all the process of coding, including starter codes, refinements, and final selection and definitions was made by four of the authors and two research assistants, collectively.
The process involved data familiarization, open coding, listing features, defining codes, and final coding. To ensure reliability, four authors and two research assistants analyzed one-third of the data over 12 meetings totaling 23 h and 48 min. The two-thirds of the data were coded by two pairs of coders. For the rest of the corpus, the coders of these pairs were interchanged.
Results and Discussion
In this section, we describe our results, which allow us to understand how entrepreneurs revise their PS to communicate and create value for their audience at an early stage while trained in an accelerator program. As stated in the methodology section, we focus on the difference between the original version and the final version presented at the Pitch Day. After reviewing quantitative results, we qualitatively examine three selected cases to illustrate how relationships among individual edits yield changes in argument.
First, let's consider descriptive results regarding changes in the number of slides dedicated to the PS. After the editing, the total number of slides of the PS did not vary much (222 vs. 217). On average, Version A of the PS represented 22.67% of the total slides of the pitch deck (222 of 979 slides), and Version B, 21.42% (217 of 1013 slides). Thus, PS represents around one-fifth of the whole pitch deck. This affirms the consensus in entrepreneurship communication research: that startups regard the PS as a crucial element to communicate and create value for the audience. Of the 66 pitch decks analyzed, 31 reduced the number of slides of the PS, and 31 increased them in the second version, while four pitch decks did not vary in terms of the number of slides of the PS in both versions. On average, when decreased, the reduction was in 6.25 slides, but when increased, it was in 17.37 slides.
Revision of Pitch Deck Components
This section discusses our first research question: which components of the pitch deck are most modified between versions? The analysis revealed that all components, except existing solutions and weaknesses, decrease in Version B. In Table 4, the raw and relative frequency of the components in both versions is presented (the total pitch decks are 66 for both versions).
Problem-Statement Components Per Pitch Deck (Between Version A and Version B).
Overall, these results support the idea that, when constructing the PS, entrepreneurs tend to focus primarily on contextualizing the issue and emphasizing customer pain points—elements that appear with the highest frequency. In contrast, business-related information, such as statistical evidence or insights into the target market, is less frequently included. Why is there more emphasis on the problem–solution narrative than on business-specific data? As previously discussed, early-stage entrepreneurs often operate in highly uncertain environments (Aldrich & Fiol, 1994; Cornelissen & Clarke, 2010; Parhankangas & Ehrlich, 2014 ), where access to detailed business intelligence—such as market statistics, competitor analysis, or validated customer segments—is limited or still under development. These findings provide meaningful insights into the trajectory of startup development, particularly in relation to early-stage value creation. While founders often demonstrate a solid understanding of the technical strengths of their offering, they frequently struggle to mobilize entrepreneurial knowledge, such as market positioning, competitive landscape, and other contextual data.
As observed, among the different components, the “existing solution and its weaknesses” is the only one that consistently expands across successive slides. This may be due to the type of information required at this stage of the startup lifecycle, which calls for a more detailed description of how the proposed solution differs from other similar alternatives. Such information is crucial for validating product/market fit, which is the main objective of the training program within the accelerator framework.
When analyzed by the volume of slides of each component, results tend to follow the same pattern. In Table 5, the raw and relative frequency of slides per component is shown (total number of slides is 222 in Version A and 217 in Version B, the categories are not exclusive).
Number of Slides Per Component in the Problem Statement.
Context and pain points occupied most slides and experienced the greatest reduction in both versions. However, despite this decrease in slide count, they remained the most prevalent elements, albeit extensively edited. This confirms the importance of describing pain points to legitimize the startup at this early stage as a provider of solutions to a customer's problem, such as frustration, barriers to progress, or uncertainty (Osterwalder et al., 2014). Surprisingly, the “existing solutions” section underwent only minimal variation. This observation reinforces our interpretation that, when refining the PS, entrepreneurs retain this component as a foundation for restructuring the problem.
Revision of Media Resources
Now, we turn to RQ2: which types of media resources (e.g., text, charts, images) are most frequently revised? One strategy used by early-stage entrepreneurs consists of editing media resources. Next, we describe these resources in the original Version A and Final Version B. Table 6 shows the raw frequency of media in each component of the PS of Version A.
Media Resources Per Component in Version A.
The analysis of Version A shows that when founders represent the PS in their pitch decks, they mostly use, in decreasing order, general descriptions (N = 113; 27.49%), digital images (N = 75; 18.25%), charts and numbers (N = 63; 18.25%), and headings and subheadings (N = 54; 13.14%). In Version A, the founders crafted the PS using more varied media resources in the most prominent components of the pitch decks. This suggests that context and pain points, as the most relevant structures in the PS, are media resource-independent. Other components do not include all the resources available; instead, fewer and less varied media are used. As expected, there are evident patterns in the use of media in specific components. For instance, charts and numbers are important in statistical evidence, but clearly nonexistent or very uncommon in the target market and in existing solutions. Other relationships, which are not so evident, are the role of icons and logos in existing solutions and the use of photos in the target market.
Table 7 shows the media editing strategies that founders display in Version B.
Media Resources Editing Strategies in Version B.
In the final Version B, founders edited the PS in pitch decks using three main strategies: adding, deleting, or substituting. In the case of adding, for instance, some entrepreneurs included charts and numerical data in the statistical evidence component. For deleting, we observe that certain elements—such as icons in the same component—were simply removed. Regarding substituting, entrepreneurs often revised the context component by replacing general descriptions with new or more refined versions of those descriptions. Overall, the most frequently used strategy was substituting, suggesting an effort not just to condense but to rearticulate and enhance the clarity of the problem narrative.
Identifying the differences between Versions A and B allowed us to identify the main strategies used by entrepreneurs to edit the PS:
Number of slides addressing the PS: About 20% of the pitch deck's slides are dedicated to establishing the problem. Decrease of PS slides: The number of slides dedicated to the PS tends to decrease in Version B. Increase in slides describing the weaknesses of current solutions: While all components tended to decrease in Version B, the description of current solutions and their weaknesses tended to increase. Use of context and pain points descriptions: These descriptions are the most stable components of the PS. Addition of media: Charts and numbers in statistical evidence and general descriptions of pains are added.
Relations Among Edits: Threats to Argument Coherence
As discussed above, we can see general trends in edits across the dataset. But to answer RQ3—How do relationships among individual edits yield changes in argument?—we must examine relationships among edits. To understand how these editing choices interrelate, we turn to a qualitative content analysis.
To conduct this analysis, we selected three disparate pairs of decks, illustrating clear changes in argument reflected across our larger dataset. In examining the relationships among edits, which were made under severe space constraints (decks were limited in length) and time (decks had to support a 3-min presentation), we identified tradeoffs that founders made. These three cases illustrate different approaches to editing and their effects on argument coherence, specifically problem–solution fit.
Case 1: Cutting content load—by omitting the problem. Case 1 (Figure 1) involves a solution designed to help client companies structure their data for decision-making. The initial deck (A) presented four slides with comparative textual and visual complexity. The revision (B) cut the number of slides to two, and streamlined the content by reducing the textual and visual complexity, yielding just text. While this editing pattern could conceivably enhance readability and concision, in this case, the revision actually omitted the problem–solution segment entirely.
Problem statement before (A) and after (B) for Case 1.
In Version A, the PS is described across four slides, incorporating both user pain points and user reactions. Slide 2, “Problem #1: too many options,” presents the problem through a metaphor in which a supermarket client is depicted as struggling to choose the best bottle of juice in the juice section (2). Slide 3, “Problem #2: too many opinions,” illustrates the problem by means of a screenshot of an e-commerce website in which some products for sale have received several evaluation comments (3). Slides 4 and 5 depict a client's reaction using a meme image and a stock image in which a woman expresses frustration. Critically, this slide sequence emphasizes the emotional aspect of pain points: feelings such as stress and frustration.
In Version B, the PS section was reduced to two slides, with no images. Through this editing, the founders deleted both the metaphor of buying juice and the emotionally evocative images. Instead, Version B represents the problem via two declarative sentences elaborated as “facts” playing the role of contextual information. These slides no longer explicitly reference a problem, instead implying it: Slide 2 implies that the Internet is full of unstructured data (“Internet is a lot of unstructured text”), while Slide 3 implies that machines are needed to process this information (“Human brains can process limited information”).
In Version B, the founders cut the content load by omitting an explicit description of the problem. The result did indeed reduce the number of slides and the density of information per slide. But in doing so, it lost both the concrete metaphor and the emotional evocation of Version A. In editing out these interrelated elements, the founders omitted their arguments about pain points (why must unstructured text be processed?) and current solutions’ weaknesses (what are the limits of current technology?). In Version B, the problem is implied but unstated. That is, problem–solution fit is unclear: Without clearly establishing the problem, the startup has trouble arguing why this solution is needed or why it must take the shape that it does.
Case 2: Cutting text details—and losing the problem's causes. Case 2 (Figure 2) involves a technical assistance startup that offers maintenance services. In Case 2, the revision shifts the focus from explaining the underlying causes of the problem to highlighting its consequences. This case is about an entrepreneurial project offering technical assistance services. Like Case 1, Case 2's slide count is reduced in Version B.
Problem statement before (A) and after (B) for Case 2.
In Version A, slides represent the PS in two slides, explicitly identifying the “Problem” (as can be seen in the slide heading) and explicitly declaring the effects of the problem—a critical move in PS (Pellegrini & Hunter, 2024). Both slides describe the problem using declarative sentences where a specific entity (i.e., Small and Medium Businesses) is described in terms of two main “possession attributes” (using “have”). In Slide 3, the attribute corresponds to “a cost overrun of 25% by machine failures” (i.e., current solution weakness), where “machine failures” functions as the cause of the problem. Interestingly, the attribute is visually reinforced by a pie chart in which one-fourth of it represents the costs. In Slide 4, the attribute corresponds to (the lack of) “effective maintenance programs,” a phrase which is further elaborated by two textual features: they tend to be “corrective” and “not optimized preventive.” These attributes, corresponding to current solutions and their weaknesses, are put in a relational structure through the verb “mean,” where information on the left is the cause of the information on the right. Linearly read, it means that the lack of effective maintenance programs involves “lots of resources, time, and money.” In terms of color, yellow depicts the new information offered to the audience (on the right side). Although it seems rather simple, the composition in Version A is quite complex.
In Version B, declarative sentences containing attributes and causal relations are replaced by two noun phrases (“Waste of time and resources” and “SME 25% over cost”) whose distribution implies that they are semantically related: A means B (Figure 1). The relation is implicit. Text in Version B is visually reinforced by (or translated into) icons depicting money, time, and percentages.
In this revision, the founders not only chose to reduce the number of slides, but they also chose to reduce the amount of text, replacing sentences with noun phrases and icons. The resulting slides focus on pain points, but omit the causes of the problem. Thus, the audience may not understand that the core problem is the lack of maintenance programs. This issue is key, since the startup offers technical assistance services—and without establishing the causal connection, the startup has difficulty establishing that their solution is appropriate. Again, the problem–solution fit has become less clear through these edits, impacting the argument's coherence.
Case 3: Cutting irrelevant information—and making the problem implicit. Case 3 (Figure 3) involves a service for centrally managing air conditioning (or “aircon”) systems. The revision improves the coherence and flow of information but makes the problem less explicit. Like the other two cases, Case 3's revision has fewer slides than the original deck.
Problem statement before (A) and after (B) for Case 3.
Version A includes an appeal to the audience in Slide 2: “In this last week have you felt cold?” The audience is invited to imagine being cold through the stock image of a baby wearing several coats, suggesting that one solution to face cold temperatures is to wear lots of layers. In Slide 3, no heading is provided; we can infer that the presented images correspond to two other solutions to withstand cold: “kerosene and gas.” Both are depicted through heater images presented with a ban symbol. These solutions based on fire heating are negatively reacted to in two ways: textually, with the phrase “no comments…,” and visually, with a cartoon image of a caveman dancing and warming up in front of a bonfire. This accounts for a personal standpoint regarding the use of current solutions, probably considered old-fashioned. Finally, Slide 8 seems unconnected from the previous slides: A sentence broken into 10 lines, in pink caps, is presented (“… and there are millions of aircon already installed in universities, schools, offices and others around the world that are waiting to be managed”). It is not clear if the text presents a context (the fact that there are plenty of aircons) or a pain point (they should be managed). If it is to be interpreted as a pain point, then the problem remains unclear since no information about why the aircons should be managed is provided.
Version A thus Presents a Problem of Coherence. Why are Aircons Waiting to be Managed? Why Do People Act like Cavemen?
In Version B, the problem is drastically reshaped through edits. The information in Version A, representing kerosene and gas heaters as inconvenient alternatives (“no comments”) is changed in Version B by Slide 2, a text-based slide asking “why heating/cooling solutions are still so ‘old-fashioned’?” These “old-fashioned” solutions (kerosene and gas heaters) remain implicit. To reinforce the idea that people prefer old-fashioned or traditional heating methods, a new piece of statistical evidence is added: “In Chile <1% population has AC.” But this is not the problem solved by the technology! No more ACs are needed. To clear this up, Slide 3 specifies the topic of the problem with the heading “Facts about climatization.” Although this slide allows the reader to understand the problem, this problem is not explicit. Indeed, using nominal phrases, climatization (the targeted current solution) is described as “expensive,” and with “no energy management” (their weaknesses). We can infer that that's why people go for cheaper old-fashioned alternatives.
Finally, while Version A attempted to include the number of ACs to be managed, Version B specifies the target market with statistical evidence about the technology use in the local setting (Chile).
As we saw, Version A presented a problem of argument coherence. Version B does not repair this problem—it still implies rather than states the problem—but it does identify a target customer. The edits, aimed at establishing the market, did not address problem–solution fit, and thus the argument's coherence was not improved.
Losing argument coherence due to atomistic edits. How should these clusters of edits be interpreted? The reduction of components, slides, and media resources may be explained by time–space constraints, which can be a general rule that acts on most genre revision cycles, but especially, in entrepreneurship types of texts. However, as we have demonstrated, these related edits imperiled the pitch arguments’ coherence by either obscuring or not improving problem–solution fit.
In Case 1, the founders replaced all images with declarative sentences. These edits reduced content load, but also cut a metaphor and emotional appeals. Without these, the audience may not understand the problem and the pain associated with it. Without this contextual information, the deck's PS lost some of the elements that we expect in a PS: pain points, current solutions, and weaknesses.
In Case 2, founders replaced complex sentences with noun phrases supported by visual figures. These edits reduced complexity and the number of slides, but they also eliminated causal arguments. Without understanding the cause of the problem, the audience may have difficulty understanding how the startup's offering will be able to solve it.
In Case 3, founders replaced images of older technologies with text. In doing so, founders replaced implicit connections between images with less implicit connections among statements. Although the edits still implied rather than stated these relations, they did refer to the main problem: the lack of AC management.
In our quantitative analysis of individual edits of components and media resources, we found that PS of revised pitch decks tended to have fewer slides and components while adding information on the weaknesses of current solutions as well as additional media. In this closer qualitative analysis of three decks, we also see that the relationships among edits impacted the overall argument of these decks’ PS, sometimes in ways that made problem–solution fit less coherent.
Conclusion and Implications
In the Introduction, we mentioned a startup founder who thought the best way to edit his pitch was by A–B testing each individual component. This example shows the weakness of the learning loop approach that Lean Startup promotes: Such editing choices are not just individual or atomistic, but rather interrelate to impact the coherence of the larger argument.
Editing is crucial for founders, especially in the PS, which is a fundamental pillar for the pitch. Early-stage technology startups often operate with immature or experimental products, lack validation from real users, and face uncertain market demand. Thus, a clearly articulated and strategically revised PS is crucial for signaling investors that a problem actually exists, one that customers will pay to solve, making the startup's solution worth investing in. While several studies have examined the rhetorical and discursive features of entrepreneurial pitches, few have investigated how these narratives evolve through processes of revision and editing. By analyzing pitch decks with multiple versions—such as Versions A and B—this study offers a valuable lens for understanding the relevance and communicative implications of specific changes. Our findings suggest that revisions in pitch decks should not be seen merely as atomistic improvements in clarity or design, but rather as interrelated to change the broader argument of the PS. These interrelations among edits can shape how the problem is framed as part of a problem–solution pair, and, ultimately, how coherently the pitch is structured. For this reason, the revision process deserves greater attention from mentors, communication experts, and entrepreneurs alike, as a key site where entrepreneurial value is actively constructed. Problem–solution fit is not just important to the entrepreneur's conception of their business (Tripathi et al., 2019), it's also critical to the argument coherence of the pitch.
With this in mind, SUP and similar accelerators can help startups by providing more conceptual grounding to help them choose individual edits and think through how those edits affect argument coherence. We see the following implications for such entrepreneurship education programs.
Components. Through inductive and deductive coding, we have developed a set of components (Table 2). We do not see these as the final word, but as a starting vocabulary that could help entrepreneurship education programs to describe components for founders to edit. In learning to name these, founders can conceptualize them and identify opportunities to edit them.
Media resources. Similarly, we provide a starter set of media resources (Table 3) based on the research literature as well as the decks we examined. This set can be expanded further (e.g., by including video or animation). Again, in learning to name these, founders can conceptualize them and identify opportunities to edit them.
Relations among edits. Most critically, founders need grounding in how to look past individual edits to understand how they cohere in a larger argument. With this grounding, they can go beyond atomistic A–B testing to consider how editing choices can yield better coherence and demonstrate better problem–solution fit.
In the Appendix, inspired by Margherita and Verrill (2021), we provide a heuristic based on components and media resources, aimed at helping accelerators guide startups to apply conceptual language to their PS. The “Objectives” column lists argument objectives that can guide startups in making their arguments, arguments that they can build through selecting components from the other columns. Through this heuristic, startups can be guided as they think through the interrelations among parts of the argument that the PS makes.
Footnotes
Acknowledgment
We are thankful to Startup Chile for giving us access to data to conduct this research.
Ethical Approval and Informed Consent Statements
The project in which this specific research was embedded was declared exempt by the Institutional Review Board of the first author's university. The examples we show are anonymized, although following the procedure we conduct, they are considered public information.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Chilean National Fund for Scientific and Technological Development Fondecyt under Grant 1170133. The work of Omar Sabaj was supported by DIDULS/ULS through the project under Grant PTE213269.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Author Biographies
Appendix
Heuristic for Editing Components and Media Choices.
Component
Objectives
Edits: Options to Achieve These Objectives Concisely and Consistently Via Media Choices Such As …
General Descriptions
Headings and Subheadings
Interpersonal Appeals
Charts and Numbers
Digital Images
Icons and Logos
Photos
Cartoon Characters
Geometrical Shapes and Diagrams
Context
Audiences can understand the problem through the pitch deck alone.
Audiences can understand who encounters the problem.
Pains
You describe the pain that the target market encounters.
You establish the severity of the pain.
Statistical evidence
You provide statistical evidence in addition to anecdotal evidence.
This statistical evidence demonstrates the target market size.
This statistical evidence demonstrates the severity of the problem in terms of monetary impact or other relevant metrics.
Existing solutions
You demonstrate that the target market has tried and failed to address this pain via other solutions, either those currently on the market or idiosyncratic attempts.
If applicable, you describe other possible solutions that could partially solve the problem.
Target market
You demonstrate that the target market exists.
You demonstrate that the target market is large enough to support a solution.
You demonstrate that the target market faces severe enough pain that they will pay for that solution.