Meaningful Innovation: Ethnographic Potential in the Startup and Venture Capital Spheres

Aiming for Disruption

Innovation is defined in different ways depending on the context and the author. At its most basic level, innovation refers to newness in form or approach (Van de Ven 1986). There have been countless studies of innovation, viewing the concept both from a product perspective (the innovative outcome or result) and from a process perspective (how one innovates) (Bundy 2002). Here, I want to unpack the two more closely.

From a product, or result, perspective, there are many types of innovation, discussed along a variety of spectra. But the most discussed and frequently debated conceptualization of innovation in the current technology sphere is disruptive innovation — those innovations that disrupt an existing market or displace an earlier technology. Though the phrase disruptive innovation was coined by Clayton Christensen in the mid-1990’s (Christensen 2013), the concept of innovation through ‘creative destruction’ was first popularized through economist Joseph Schumpeter’s theories from the 1940s (Schumpeter 2013). Schumpeter described this ‘revolution from within’ as an inherent part of innovation. In today’s startup sphere, this focus on disruption is very much considered the ideal. I will note here that much research has noted the dangers and downfalls of a focus on disruption, and, more broadly, that innovation does not mean the changes it creates are necessary or positive (Abrahamson 1991). I will take up these issues again later. What is important here is that disruptive innovation is currently a main goal for startup founders and the venture capitalists who fund them alike. According to Paul Graham, founder of Y Combinator, startup ideas should be risky and actually even repel you (Graham 2012). There is a pervasive attitude that startups should be disruptive, contrarian. This focus foregrounds all that follows.

This focus on disruption, then, creates a central challenge for the startup community writ large. The startup sphere has a tendency to aim for innovations that will have a revolutionary impact to the market — disruptive innovations — but at the same time, many of the innovation processes emphasize iterative, small feedback loops to refine ideas and products, based on feedback from those who would use the product– early adopters. This creates a sort of dichotomy of focus on revolution and evolution from the outset. I frame this dichotomy in terms of Norman and Verganti’s (2012) discussion of radical versus incremental innovation. By their definition, radical innovation is “a change of frame (‘doing what we did not do before’),” while incremental innovation is “improvements within a given frame of solutions (‘doing better what we already do’).” Norman and Verganti argue that radical innovation is “surprisingly rare” and requires agents of meaning or technology change (2012, p. 6). In their point of view, most innovation is incremental. And this is indeed what I have witnessed in the startup community.

A Focus on Diffusion

Over the past few years, I conducted an in-depth research of technology startups and startup accelerators situated in different locations globally. Much of this research was conducted in situ through ethnographic fieldwork within Silicon Valley and over the several-month courses of accelerator programs in Singapore and Buenos Aires. This provided a rich view into the day-to-day workings of these accelerators and those affiliated with them– founders, funders, mentors, and more. Additionally, I have conducted interviews in Silicon Valley and abroad with a variety of people playing various roles in the startup community. Through this, I have learned the processes and practices of Silicon Valley and global startups—how they innovate. And what I have witnessed is a counterintuitive focus on disruption and diffusion simultaneously.

The process of innovation as it has been conceived historically consists of three general, yet overlapping phases: invention -> innovation (use) -> diffusion (King et al. 1994). There are many approaches from a variety of schools of thought on the flows and influences of innovation that critique such a linear approach and present other approaches, but diffusion is always considered to occur late in the process. I argue that the model of innovation being spread since late 2000’s has shifted that flow, making it looped or even backwards. And by way of the global reach and connectedness of the tech startup community and certain funding mechanisms, this is fundamentally changing the flow of innovation from one that begins with an invention or product to one that begins with a focus on global diffusion. And this inherently changes the types of “innovations” that make it the market.

In the first dotcom boom in the late 1990s, it was very expensive to build a software product. Startups had to have an initial product or an idea that seemed really good. Then they would get investment a priori to build it or to continue building it. With funding, they would then build it and then see how well it diffused. Today, it is cheap to build software, so you can easily start to develop something. But now, in order to get funding, you have to show how it would scale, not just guess how it would. Metrics showing traction and validation are key. This focus is oriented by the Lean Startup model and Customer Development practices, which rely on several concepts that originate from Rogers’ work on diffusion of innovations (DOI).

In Lean, an MVP need not be a functioning product. It can, and often is, still in the idea phase. As Steve Blank, the creator of Customer Development says: “You’re selling the vision.” The idea or experiment is tested with “early adopters,” a phrase coined in Rogers’ DOI work. Validation of the idea relies on tracking metrics and creating a “funnel” of potential customers. The goal is to exhibit the ability to “Cross the chasm,” i.e. move from early adopters to the mainstream market— which is based on the graphical depiction of innovation diffusion over time in Rogers’s work.

This is not the same as going out and researching a market to develop a product. The focus, rather, is on evaluating whether the product will scale before actually fully developing it. The process moves from finding potential early-adopter customers for an idea, to refining that idea based on how they may use the product, to then developing the actual product. The potential for diffusion precedes the innovation.

This flow was not possible until recent years. First, the global, networked platform of the internet has enabled software and digital products to be globally scalable. Secondly, recent technological advances underlie the ability to measure global scalability. One can now test an idea and analyze data in ways previously not possible. Analytics tools like KISSMetrics, Mixpanel, and Google Analytics enable the development of measures to determine demand and scalability. Advertising platforms like Google AdWords and social networking sites like Facebook provide methods to experiment. And the models of Lean and of Customer Development provide the structure to follow. Teams can start with a premise or an assumption of a problem and create advertisements on Facebook or Google AdWords to target potential customers. They can then measure their interest directly based off of clicks and conversion rates and other metrics, all before committing a single line of code in the product.

The common conception of ‘startups’, the canonical literature in this area, and adoption of Lean methods and principles are all products of Silicon Valley. The terminologies and cultural views of this origin are imbued into the structures, practices, and approaches — and this includes adoption of the venture capitalist business model and its underlying goals and objectives. While VCs provide value to the innovation teams by injecting economic capital, VC funds also rely on (and expect) a return of capital via increased valuation and a future liquidity event, commonly known as an “exit.” This enables the VC to provide returns and to fund its future operations. It is a hit business; it makes many small bets with the goal of having at least one team among many become successful. Therefore, by design, the goal of VC operators is to foster an environment that produces large capital gains within a short timeframe. In practice, this translates into a culture within the startup world that promotes creation of globally-focused, scalable, and profitable businesses. It privileges global scalability (diffusion) over what the actual technology is or how it is used.

Seed-stage and early-stage startups are trying to create a product and create a business simultaneously. They are focused on doing something innovative, but also on building legitimacy and showing that they are scalable. Their survival relies on funding, and funding is rooted in both of these. They have to show investors they are building something scalable and have metrics to prove it. They also have to appear legitimate, participating in the culture and practices that are part a startup world. The incentives for mass diffusion shape the direction of the product more than developing an innovative or useful product do. Even if that is not the goal in the beginning, startups often reorient to garner continued investment from their funders, who become their advisors.

Placing the focus a priori and continuously on diffusion fundamentally shapes the types of innovations that are made. Namely, it shapes the focus toward designing technologies that are easily adopted, and that would be adopted broadly. This may influence focusing on broader problems that effect many people. It may also promote the development of products with immediate impact, which can then be built upon, contributing to cumulative innovation (Murray & O’Mahony 2007). But, prioritizing for potential diffusion, but not for radical change, notably also limits the possibility for disruptive innovation.

A Pathology of Startup Innovation

A Lean Methodology – Through my experience and close examination of startups and innovation, an underlying thread that connects many of them is a focus on Lean Startup. Lean methodologies, based primarily on Eric Ries’s works on Lean Startup (2011), still form the core of the practices most Silicon Valley startups— and some enterprises— purport to follow. The work centered on Lean Startup helped spur a startup mania around the globe; it was the first model put forth to describe startup creation as a science and has become dominant in the startup sphere.

The concept of Lean traces its roots back to manufacturing. Lean Production, a term coined in the 1990s at MIT, was initially used to describe the Toyota Production System (TPS) (Holweg 2007). TPS is a socio-technical system that combines a distinct management philosophy and practices. According to the Toyota group that teaches the system, it is based on four core principles:

• put the customer first;
• the most valuable resources are people;
• a focus on the workplace itself; and
• kaizen, meaning “good change” in Japanese, which Toyota uses in the context of “continuous improvement” as a philosophy.

TPS describes itself as “a culture of problem solving at every level of the organization,” and the necessarily skills are learned by doing, not by concept. Following from Lean Production, others have borrowed the term to emphasize a focus on reducing waste and on continuous refinement, although the term is often confused with meaning small teams or low monetary cost.

The term Lean Startup was coined by Ries, who had experience with Agile Development as the Chief Technology officer of IMVU, a 3D social network. Through IMVU, he met Steve Blank, an entrepreneur and investor who created the Customer Development framework. In exchange for investment in IMVU, Blank required Ries to enroll in Blank’s entrepreneurship class at UC Berkeley. Ries was heavily influenced by the course and combined his experience with Agile with Blank’s methodology to create a continuous deployment concept that is heavily influenced by TPS. He primarily focused on two key concepts: Customer Development (Blank & Dorf 2012) and continuous deployment (Maurya 2012).

Lean Startup is thus intended to be a model of innovating rooted in experimentation and rapid iteration. The entire product development cycle is about building a hypothesis, testing, learning, and iterating on it. Using the Build-Measure-Learn framework, a startup theoretically can focus on reducing the time and labor involved in developing the product. In analyzing complexity, fast iteration almost always produces better results than in-depth analysis (Sessions 2006).

The Minimum Viable Product (MVP) is the bare minimum product (or non-technology-based experiment) a team can build to use and test a number of assumptions. By testing this with early adopters, startups can continue to iterate, using Agile development practices such as Scrum or other Kanban (a signboard/scheduling system) principles to conduct short product development sprints. The form of the product may change throughout the Lean process, but the intention is that motivation and overarching vision of the team should remain intact, that is, unless experimentation disproves assumptions. In that case, they pivot. That means they change their focus or strategy based on these learnings. This is where Lean fits into the larger Customer Development framework.

Customer Development is essentially the startup version of user research. Steve Blank, who started Customer Development has defined a startup “[as] an organization formed to search for a repeatable and scalable business model” (Blank & Dorf 2012). His conceptual practice, Customer Development, focuses on this goal and includes four distinct phases: Customer Discovery, Customer Validation, Customer Creation, and Company Building. The first two phases, Discovery and Validation are in a loop, and only when customers are “validated” does the loop breaks off into a linear progression of Customer Creation and Company Building. The process is half of the Lean Startup methodology and parallels the hypothesis-driven Lean Startup approach. The Build-Measure-Learn cycle is centered in the first two phases, and emphasizes validation. Validation is a significant event, although it perhaps gets less and less significant to the core value of the business as iterative cycles continue to improve peripheral aspects. A key event that Lean processes create is forcing a pivot. The team moves into an entirely new direction strategy-wise. Pivoting requires drawing insight from data collected from experimentation, both quantitative and qualitative, and then building a new hypothesis. The methodology also popularized terms like Problem-Solution Fit and Product-Market Fit, creating common startup languages for entrepreneurs, investors, and stakeholders (Blank 2013; Cooper & Vlaskovits 2010). In short, these processes are very focused on incremental improvement, and also oriented toward mass-diffusion.

Beyond Design Thinking – While Lean processes are still dominant, other approaches, particularly from design thinking, have begun to be spread more broadly in the startup community. Design thinking is rooted in combining the context of the problem and empathy—in many ways, not dissimilar from what Customer Development preaches. However, the philosophy and the approach are more grounded in data. This grounded-theory approach provides a structure for patterns to emerge; it allows innovators to arrive at conclusions based on observations early. Thus it sets the stage for alternatives to be examined and experimentation and metrics to be used in a valuable way. In a complex environment, this allows experiments to be more focused on producing understanding, which Anderson et al. suggest is key: “the system is now too complex for a prior [sic] comprehension and thus the product launch is itself an experiment about order or arbiter of order” (2013).

As I have suggested before, combining elements of Lean and design thinking seems to have potential for startups (Haines 2014), and others have noted the potential synergies of these approaches (Müller and Thoring 2012). Lean focuses more closely on business value while design thinking focuses more on developing the right product for the end user. Together, they are useful in terms of thinking about how to provide value for both stakeholders and users at the same time.

But while some “design thinking” methodology has started to make inroads in early stage technology startups, there seems to be much more to explore in terms of approaches. And in particular, it seems we need to go beyond the simplistic tool-based discussions. For startups, it is important to consider other practices that are not currently mainstream within the startup world. As Norman and Verganti suggest, a key to pursuing radical innovation seems to be in conducting research that helps interpret what is meaningful to people—or what could be meaningful. It seems that ethnographic approaches—and ethnographic thinking, as Hasbrouck has articulated (2015)— are uniquely suited to help teams understand meaning and shape meaning through new technologies or business models. In other words, ethnography is well oriented to support disruptive innovation through developing and asking the right questions, pursuing ongoing inquiry, and providing rich interpretations.

A Pathology of Venture Capital

How VC Works – Access to capital is especially important at all stages, but especially the seed stage of a technology venture, and Silicon Valley provides the most VC funding globally (Kramer & Patrick 2014). This funding occurs through established VC firms and corporate VC arms, as well as through wealthy individuals known as “angel investors.” The origins of this began in the 1950s, when “The Group,” a small network of young investors began pooling investments in technology-focused startups in Palo Alto, becoming key players in the growth of the VC community there (Kenney 2000). By the 1970s, successes had spurred greater investment, and successful entrepreneurs started to become venture capitalists after exiting, organically creating the “virtuous cycle” of capital funding in Silicon Valley. This continues today, with examples such as the PayPal Mafia, a group of former PayPal employees who went on to invest in (and found) many other successful startups (Lacy 2008).

This virtuous cycle explains why there is so much more seed capital available in Silicon Valley than other places, where people tend to put money in stocks, bonds, real estate, or other more stable ventures. VC investments are notoriously risky, with average returns well underperforming other investment vehicles (Huntsman & Hoban 1980). Entrepreneurs-cum-investors, however, tend to have a higher risk tolerance, which in theory helps spur innovation at a much higher rate than corporate R&D (Kaplan & Lerner 2015). According to previous research, VC tolerance for failure help startups “overcome early difficulties and realize their innovative potential” (Tian & Wang 2011). VC firms bridge the gap between entrepreneurs and financiers, dealing with “moral hazard” and asymmetric information (Lerner & Tag 2013). They screen startups, create contracts, stage funding, and closely monitor and often advise the startups they fund. But the way they make money underlies some of the main challenges the startup world is facing.

Venture capitalists make money through financial engineering. They find companies that are undervalued, that have great potential for increasing in value dramatically. They invest at a low price, and eventually aim to sell stocks at a high price as the company garners more investment, grows, and possibly exits – getting acquired or going public. Other, recent changes to structuring capital have also enabled angels to make more investments. Convertible notes enable them to invest in the form of a loan that converts to equity once another round of funding is raised (Feld & Mendelson 2011). This encourages investing in very early rounds, when the valuation of a startup is uncertain and unstable. About 80% of a VC funds’ returns come from 20% of it’s investments (Rachleff, 2014). Increasingly, for investors, it is a “home run” game. They depend on outliers, investments with the potential for extreme returns within five years, usually larger than ten times the initial investment. Essentially, VC funds make lots of small bets in hopes that one will become the next “unicorn” among the likes of Google, Facebook, or Uber. This also means that a lot of those investments fail. In fact, VC funds earn capital gains from only a small portion of their portfolio companies; typically, more that 75% of investments are written off (Hochberg et al 2007). Exits from only a very small percentage of top companies are what drive returns for the majority of the VC industry (CB Insights 2015).

The funds typically range from $50 million to $100 million for seed-stage focus and $100 million to $500 million for later-stage startup investment, with a 10-year lifespan. Partners who manage a VC fund make money from managing these funds in two ways. First, they typically charge 2% of a fund in management fees, so managing a $100 million fund garners them $2 million over 10 years, or about $200k a year. They also make 20% carry— 20% of the profit from the fund. This structure incentivizes them to raise and manage multiple funds simultaneously, and allows them to lock in high levels of personal income, even if they fail to return investment capital to the limited partners who invest in the fund (Mulcahy et al 2012). There is also pressure to deploy at least a third of a fund within the first 3 years of its 10-year cycle so that returns can be realized during the lifetime of the fund. This means there is a rush to make decisions on investing tens of millions of dollars. This rush to make decisions is not good for finding and fostering innovative startups, but notably, it’s also not good for the VC funds themselves..

Challenges for VC – The way VCs are structured, as described above, leads to several major challenges for the VCs themselves. First, there is the virtuous cycle. The fact that many VCs have emerged from successful startup founders themselves means that, yes they have a tolerance for risk and good business insights to a large degree. But that doesn’t provide them broad background and knowledge into domains and markets that they may want to invest in. Second, the structure of trying to get the “home run” investment that counterbalances the others that fail leads to a hunt for “unicorns,” or at least trends following in the footsteps of unicorns. While some VCs have overarching theses around how they aim to invest and there is some level of due diligence before selecting startups, there are no methods, broadly speaking, for doing things in a systematic way. And, importantly, the pressure to deploy exacerbates things. Most invested capital needs to be deployed in the first few years of a 10-year fund cycle, leading to rushed decision-making. This rush to find something links back to the issue of the virtuous cycle as well, as VC partners look to entrepreneurs they know to find something fast. This stifles funding for truly innovative emerging startups who may lack connections.

But beyond stifling innovation, and more to the bottom-line for investors, VCs in general just don’t perform well for the risk they entail. VCs have not out-performed the public market. In fact, a recent study showed that only 20 of 100 funds generated returns that performed 3% better than the public market; the average VC fund fails to make a return on investor capital (Mulcahy et al 2012). Between 2004 and 2014, venture capital as an asset class only slightly out-performed the S&P 500, with an average annual return of 8.1%, compared to 5.7% (Cambridge Associates, 2013). In general, this performance is mostly accounted for by the top performing firms like Andreesen Horowitz, Accel, Benchmark, Greylock, Kleiner Perkins Caufield & Byers, Sequoia, Union Square, and Y Combinator. The top approximately 20 VC firms, account for 95% of the return in the industry (Rachleff, 2014). The others struggle to have the same sorts of networks to source the best startups.

In short, returns are unsatisfactory industry-wide, and venture capital underperforms as an asset class (Kedrosky 2009).

To put things bluntly, as Josh Kopelman, a well-known venture capitalist did: “[F]or an industry that funds innovation – it really doesn’t have that much” (Griffith 2013). In part, this seems due to the lack of structure and of systematic processes within the VC community. While some of the top VC firms have specific theses about the way they will invest or the domains in which they are keen to follow trends, much of the decision making relies on thought leadership and connections. But beyond these informed opinions, there are no general systems or methods for exploring potential domains, finding startups and making decisions on funding them. It seems that it may well be of interest to utilize research more adeptly within these funds, and to have a more meaningful process. VC partners have the capital available to subsidize research, and the returns from having a more informed and thoughtful process would be worth the investment, especially. It seems that incorporating ethnographic thinking and research would be a step toward creating a more innovative VC model, which in turn could fund more innovative startups and hopefully foster more sustainable ones.

A More Meaningful, Ethnographic Approach

There are several overarching challenges for the VC model– in identifying areas of opportunity and trends, and in finding, funding, and fostering more innovative startups.

Identifying areas of opportunity is often based on hunches, rather than research. And in particular, there is a lack of knowledge about different markets globally and burgeoning areas that are spreading tech innovations into other domains and industries and markets. In finding new startups– what’s known as “deal flow”– one of the biggest challenges is that VC firms rely on their internal knowledge and connections rather than solid ideas, and decisions are usually reactionary and informed by a herd mentality. Making decisions around funding startups usually centers on hollow metrics. While there is a level of due diligence, the data sources are not great and lead to funding startups that may not truly be innovative. And finally, fostering startups tends to be an afterthought, particularly in looking at broader impact, values, and sustainability. VC investment controls what startups succeed in large part, so these issues are paramount to fostering innovation in the startup world.

Beyond a Realist Ontology

If we are indeed studying something that is “real” for the purposes of innovation, we are also intentionally trying to understand how we can change it and thus acknowledge the world as socially constructed; we are actively constructing reality. This calls into question the ‘realist’ ontology and epistemology so common in conversation around the predominant folk model of ethnography in innovation. The power of ethnographic methods and ethnographic thinking is not about finding new territory to colonize for startups and for VCs. It is in the rich, reflexive, deconstructionist, interpretivist perspective it provides us. In imbuing an ethnographic mindset into the main elements of the startup sphere, the goal is not to create disruption for the sake of a profitable “disruptive innovation” orientation. It is to make innovation more meaningful. To drive value. Ethnographic approaches enable this in several ways: analysis of complexity, anticipation, attunement, advocacy. The tools inherent in an ethnographic methodological approach allow for analysis of complexity in uncharted terrains. The focus on studying socio-cultural contexts and their dynamics enables would-be innovators to anticipate change in social meaning. An interpretivist mindset underslies the ability to become attuned to a domain, a market, and importantly, an individual human who may be a user. And, finally, being reflexive allows us to understand value and to advocate for those values that make for positive change, not just disruption.

Julia Katherine Haines conducts research at the nexus of technology, innovation, and human practices. She received her PhD in Information and Computer Sciences from University of California, Irvine. She has an MS in Human-Computer Interaction and an MA in Social Sciences. Julia is currently a User Experience Researcher at Google.