Linked but Desynched: An OODA Analysis of Associated Entrepreneurship Accelerator Programs

Pitching and Venture Development

New ventures must develop both an offering (a product or service) aimed at market need and a business model that can sustainably deliver that offering. Furthermore, they must pitch their offering and business model to stakeholders (potential investors, partners, suppliers, mentors, team members, and customers) whose support they need in order to start and sustain that venture. This work of pitching has historically been supported by a genre called the business plan. As Blank (2013b) argued, a business plan is “a static document that describes the size of an opportunity, the problem to be solved, and the solution that the new venture will provide”; it precedes the product on which the business is built, assuming that “it’s possible to figure out most of the unknowns of a business in advance, before you raise money and actually execute the idea.” He charged that a static business plan prevents the entrepreneur from collecting customer input from the market—until “the sales force attempts to sell it,” at which time “entrepreneurs learn the hard way that customers do not need or want most of the product’s features” (p. 5).

The business plan, according to Blank (2013b) and other devotees of the lean startup methodology (e.g., Blank & Dorf, 2012; Maurya, 2012; Ries, 2011), is too static, slow, and, most important, insular to deal with the rapidly changing landscape of business competition. It assumes isolation rather than interaction with the business environment. Thus, it yields both a product and a business model (i.e., a way of sustaining the product through reliable income streams) that have not been sufficiently tested. Because of these drawbacks, lean startup advocates have suggested either replacing or preceding the business plan with more nimble descendant genres focused on continual interaction. These genres include the pitch, a live performance (backed by a slide deck) meant to offer the basics of the business argument and gather feedback during a question-and-answer (Q&A) session; a business model canvas, a heuristic meant to display the business model with its many complex connections; and a minimum viable product, essentially a functional prototype of the envisioned system. All three of these descendant genres offer hypotheses, test them with rapid audience feedback, and iterate them to yield a compelling claim for value (i.e., value proposition), a functional business model, and a product that can realistically provide value for the audience. They replace isolation with interaction, encouraging entrepreneurs to “get out of the building” (Blank & Dorf, 2012), interact with potential customers, and cocreate value with them. Blank’s lean startup inspires the approach for many entrepreneurship training programs (Mansoori et al., 2019), including the National Science Foundation’s I-Corps Go program and SEAL. Indeed, Blank was integral in creating the I-Corps program for university innovators.

Such entrepreneurship training efforts include accelerators and incubators. As Cohen (2013) argued, these two types of efforts have different structures and purposes (p. 20):

An accelerator is “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). It is short-term (between 3 days and 3 months), competitive, and cyclical, involving intense mentorship and a seminar-style education; it addresses early-stage ventures, when the entrepreneur is still trying to figure out a business model and identify a customer segment. Accelerators are frequently structured as pitch competitions, with the entrepreneurs competitively pitching their venture at the end of the program. Examples include SEAL, I-Corps Go, and Cleantech Open. Additionally, some university programs function as accelerators, such as the University of Texas Longhorn Startup, Texas Tech Accelerator, and Baylor University BRIC LAUNCH. SEAL has traditionally been a 9 to 11 week program, but in 2018, it was shortened to 7 weeks.

Accelerators and incubators thus vary in their outcomes, lengths, cycles, and decisions—not just between the two groups but also within groups. For instance, the accelerator Three-Day Startup teaches the rudiments of entrepreneurship through highly structured, hands-on activities, resulting in a pitch that likely does not represent a workable business. In contrast, the SEAL accelerator also teaches the rudiments of entrepreneurship but structures longer conversations with mentors, resulting in a public decision: a go or no-go pitch in which the founders decide whether their proposed venture is workable. And the I-Corps Go regional accelerator lasts 3 weeks, focusing on the basics of lean startup and preparing ventures to perform 100 interviews of potential customers—interviews they must complete before they enter the 7-week I-Corps Go national program.

Yet accelerators and incubators are interlinked. Sometimes this linking is ad hoc, as when a firm enters multiple accelerators and incubators. Sometimes it is formalized, as in the SEAL–I-Corps Go linkage. This interlinking means that accelerators and incubators are often symbiotic. Incubators use accelerators to identify plausible medium-term ventures to enroll, as well as to offload specialized structured education for the ventures that have already enrolled. Additionally, young ventures often seek accelerators on their own: As SEAL’s program director for ecosystem development told us: “With them it's just kind of like, “I need a program.” You know people who want to sign up for a marathon, but I need to join a running group or a program.” Consequently, ventures must follow the cycles of each program in which they are involved, absorbing different educational programming and mentorship and producing documents, reports, and especially pitches on different schedules for different audiences.

In the literature on pitching, this interlinking across programs has been lightly mentioned (e.g., Spinuzzi et al., 2018), but studies have mainly focused on pitching solely in the context of the incubator or pitch competition (see Sabaj et al., 2020, for a review; e.g., see Cabezas et al., 2020; Galbraith et al., 2014; Spinuzzi et al., 2018; Spinuzzi et al., 2016). Their central focus has been on examining a venture’s more or less linear journey from idea to incubation to exploitation (see Figure 1; cf. Spinuzzi et al., 2020; Vogel, 2017), rather than on accounting for the more complex, networked relationships across the programs through which a venture might pass—programs with varying outcomes, lengths, and cycles.

OPEN IN VIEWER

Figure 1.

The Vogel cycle, with SEAL superimposed (adapted from Vogel, 2017, in Spinuzzi et al., 2020, p. 104).

To better understand the entrepreneurial journey—specifically how new ventures navigate varying programs—we need a framework for understanding how these ventures are involved with varying, often uncoordinated cycles of development. Blank (2013a) has suggested one such framework: the OODA Loop.

The OODA Loop

Blank (2013a) attributed the lean startup’s focus on rapid development and feedback to the OODA loop, which he drew (loosely) from the work of warfare theorist Boyd (2018). As Blank and other lean startup advocates have argued, the business plan is too static, slow, and insular to deal with the rapidly changing landscape of business competition, so it should be replaced with more nimble, descendant genres focused on continual interaction: the pitch, the business model canvas, and the minimum viable product. These descendant genres are meant to offer hypotheses, test them with rapid audience feedback, and quickly iterate them, replacing isolation with interaction. Blank justified the lean startup approach by comparing it with “a U.S. warfighting strategy known as the ‘OODA Loop' articulated by John Boyd and adopted by the U.S. armed forces in the second Gulf War” (p. 27). He promised that “you will use the military concept of OODA (Observe, Orient, Decide, Act) by moving and responding to competitors and customers at a tempo much faster than your competition” (p. 221).

Boyd was a fighter pilot for the U.S. Air Force who later studied as an engineer and then became an autodidact military theorist whose readings included Tzu, Clausewitz, and Mao but also Lenin, Maturana and Varela, Bateson, Polanyi, Kuhn, and Popper. In his biography of Boyd, Osinga (2007) told us that “some regard Boyd as the most important strategist of the twentieth century, or even since Sun Tzu” (p. 3), but “on the other hand, his work has invited dismissive critique.” Complicating Boyd’s legacy is the fact that his body of theoretical work is made up of “four briefings and an essay” (p. 1); the briefings are slide decks that Boyd iteratively revised throughout the 1970s, 1980s, and 1990s. These slide decks were meant to be presented rather than read as stand-alone documents, thus posing a problem for those who want to better understand Boyd’s thought.

Boyd made several contributions—for instance, his Energy-Maneuverability (EM) Theory led to the development of the F-16—but is perhaps best known for his concept of the OODA loop, which depicts a cycle in which an agent (an individual or an organization) must observe, orient, decide, and act in an adversarial environment (see Figure 2). An example is that of fighter pilots, who must observe (sense) their environment, orient to (analyze, synthesize) a threat (an enemy pilot), decide how to address the threat, and act on that decision. The pilot then observes again, beginning a new cycle. More broadly, the OODA loop presented a systemic “model of individual and organizational learning and adaptation” (Osinga, 2007, p. 235), a model that Boyd applied to individual pilots, units, brigades, and entire armies.

OPEN IN VIEWER

Figure 2.

The expanded OODA loop (figure by Patrick Edwin Moran, used under Creative Commons license CC BY 3.0, https://commons.wikimedia.org/w/index.php?curid = 3904554).

The OODA loop, then, is conceptual and not really indexed to a fixed cycle. That is, it is not an operational but rather an analytical aid for separating different things that might overlap. For instance, observers do not stop sampling the environment in order to feed observations forward into an orientation module, nor do they stop orienting and then decide. But they do have to orient based on observation, decide based on an orientation, and act based on a decision. Thus, OODA, like other simplifying schema such as activity systems (Engeström, 2016), provides related analytical concepts for understanding cyclical decisions and acts.

Critically, Boyd argued that “we should operate at a faster tempo than our adversaries or inside our adversaries [sic] time scales … such activity will make us appear ambiguous (non predictable) [emphasis added]” (Osinga, 2007, p. 27). This insight was the kernel that led to his elaboration of the OODA loop. A faster tempo is often better—not to outpace the adversary but to confuse and disorder the adversary. (This outcome can be achieved in other ways as well, such as stealth and deception, but speed is the most salient element in an aerial dogfight, the condition under which Boyd began developing his theory.) In this antagonistic scenario, the goal is disrupting the adversary’s understanding of the problem: changing the state of things between the adversary’s OO and DA while simultaneously guarding against the adversary’s disrupting your own OODA loop. If your OODA loop is disrupted, strategic thinking becomes impossible because you are reacting to outdated information.

OODA thus describes a systems model of learning that is tactically oriented, one that is grounded in individuals' experience but that applies to organizations as well. As Osinga (2007) argued, the OODA loop

represents and means more than a decision process, and the model contains more for victory than information superiority and speed. The OODA loop is much less a model for decision-making than a model of individual and organizational learning and adaptation in which the element of orientation—made up of genetics, experience, culture—plays the dominant role in the game of hypothesis and test, of analysis and synthesis, of destruction and creation. (p. 235)

For Boyd (2018), this systems model comes from his reading of thinkers such as Bateson (1972, 1979) and Maturana and Varela (1987/1987/1998) and reflects their emphasis on organisms and environment in homeostasis. OODA models an organism’s position within an environment, with actors pursuing and iterating their own objectives—objectives that may converge.

As Osinga (2007) argued, Boyd saw interactions as binding social systems together (p. 192). He portrayed this binding interaction with the terms “organism” and “environment,” meaning them literally but also scaling them to organizations. He clarified these organism–environment relationships in terms of feedback (p. 234) and mismatches (p. 349), both of which allow the organism to maintain equilibrium with the environment. Without continual interaction with the environment, the system will collapse: The system's internal instabilities can be addressed only through interactions with the environment (p. 330). Thus, Boyd (2018) saw adversaries as being locked in a strategic game of “interaction and isolation … in which we must be able to diminish [an] adversary’s ability to communicate or interact with his environment while sustaining or improving ours” (p. 286). In this adversarial game, “interaction permits vitality and growth while isolation leads to decay and disintegration” (p. 284).

That interaction, then, could happen across OODA loops, and interferences across these decision loops could induce isolation. Boyd (2018) noted that, in military strategy, each level of a military’s hierarchy has its own OODA loop, which interlocks with loops at other levels:

Each level from simple to complex (platoon to theater) has their own observation-orientation-decision-action time cycle that increases as we try to control more levels and details of command at the higher levels. Put simply, as the number of events we must consider increase [sic], the longer it takes to observe-orient-decide-act. (p. 90)

Since this interlocking can slow or disrupt decision making, Boyd advocated for centralizing command but localizing control. In this approach, decision makers have a common outlook (p. 92) in which the “what” is agreed on but the “how” is left to the officer's discretion (p. 94).

Entrepreneurs such as Blank (2013b) have seen a parallel with entrepreneurship, which (perhaps like warfare) must take in feedback about the environment, strengthen itself, and prevent mismatches. For example, technology entrepreneurs want to “fail faster” (i.e., rapidly gather and adjust to feedback, by testing and discarding hypotheses in a low-risk context), using pitches, business model canvases, and minimum viable products instead of a more static 5-year business plan.

The OODA framework, however, has been criticized as simplistic and mechanistic. As Breton and Rousseau (2007) argued, the OODA representation of decision making does not adequately address interactions in complex environments. Specifically, they argued that the level of cognitive granularity is too low to identify design requirements for decision-support systems. And although the OODA loop is sometimes portrayed as having multiple loops, it still suggests a unidirectional sequence, one that is not dynamic enough to account for ongoing decision making and does not adequately illustrate iterations within and between phases (p. 243). Bryant (2006) similarly argued that the OODA loop overemphasizes data collection, ignores the role of “top-down” cognitive processes for making sense of perceptions, “does not hint at the necessary dependence of perception on preexisting knowledge and concepts” (p. 186), and provides no explicit role for plans, intentions, or goals. Because it attempts to analytically separate two actions that are empirically inseparable—observing and orienting—the OODA loop cannot account well for sensemaking (p. 187).

Applied to entrepreneurship, these limitations mean that OODA is oriented to tactical competition rather than strategic cooperation and cocreation, which entrepreneurship also relies heavily on (Vargo & Lusch, 2004; cf. Spinuzzi et al., 2016, 2018). Put differently, while OODA can describe the tactical, reactive aspect of entrepreneurship, which involves testing and comparing in order to make microadjustments, it cannot necessarily describe the movement toward a strategic objective, which involves developing road maps and identifying a vision for the venture (a limitation we discuss further in the Conclusions and Implications section).

Nevertheless, the OODA loop gives us a way to conceptualize destabilizations that occur through rapid conflicting feedback—an advantage that it has over other social theories that have been applied in technical communication research, such as activity theory (cf. Spinuzzi 2017), actor-network theory, and distributed cognition. In this study, we apply OODA to the rapid conflicting feedback and resulting destabilization that entrepreneurs experienced in interlinked entrepreneurship programs, centering our investigation on the public go or no-go decision that these entrepreneurs were asked to make on Demo Day.

Site

Our study sites include SEAL, Regional I-Corps Go, and other programs.

SEAL. SEAL is an annual summer program designed to help student teams identify and address threats to their new technology-based ventures (Spinuzzi et al., 2018; Spinuzzi et al., 2020). In SEAL, teams examine market interest, technology fit and function, and the ability to create a differentiated value proposition. These teams identify key challenges, test business and technology claims in the marketplace, define their value propositions, and communicate their decision to launch (go), stop development (no go), or change strategy (pivot). During SEAL, teams explore whether a business can be built around a technical innovation (rather than start with a business problem and develop a product to address it). That is, rather than incubating the technology, SEAL incubates the concept of converting the technology into a business proposition. Firms then leave the program with clarity on whether the resulting business is worth pursuing (see Figure 1). Although SEAL originally served just undergraduate student teams at the University of Texas (including students who had been involved in UT’s Longhorn Startup), in 2017, it opened the accelerator program to teams from other universities and teams that included university faculty (see Spinuzzi et al., 2018). SEAL is just one program run by the ATI. In 2018, the ATI also ran the regional I-Corps Go program.

Regional I-Corps Go. Regional I-Corps Go, a 3-week program supported by National Science Foundation (NSF) funds, taught the fundamentals of “LeanLaunchpad methodology” to “early-stage startups,” according to a one-page document that ATI sent to potential entrants (see Table 1, Document D09). These start-ups had to be either unaffiliated with a university or based on intellectual property that a university did not own (although the start-ups could be university spin-outs), and they had to be SBIR Phase 0 teams — that is, they planned to write SBIR proposals to fund their technology development. A regional I-Corps Go flyer (see Table 1, D10) stated this:

During the program, teams will be introduced to the fundamental I-Corps principles, helping teams explore the potential value of their research or innovation, and quickly and effectively validate their commercialization strategy. … Teams will be introduced to the I-Corps approach, and learn about business model development and the customer development process. Teams will also spend time outside the building, talking to customers, partners and competitors, and testing hypotheses. At the conclusion of the program, teams will present their findings from the customer development process and receive real-time feedback from the I-Corps teaching team.

OPEN IN VIEWER

Table 1.

Data Collected for This Project.

Category	ID	Data	Number
Interviews	IIT 01-11	Initial interviews with teams (see Table 2)	11
	FIT 01-11	Final interviews with teams (see Table 2)	9
	ID 01-03	Interviews with program directors (see Table 3)	3
Video pitches		Initial pitches	18
		Final pitches	17
Documents: SEAL program	D01	SEAL kick-off pitch schedule	1
	D02	Final SEAL kick-off pitch schedule	1
	D03	Financial modeling (slide deck)	1
	D04	IP law (slide deck)	1
	D05	Sales, marketing and go-to-market strategies (slide deck)	1
	D06	SBIR and STTR grants (slide deck)	1
	D07	Attendance	1
Documents: I-Corps	D08	I-Corps (SEAL) report.docx	1
	D09	Go_onepager.docx	1
	D10	Regional program flyer_UT winter 2018	1
	D11	Flyer_go spring 2018.pdf	1
Documents: Summary reports	D12	2018 SEAL summary.docx	1
	D13	I-Corps Go (SEAL) report.docx	1
Surveys: Mentor feedback	S1	2018 ATI SEAL-I Corps kickoff feedback form responses.xlsx	1
Surveys: Team feedback	S2	weekly surveys	5
	S3	Summary	1
Observation notes	O0	Kickoff	1
	O1	Financial modeling	1
	O2	IP law	1
	O3	Maker spaces, prototyping and engineering	1
	O4	Sales, marketing and go-to-market strategies	1
	O5	SBIR and STTR grants	1
	O6	Fundraising	1
	O7	Storytelling	1
	O8	The Austin entrepreneurial ecosystem, other accelerators and resources	1
	O9	Decision day	1

Two separate regional I-Corps Go programs (in January–February and in April–May) sourced a total of 10 teams—which were also expected to participate in a 7-week national I-Corps program in Fall 2018—but between those two programs, they were required to participate in SEAL 2018: “Teams will participate in a 12-week summer program aimed at helping teams develop pitches and obtain funding. 1–2 in-person days each week in Austin are required during this time” (see Table 1, D11). ATI linked the two programs to reduce the time, effort, and funding needed to run separate programs. Furthermore, Firm 5 had participated in two regional I-Corps programs; Firm 8 (a firm in the ATI portfolio, i.e., a longer term resident of the incubator), which had dropped out of I-Corps, planned to participate in programs such as Cleantech Open and entered SEAL 2018 in preparation; Firms 9 and 10 (both I-Corps teams) had already participated in Cleantech Open before SEAL; and Firms 4 and 9 had not participated in I-Corps Go yet but planned to do so soon. After the end of SEAL 2018, the program director described how this interlinking complicated SEAL:

Historically, the first eight years, or the first, seven years, it was all UT-only teams. Then for two years, it had teams from other universities but they're still college students, or very recent grads. This year, the bulk of the teams were not students or recent grads, or faculty members. In some ways, an entrepreneur is an entrepreneur, especially at the early stages, but it's a different feel when you have—yeah.

I-Corps Go entrants are thus different from traditional SEAL entrants in terms of affiliation (nonstudent, nonacademic), experience (typically existing firms), and funding sought (SBIR as well as investors). Further, unlike SEAL entrants, I-Corps Go firms were committed to the national I-Corps Go program after SEAL, so they already had made a “Go” decision. As the program’s director for ecosystem development told us later, “Some of them are down the road too far” (beyond the go or no-go decision) and need an incubator, not an accelerator. On the other hand, the SEAL 2018 director argued that “there’s a series of go–no-go decisions” in any business, so SEAL would still be applicable.

In addition, I-Corps Go assumed that firms would spend the time between regional and national programs in customer discovery: “Using the LeanLaunchpad methodology, I-Corps teaches entrepreneurs how to ‘get out of the building' and talk to customers to identify the best product–market fit” (see Table 1, D09), so these entrants had a better understanding of customer discovery than did traditional SEAL participants, according to the director of the circular economy and materials portfolio. At the same time, I-Corps Go teams were expected to interview at least 100 potential customers while participating in SEAL, which did not focus specifically on customer discovery but rather on business fundamentals (see Table 1, D01-D07, O0-O9). As the SEAL 2018 director reflected later, “There’s a lot put on the back of SEAL this year.”

Other Programs. As the director for the circular economy and materials portfolio alluded, entrepreneurs often sign up for programs, just as marathon runners often sign up for running groups and running programs. One I-Corps Go entrant told us that he had signed up for 37 pitch competitions before entering this program. Another entrant attempted to enter a local pitch competition for one idea while developing another idea for SEAL.

Data Collection

This program was declared exempt by the Institutional Review Board of the University of Texas. We collected the following data for this project (see Table 1):

Interviews with team leads: initial audiotaped interviews with 11 team leads as well as final audiotaped interviews with leads of 9 of these 11 teams. (The other two teams did not respond to requests for a second interview.) Initial interviews occurred in the first 2 weeks of the program whereas final interviews occurred in the last 2 weeks.

Sampling

SEAL involved 21 teams. We selected 11 of these teams to examine (see Table 3), taking a striated sample of every other team from a table ordered by investment potential as assessed by mentors (IIP score; cf. Spinuzzi et al., 2018). This sampling strategy yielded a mix of different investment potentials as well as participants in I-Corps Go (6 teams) versus outside I-Corps Go but in the local accelerator (1 team) versus traditional student teams (4 teams). ² We reduced data further by coding them and investigating specific themes.

OPEN IN VIEWER

Table 2.

Program Leads.

Participant	Title	Interview Length (Minutes:Seconds)
Director 1	Director, SEAL 2018	23:26
Director 2	Director, circular economy and materials portfolio	50:00
Director 3	Director, biosciences portfolio	24:19

OPEN IN VIEWER

Table 3.

Selected Teams Participating in SEAL.

Firm ID	IIP Score	Sector	Description	Interview 1 Length (Minutes: Seconds)	Interview 2 Length (Minutes: Seconds)	I-Corps?
01	2.00	Finance	Credit scoring and lending services	18:20	21:48	No
02	2.60	Transportation & mobility	Predictive analytics simulator for traffic issues	16:48	37:40	Yes
03	2.67	Health care	Mobile-first health informatics system for developing communities	14:06	--	No
04	3.00	Bio/health sciences	Monitoring of electrosurgical hemostasis process	11:15	23:07	Yes
05	3.17	Bio/health sciences	Preventing vehicular heat stroke of infants left in cars	12:24	14:34	Yes
06	3.25	Bio/health sciences	Remote behavioral monitoring for families and caregivers	22:31	14:45	Yes
07	3.50	Miscellaneous services	Citation machine for law documents	11:10	15:30	No
08	3.60	Water & food	Smart submetering for water	10:24	8:21	Yes
09	4.00	Energy	Heat pipes for server racks in data centers	22:27	10:25	No a
10	4.25	Water & food	Produced water filtering and cleaning	21:57	--	Yes
11	4.50	Miscellaneous services	Software to capture employee feedback on hiring candidates	12:09	17:00	No

Note. IIP scores are ordered from lowest investment potential (2.00) to highest investment potential (4.50).

^a

This team was in a technology incubator and had initially been in I-Corps Go but dropped out. The SEAL 2018 program director remarked, “I'm not sure how much incremental value SEAL provided versus ATI, but they're going to go through Cleantech Open. They were looking for that value.”

Data Coding and Triangulation

We transcribed all interviews, then coded and triangulated the data.

Coding. We limited coding to three data sets: (a) initial interviews with firms, (b) final interviews with firms, and (c) interviews with program leads. Coding was nonexclusive. Cochran coded entries under Spinuzzi’s direction, initially using descriptive starter codes (Miles et al., 2014) based on central concerns, specifically teams ‘experiences and expectations of the program. Once starter codes were applied, Spinuzzi performed open coding (Corbin & Strauss, 2008) to inductively identify recurrent themes related to teams' perceptions of the program, its clarity, its criteria, and its central go or no-go decision. The Appendix shows selected starter and open codes and gives an example of each.

Once we identified themes in codes, we used other data sets to confirm and illustrate them, examining how observations, documents, surveys, and pitches related to the interview data.

Triangulating. We also triangulated data sets, comparing interview statements with each other and with other data. Specifically, we triangulated firm interviews with

each other

Decide

Even SEAL’s program directors seemed to have different ideas about what the program offered in decision points: Director 1 asserted that the program was always applicable because firms faced an infinite number of decision points: “There's a series of go/no-go decisions that you make for infinity. … It's just a matter of winding up; is there a near one coming up that this type of structure can help the team tackle?” In contrast, Directors 2 and 3 thought that SEAL was more applicable to firms that had not yet begun customer acquisition (i.e., had not turned their venture concept into a venture opportunity to be exploited; see Figure 1).

Despite SEAL’s go or no-go premise, firms could and did make various decisions at the end. But firms did not necessarily distinguish between these decisions, sometimes characterizing themselves in multiple categories—the different decisions were not necessarily well bounded. These decisions included three that were mentioned in the SEAL program documents (“go,” “no go,” and “pivot,” as described in introductory slide decks for kickoff and Demo Day) as well as three other outcomes: “We’ve already been funded,” “We can’t say,” and “Nobody told us.”

Making a Go Decision. Across SEAL, 12 of the 21 teams made explicit “go” decisions (including six of the teams in the study sample). These included nine of the 11 I-Corps teams (including four of the six I-Corps teams in our sample: Firms 4, 5, 6, and 9) and three of the 10 non-I-Corps teams (including two of the five non-I-Corps teams in our sample: one student team [Firm 11] and one nonstudent team [Firm 8]).

The one student team in the sample that made a go decision, Firm 11, explained that they had come into SEAL with “a pretty polished product” and that they had spent the summer “figuring out our sales approach” and developing a focus in communicating its value proposition (Interview 2). They had declared at the beginning of the program that they intended the firm to be a go (Interview 1), adding that they had not even realized that the go or no-go decision was part of the program’s framing. They were already beyond their go decision. They had planned to test their business and develop an expansion plan.

The one nonstudent, non-I-Corps team, Firm 8, had also not seriously considered a no-go decision because it was already in the Austin Technology Incubator and was using SEAL as an opportunity to prepare for another competition, Cleantech Open, and further develop a business that had already been established. Like Firm 11, Firm 8 had already passed the go decision and was obligated to it due to its investors, customers, incubator, and commitment to Cleantech Open. A team member expressed confidence that the firm had validated its venture concept and was ready to exploit it: “It's a Go for us because … we've done enough validation work. We're far enough along and have done enough of the work to feel like we're confident and that we can take this to market and be successful in one of the segments” (Interview 2).

The four I-Corps teams in our sample with a go decision all claimed that the SEAL go decision was by default because SEAL was a step toward what they considered the actual go or no-go decision, which would take place in the I-Corps national program. A Firm 5 team member put it like this:

I'm confident I'll be able to say we have a green light, we're going to continue, go through this national program, and that we're almost going to have another red light/green light after nationals. That's really our true red light/green light test. … I don't think a single team will say red light just because I heard [a SEAL director] talk about this. … We had the highest or the biggest number of funding teams, highest funded cohort ever of SEAL. That was mainly because of the NSF teams. It's really hard to say a no-go for NSF teams because nationals hasn't started yet, and the point of nationals is to figure out the best path to success. (Interview 2)

For these teams, then, the SEAL go decision was really a false decision, one that was overridden by their commitments to I-Corps.

Making a No-Go Decision. In contrast, only one of the 21 SEAL teams explicitly made a no-go decision: Firm 3. Firm 3 was a student team proposing a new product to be marketed by an existing firm (one that was successfully developed in a previous year of SEAL). Thus Firm 3’s no-go decision applied to its proposed new product rather than the firm itself.

But we might also count as making no-go decisions the three student teams that “just kind of imploded” (Director 2 interview). One of these teams was Firm 7 from our sample. The program director remarked that in this case, “the pressure test of SEAL actually worked on the team dynamics.” This firm intended to decide go as late as our second interview, the week before the Demo Day pitch. But by Demo Day, our interviewee concluded that her business partner simply was not as committed to the firm as she was. The program director viewed these implicit no-go decisions as a positive outcome—an outcome that SEAL was geared to produce. It was better for them to find out now than later, once they had sales and commitments.

All of the no-go teams, then, were student teams. For them, the end of SEAL marked the potential end of the journey, a decision about whether or not to execute their venture (see Figure 1). Since they did not have upcoming commitments, as the I-Corps teams did, they could make a genuine decision.

Making a Decision to Pivot. A pivot involves team members' rethinking key parts of their firm before continuing and may include fundamental moves such as reorienting to a different product or market or moving to a new business model. About SEAL 2018, Director 1 told us that “there are three pivots that were noticeable. Lots of tweaks, I'm sure, but there’s only three top-level pivots. Out of 21, that's what I remember. … Or maybe four.” We counted three.

Two teams explicitly announced a pivot in their Demo Day pitch: Firm 1 and 106, both student teams. Firm 1, which was in our study sample, pivoted—in the view of the SEAL directors—because the team was large and had weak internal cohesion and an insufficiently developed vision of the proposed service. But we had trouble understanding Firm 1’s explicit reasoning, and we agree with Director 1 that “the presentation at the end was incoherent.”

One team implicitly pivoted: Firm 2, an I-Corps team, which Director 1 characterized as “way too early.” As a member of Firm 2 told us, “We've already pivoted twice, and we've sort of pivoted a third time. We don't know what's going to happen. It's a go in that we're going to continue customer discovery, but we don't know what's going to happen in the end once we're finished with the I-Corps Go process six weeks from now.”

Director 1 remarked that it was harder for more established firms to pivot:

Some of the companies have raised enough money [emphasis added] where they're going down the pathway and they've already convinced themselves of what they're doing. They've had investors validate that. It's harder to iterate and harder to pivot at that point in time.

We think the emphasized points are key. A firm pivots when it finds that its envisioned configuration does not sufficiently address the market, problem, or other needs—when that configuration has been tested and found wanting. But as Director 1 argued, many firms were not in that position. On one hand, some firms had already had their decisions validated: They had raised money and thus had committed to a path. On the other hand, I-Corps teams by definition could not make a true pivot decision yet because they were supposed to be validating their configuration via customer interviews—interviews that should have been completed before they entered the national I-Corps program. The Firm 2 member alluded to the latter issue when saying, “It's a go in that we're going to continue customer discovery.”

We’ve Already Been Funded. Two teams, neither of which were in the study sample, received funding during SEAL, making the go or no-go decision moot—although Firm 110’s slides also declared a Go. Firm 110, an I-Corps team, outlined how it had managed to grow strictly through sales rather than relying on investment; its concluding slide was oriented to other participants, telling them, “Key lesson: You don’t need investors to crush it.” Firm 107, a student team, did not bother showing up for Demo Day: After receiving investment early in the summer, the firm ceased all SEAL activities (interviews with Directors 1 and 3).

We Can’t Say. Again, these decisions are not exclusive. Two teams, both in I-Corps and both in our sample, declared in their second interviews that they could not really make a decision, and they both avoided claiming a decision on decision day. Members of Firm 2, which pivoted, also declared that they could not make a decision because they were committed to the national I-Corps program:

We're going to find out if this product is worth pursuing or not. We're going to find out more about how you would sell it. … We're in the stage with I-Corps. We're supposed to not know what things are yet. We're still supposed to be learning. If we say we know what we've got, then we're not doing the program right. … We don't know for sure yet if we're going to have something that makes money or not because we're not going to do it as a hobby. (Interview 2)

Firm 10 similarly claimed that the go or no-go decision would be determined in the national I-Corps, not in SEAL, so its presentation would outline “next steps” (which it did).

Nobody Told Us. Finally, three of the firms in our sample said that although they would declare a decision, they had not realized that SEAL’s core rationale was the go or no-go decision. (SEAL’s advertising said that SEAL participants would “tackle the most difficult, deal-killing questions of their new ventures,” but it did not explicitly lay out the end goal of a go or no-go decision.) For instance, although student team Firm 11 declared a go decision, one team member remarked that “maybe it would have been different if our expectation originally had been set to ‘This is the no-go or go.’ When that was said on the first day I was like, ‘Oh, oops.’ … the fact that we didn't know this was a ‘no-go/go.’ If that's the core tenet of the program, that's surprising that we had no idea” (Interview 1). And a member of I-Corps Firm 6 said, “I don't even think they asked for a go/no-go discussion” (Interview 2). Similarly, a member of Firm 5 told us, “I know NSF was kind of test-piloting this. I had no idea what I [was] getting myself into. They just were like, ‘Hey, we got this cool partnership with SEAL.’ They didn't tell us what SEAL was. … Then, we sit down the first day and they are like, ‘OK, it's red light/green light, mini-accelerator’” (Interview 2).

Observe and Orient

In SEAL, firms attempted to observe how the program worked and how it related to other programs. Yet these observations were often fragmentary, and firms had different models in mind. Student teams in our sample (Firms 1, 3, 7, 11) had previously been involved mainly in university-based programs and thus had only a vague idea of how the I-Corps firms were involved. But I-Corps teams in our sample (Firms 2, 4, 5, 6, 9, 10) did not seem to understand the relationship either:

I didn't know much about the SEAL program. My point in going to the SEAL program was, it was part of the NSF I-Corps program. What the UT did was they have a hybrid where chemically the I-Corps was mainly for the academic teams. They want to expand it to nonacademic teams. What they did is they created a power program called I-Corps Go. They're mixing the NSF grant for successful regional teams to also go through the ATI SEAL program. We were automatically enrolled in this. (Firm 10, Interview 1)

As a member of Firm 9 said, “a lot of the companies or ideas coming into SEAL are in a significantly earlier stage than we are” (Interview 2). And Firm 6 interpreted SEAL as serving “potential start-ups, fledglings like ourselves” (Interview 2). Earlier, a Firm 6 member had also asked the interviewers, “Has SEAL been exclusively for UT students? I'm sure it's developed for the diversity of the folks going through SEAL” (Interview 1).

The firms' understandings of SEAL were partly based on their own experiences and journeys. Firms mentioned participating in a total of 21 named programs, accelerators, incubators, events, and entities—as well as the unspecified “37 pitch competitions” in which Firm 2 had participated (see Figure 3). Firms oriented to SEAL in relation to these different experiences. On one hand, student teams understood SEAL as a student-focused program and were confused about why it involved so many nonstudent teams: “One thing that struck me really hard, I was so confused, is that SEAL is a student entrepreneur launch accelerator. We are probably the only one or two student teams. I'm confused. What the heck is going on?” (Firm 1, Interview 2). On the other hand, I-Corps teams did not have a strong idea of how SEAL was meant to help them, with one team bluntly declaring that “They put us into [SEAL] as a placeholder” (Firm 9, Interview 2). As Figure 3 shows, several firms—including the six I-Corps firms and the one ATI firm in our sample but none of the student teams—followed a pathway from one experience to SEAL to another experience (bold arrows). Thus, they interpreted SEAL as part of a larger, structured journey.

OPEN IN VIEWER

Figure 3.

Programs, accelerators, and incubators in which the SEAL firms reported participating.

Patterns of Conflict Across Programs

In one OODA briefing on the patterns of conflict, Boyd (2018) advocated for grand tactics to be used against adversaries in warfare. One of these grand tactics is to “enmesh adversary in an amorphous, menacing, and unpredictable world of uncertainty, doubt, mistrust, confusion, disorder, fear, panic, chaos . . . and/or fold adversary back inside himself” (p. 152). Here, Boyd was describing deliberate isolation of an adversary from its environment, a move that makes decision making impossible. But as our results suggest, some firms in SEAL similarly had trouble making a specific decision, and they similarly described encountering “uncertainty, doubt, mistrust, confusion, disorder” (although none of them mentioned fear, panic, or chaos). These firms often attributed the uncertainty, confusion, and disorder to SEAL itself: In the summary report (see Table 1, D12), for instance, one firm claimed that “there was a lack of cohesiveness” in SEAL. Firms 1, 2, 5, 6, 8, and 11 characterized SEAL as disorganized or discohesive, ³ with Firms 2, 5, 6, and 8 specifically complaining that pitch criteria were not clear. And as we mentioned, some firms stated that they either had not been told that they should make a go or no-go decision or felt unable to make such a decision due to commitments to other programs.

This inability to make a go or no-go decision—the entire point of SEAL and the principle around which it was structured—is concerning:

The secret sauce behind SEAL is having teams take a hard look at either different projects or . . . their company on the whole and being willing to kill it if there's not a path forward for them. That requires a different level of analysis. If companies are coming in with something built in that incentivizes their participation and not killing it, I don't know how authentic that examination was all the time. (Director 2)

Yet in terms of both their positioning within the program and their stage of customer validation, the I-Corps teams were incentivized not to kill their projects. In fact, more broadly, SEAL’s rationale and structure faced interference from the conflicting intentions of I-Corps and other programs.

The Framing of I-Corps Overrode SEAL’s Rationale. As we have seen, the framing of I-Corps overrode SEAL’s rationale by making a decision other than go impossible. (Although one I-Corps firm did pivot, its members did not declare that pivot in their Demo Day pitch.) So what did the I-Corps teams think SEAL was good for? They interpreted SEAL differently: One (Firm 9) considered SEAL to be an irrelevant “placeholder,” two (Firms 2 and 9) interpreted it as an opportunity to do customer discovery, and one (Firm 6) considered it to be a space for improving pitches and acquiring customers. None described it the way that SEAL itself did, as an acceleration and launch program oriented around the launch decision.

The Framing of Other Pitch Competitions Overrode SEAL’s Structure. All of the firms, both I-Corps and non-I-Corps teams, had some experience with other entrepreneurship programs, most of which were structured as pitch competitions. At least two firms believed that the point of SEAL was to hone a pitch that would be perfect for any rhetorical situation. As mentioned, a Firm 9 team member said he had honed his pitch over 37 such competitions, using A-B testing to determine what words and gestures he used. He clearly thought that a pitch should be in one genre and could be honed for all audiences. Similarly, Firm 8 wanted to see “a consistent model [of pitch] of a single topic” and complained about not getting “consistent feedback,” having never been offered a workshop “that was just specifically on getting it to a final pitch to where it's perfect.” Unfortunately, this arhetorical understanding of the pitch genre does not support the different functions and audiences to which pitches must be oriented.

Firms 2, 5, 6, and 8 complained that SEAL’s pitch criteria were not clear. In addition, Firm 11 had not heard about the decision orientation of the pitches. Firm 2’s two participants discussed the reactions to their first pitch: Firm2A:

I found that the questions [on which the pitch was evaluated] were disjointed from what we were asked to do. What we were asked to do was not an investor's pitch, but the questions that they were asked to evaluate were based on an investor pitch. … Some of the people just wrote, “They aren't even supposed to do this yet.”

Firm 6 complained that “looking at what was sent out in terms of what to prepare to present didn't really match up with what it looked like the judges were judging on” and that they were “not sure, on the first pitch, what they were really asking. It turned out that what they asked everybody to write up was very different than what we talked about.” And Firm 5 complained that “all they told us initially was there's this seven to eight-minute pitch and to pitch. … It wasn't until the night beforehand we got something, then they extended the deadline to two more days. But what he wanted us to pitch on was just four things” (Interview 1). In short, absent explicit advice about how to structure their pitches, the firms based their decisions on their experience with other programs—programs with different orientations.

Conclusions

To an extent, we agree with the teams that the SEAL program was described in confusing and unclear ways. SEAL’s messaging was not strong in its materials—although it did mention having teams make decisions, and the go or no-go decision was implied in that the last letter of its acronym stood for “Launch.” But SEAL had used similar messaging and had structured its program in substantially the same way in previous years, and firms in those years seemed to be less confused about its orientation. The main difference in 2018 was that, for a substantial number of teams, SEAL was positioned as part of I-Corps Go. In retrospect, its go or no-go decision was incompatible with that larger program.

By applying the OODA loop, starting with the visible decision point, and tracing back to the observe and orient points, we found that teams had trouble making the envisioned decision because their participation in other programs (especially I-Corps) made that decision impossible. Furthermore, the orientations that the teams learned in other programs did not match SEAL’s orientation. Teams reacted by attempting to interpret SEAL as part of a larger journey across programs. In doing so, they encountered difficulty in understanding what their pitches were supposed to do, what criteria they should meet, and how they would be evaluated. In OODA terms, they had trouble “orienting” appropriately, leading to different or delayed “decisions” due to the interference between the differing goals and cycles across the linked programs. As a Firm 6 member said, “They're happening over and over again, and they're happening all together.” And because the entrepreneurs needed to be oriented to these multiple goals and cycles, they had trouble interpreting the decision point and therefore could not effectively communicate and justify their SEAL decisions.

The teams also lacked internal metrics for making that critical decision. What was the minimum metric for making a go versus a no-go decision? Ultimately, the teams based their decision on their willingness to “soldier” on—not a criterion, heuristic, or metric. Thus, for SEAL, the pitch described a decision but not a direction: It answered questions such as Where are we now? and What do we propose to do next? but not Why should we continue to exist? or What will we ultimately accomplish? The data were not funneled into a decision process in which the decision resulted from external data such as market interest or acceptance. Although the teams did successfully use the OODA process, they used it in a continuous loop, without strategic direction.

Consequently, entrepreneurs in SEAL engaged in cycles of small, tactical adjustments in the venture: comfortable microadjustments oriented to program mentors and materials, largely yielding comfortable go decisions based on successfully following the SEAL process. In contrast, larger strategic adjustments require feedback directly from the markets that the entrepreneurs seek to engage—feedback such as market interviews (e.g., the 100 interviews that teams had to conduct for I-Corps Go) or a robust market report (e.g., Spinuzzi et al., 2016). Small, tactical adjustments may have fit SEAL’s goal to “stress-test” ventures before they entered the exploitation state, but such adjustments did not match I-Corps’s objective of identifying larger (and less comfortable) strategic macroadjustments—nor did they serve other accelerators' focuses on issues such as funder interest. The programs shown in Figure 3 had different cycles in part because they had different objectives.

Implications

Up to this point, studies of entrepreneurship training programs have examined these programs separately. Consequently, such studies have tended to focus on a bounded set of objectives and how the programs meet those objectives. But as we discussed, entrepreneurs often go through multiple programs, some simultaneously. These entrepreneurs may have difficulty understanding how program goals differ—or even how pitches themselves orient to different purposes, stages, and audiences. In this study, teams' previous and overlapping entrepreneurship programs demonstrably impacted how the teams interpreted and addressed the goals in the current one. Because this is not a novel situation—entrepreneurs, like runners, tend to sign up for many programs (Director 2)—studies of entrepreneurship programs need to take such interferences into account.

In this study, we used OODA to analyze these incidents of interference, partly because it is already used in the lean start-up and partly because it was developed with such interference in mind. In such incidents, OODA provides a model and conceptual language for describing tactical interferences across cycles. It does not, however, provide an adequate apparatus for analyzing strategic decisions or understanding meaning making in social or cultural terms. For instance, the OODA model relegates “cultural heritage” to the orient part of the loop, along with “genetic heritage” and “analysis and synthesis.” This role seems radically limited and underexplored, given what we know about how cultural and social expectations impact observations, decisions, and actions. In subsequent studies, OODA could be paired with a social theory with these strengths in order to provide more insights. Doing so would likely entail developing rapprochement across the theories so that their insights could be gracefully paired. Some early work has been done to put OODA in conversation with one such theory, activity theory (Dias et al., 2018; Gonçalves et al., 2013), but the work is not yet developed to the necessary degree.

Additionally, we studied the question of interference (primarily through interviews) in a central case that was linked to other programs. Subsequent studies should, if possible, investigate connected programs in a more thorough way, ideally involving observations and artifact collection in two or more programs. More broadly, studies of business communication and persuasion may need to take into account the interfering cycles that affect the activity under study.

Finally, based on these findings, we suggest that accelerators explicitly lay out their objectives and how these objectives might relate to other entrepreneurship programs. Because entrepreneurs undertake many programs simultaneously or serially, they need guidance in how to interpret relations between and objectives across these programs. Beyond this framing, entrepreneurs could benefit from regular, specific goals that are explicitly oriented to the program’s tactical or strategic objectives.