‘The fox knows many things, but the hedgehog knows one big thing’ – Greek poet Archilochus
This ancient distinction used by Issiah Berlin in his essay The Hedgehog and the Fox: An Essay on Tolstoy’s View of History was extended to two cognitive styles by Philip Tetlock in his book Expert Political Judgment: How Good Is It? How Can We Know?.
The gist of the argument is that Foxes have an eclectic toolbox, one they apply without any concern for pet theories/concepts (they don’t have any) and as such can better adapt themselves to changing circumstances. Hedgehogs on the other hand, have a single big idea/theory/concept that they try to fit the empirical world into. They have a cognitive need for closure that is satisfied by bludgeoning reality into their favorite model/worldview.
Tetlock, in his book, goes on to empirically demonstrate how foxes have been better forecasters than hedgehogs in the arena of political science. Not surprisingly this distinction between foxes and hedgehogs makes intuitive sense and similar forecasting results (i.e., foxes perform better) are likely in fields which are high in complexity and uncertainty. The dynamic nature of such fields (including business, economics, technology etc.) makes “experts” with one big idea vulnerable to getting blindsided by unexpected big events, i.e. The Black Swan.
A recent post by Nassim Taleb highlighted the way people’s identities are wrapped up in their work and accomplishments. Nassim argues that this is not only different from the past (ancients) but also varies across the economic spectrum with only the minimum wage earners being able to separate their identities from the mode of earning a living. This undoubtedly ties into his earlier point about never taking advice from people wearing a tie. Those whose identities are close to their (knowledge) work tend to overlook the fallibility in their fields. A mind focused on falsification and professional pride are rarely seen in the same person. So how can one (especially if you are a ‘knowledge worker’) avoid this problem? I think Paul Graham had the best advice (closely tied to but broader than Nassim’s comment) : “The more labels you have for yourself, the dumber they m
ake you." Keep your identity small.
“Men may not take their god too seriously; but how they do bow to earth before their shrine of three meals a day, their adulterated white flour, their vast groaning tables of unnecessary food!”
I conclude, therefore that, fortune being changeful and mankind steadfast in their ways, so long as the two are in agreement men are successful, but unsuccessful when they fall out. For my part I consider that it is better to be adventurous than cautious, because fortune is a woman, and if you wish to keep her under it is necessary to beat and ill-use her; and it is seen that she allows herself to be mastered by the adventurous rather than by those who go to work more coldly.
Bias checklist: this post is probably riddled with recall bias, but all references were checked and re-read. A key assumption I am making is that the entrepreneurs reading this post are focused on building businesses that make operating profits (and aren’t just targeting quick exits).
Feel free to point out biases/errors you see in Comments section, i’ll come back to this topic in a future post. As always caveat lector.
A recent blog post by Paul Graham got me thinking about how growth in recent years has being positioned as THE goal of all early stage firms (i.e., startups). All good things flow from the achievement of this ephemeral target, in fact those who can help early stage firms get this growth have been labeled growth hackers. One of the fascinating things about this focus on growth above all else is that it was not always so. The purpose of this post is to (very briefly) give an overview of two distinct priorities in the pursuit of success (growth vs. profit) that have been common not just in the tech sector but also in more traditional industries (public company valuations have for a long time been split into these two categories). Ostensibly in the technology sector growth either accompanies or is shortly followed by profitability.
In order to build on these two distinctions and link them to this “growth fallacy” i’ll divide the business landscape into the two categories conceived by Nassim Taleb, namely mediocristan and extremistan , he developed this distinction in his book The Black Swan: Second Edition: The Impact of the Highly Improbable: With a new section: “On Robustness and Fragility”.
In order to show how entrepreneurs can better analyze these landscapes i’ll use some simple concepts from strategic management (value creation vs. capture and disruptive innovation) to explain why this pursuit of growth is great from the point of view of angel investors and venture capitalists but not necessarily for entrepreneurs (critical distinction that is glossed over in most posts on startup growth).
Growth Hacking: Cui bono?
The gist of Paul Graham’s growth post is that startups are citizens of Extremistan while most conventional businesses are in Mediocristan, i.e. startups are exclusively high growth ventures that are not constrained by distribution and other barriers that limit the growth of most conventional businesses. However, the lack of these barriers also means that you have global competition and a crowded landscape which makes the entrepreneurial idiosyncrasies more valuable.
These idiosyncrasies are valuable because they not only allow entrepreneurs to spot new openings in the landscape but also utilize new technologies to target these opportunities. High growth (either revenues or users in case of Y Combinator) startups will presumably after n iterations hit upon some form of success, Paul emphasizes this frequent iteration and targeting of growth in his post:
“Nine times out of ten, sitting around strategizing is just a form of procrastination. Whereas founders’ intuitions about which hill to climb are usually better than they realize. Plus the maxima in the space of startup ideas are not spiky and isolated. Most fairly good ideas are adjacent to even better ones.”
Obviously, if the primary objective of a startup is growth then targeting Extremistan landscapes makes sense. These tend to be highly scalable businesses that are (frequently) populated by competitors since barriers to entry are lower (e.g. mobile applications). However, barriers to success (= growth?) are higher simply because these fields tend to be cluttered and its harder to genuinely differentiate oneself and create customer value. Its even harder to actually capture that value because close substitutes are usually available at low cost. From the perspective of the entrepreneur, the base rate of success is pathetic.
Which raises the question: who benefits from this unbridled pursuit of growth? Here again, Nassim Taleb’s insights (more used than acknowledged) are valuable. One of the key features of Extremistan is that it is the domain of Black Swans (unexpected events with high impact that give illusion of predictability after the fact). These Black Swans can have negative impact (financial crisis of 2008 for those who didn’t see it coming) as well as positive (investing in the next big thing). Taleb’s consistent advice (in Fooled by Randomness, The Black Swan and his forthcoming book on Antifragility) has been to position oneself on the winning side of Black Swans, both by minimizing exposure to risky non-linearities and increasing exposure to their upside. The latter include venture capital/ angel investing opportunities that (for limited $ investment) have a call option on a potential bigger upside. And therein lies the rub, the answer to the question I asked earlier is that blind pursuit of growth benefits the buyers (VC or Angel investors) of these call options and not the individual options (the entrepreneur) themselves, because for the entrepreneur the average outcome of his angel investor’s portfolio is meaningless (you can drown in a river that is 4 ft. deep on average) but for the angel, that is a key metric. Thus, what works for early investors may not be optimal for entrepreneurs wanting to create real businesses.
Now I turn my attention to the basic goal of all business strategies.
If you are an entrepreneur, your goal ought to be profitable growth that accrues from competitive advantage, some have already pointed out the complex issues surrounding pursuit of growth in tech investing. But calls for pursuing revenue and/or user growth with no talk of profits merely risks repeating the errors of the last tech bubble. With few outliers (who get media coverage without any mention of base rates of success) getting all the attention, its easy to forget how much the system goes against common sense approaches to building real businesses.
The simplest and most relevant framework in this context is Clayton Christensen’s disruptive innovation. An additional benefit of using this lens to look at the growth question is that disruption is commonly invoked (often un-justified) in the internet and tech investing sector. Since his theories are so well known, I won’t rehash them here. For entrepreneurs, a good starting point is his book Seeing What’s Next: Using Theories of Innovation to Predict Industry Change
One overlooked element of the disruption framework of Clayton Christensen is his emphasis on profits (not revenue or user growth). There are a couple of reasons why this distinction is so critical for entrepreneurs:
- Pursuit of profits forces entrepreneurs to vet ideas on a critical dimension: are customers actually willing to pay for it? Early pursuit of profitability thus becomes important as a validation exercise.
- Profits imply that the cost structure is low enough for the startup to actually manage to capture value in addition to merely creating it (which metrics like user or revenue growth indicate). Value capture cannot happen unless genuine moats are present around the business. Hence, high traction in users or revenues may imply that you have reached a local maxima that is valuable but indefensible. Great for exits that may not be such good ideas for the acquirer. This factor is another reason why raising lots of venture capital in early rounds can distort the value of the opportunity since external capital can makeup for weaknesses in profitability. Groupon is a classic example of why growth (revenue or user) at the expense of profitability and competitive advantage may not be sustainable.
Which brings us back to the key difference between being a buyer/holder of multiple call options (VC/ Angel) vs. being an actual entrepreneur. If you are in the former category, pursuing multiple opportunities in Extremistan and pushing them for high growth makes sense since you only need one or two outliers to make your performance exceptional. However, if you are the entrepreneur, it is highly risky to pursue un-profitable growth (every time someone gives you the example of unprofitable startups that made 8 or 9 figure exits like Instagram, take a few deep breaths and silently repeat..anecdote…anecdote…anecdote). Ask yourself cui bono?
Bias list: In keeping with my caveat lector motto, the following may have these biases inherent in it: selection bias (small sample of scholars, not randomly selected), recall bias (I may have missed noting some critical point during the presentations).
Multiple experienced scholars shared their experiences with the dissertation process, all the great advice is to their credit, all errors are mine. This is not a complete record of the session, just my notes. Some of the scholars who presented (in order of talks given) were:
I. General observations:
1) Keep big picture in mind (Tom)
2) On oft repeated quote that “best dissertation is a done dissertation”:
a) Tom: not strictly true, it is an “incomplete idea” at best because things get more challenging for each generation of scholars in terms of expectations from a “good dissertation”. Your main goal should be to have a dissertation that leads to A-level journal publication(s) in your field.
b) Lisa: “doesn’t sit totally right with me…should be more than that.”
3) Tom had a bunch of useful advice specific to why academia is so attractive: you get to do interesting work with interesting people while having enormous personal control over your time.
II. On developing a Research Stream:
1) Tom: decide if you are a focused researcher or a sequential researcher (doing a new topic every 5 years or so). Narrated anecdote by two Organization Behavior researchers that you get 1-2 years to become THE EXPERT on a topic of your choice (“the biggest opportunity”). Don’t get sidetracked by uninteresting or petty stuff.
2) Brent suggested integration (intersections of research sets) between your research interests that are broadly allied.
III. Mechanics of the Dissertation Process: this topic was mainly presented by Lisa, notes from her presentation
1) Many things about the dissertation process are new.
2) Your dissertation “defines you on the job market”.
3) Lisa used the metaphor of “the maiden voyage” for the dissertation to reflect that this is (probably) the first full-scale research project that you lead. This maiden voyage has the following components:
a) Destination & route (i.e. topic & method):
i. In terms of topic selection key advice was to identify what you like to read about? What puzzles you? Critically, in areas that interest you, where is the conversation headed?
ii. Writing a one page description of the ideas that interest you allows you to move to the second step, namely idea vetting. Test marketing your ideas important (initially with advisor, later on with broader committee members). A doable dissertation needs to be specific, Lisa’s donut theory of dissertation, you may collect data for the broader donut but the central hole (research question) needs to be specific.
b) Your crew (i.e. committee):
i. Advisor: most important criteria is availability. Other considerations are knowledge of literature, methodological savvy, good working relationship (i.e. challenges your ideas constructively).
ii. Committee: should have members who fulfill role(s) of a research design expert, data analytic expert, literature expert and great sense of the field.
c) Your ETA (estimated time of arrival):
i. Key junctions in your journey are: first, after the topic is established – should have round 2 of your idea vetting (potential contribution & challenges?). Second, after core elements determined (before proposal defense) – determine what must go well (i.e. necessary conditions for success)?
ii. Generate momentum in dissertation work.
iii. Map out big chunks of your time for each step (BE REALISTIC).
This dissertation is a “great process model that can be modified for future voyages” (i.e., other projects in the future).
I have been a heavy consumer of blogs (mainly business and technology) for over a decade and recently began to think that I should turn my attention to producing. When I decided to write this blog, a few things about blogging (medium of expression both liberates and lassos you) stood out:
- Blogs are inherently conversational and express “the unedited voice of a single person”.
- However, this conversational and topical nature (the very things that attract an audience) of blogs (true to a lesser extent of other mediums like talk radio, TV etc.) make them highly susceptible to heuristic biases and logical fallacies.
- Biases and ease of publication (no editorial or peer review required) on part of the author combined with tendency towards confirmation biases on part of the reader means that many blogs spiral downwards into echo chambers.
As a general warning/ checklist for both my writing and readers, I have a list of (some) common biases below that have the potential to affect my writing. Since this list is based on the most common biases (full list of biases) I have noticed as a reader, they apply equally to any other blog posts you may be reading.
The Bias Checklist:
- Availability bias: refers to what comes first to mind or is easily recalled. In making this list, I am guilty of the availability bias.
- Confirmation bias: tendency to look for evidence that confirms existing beliefs/ theories. Alternative is actively seeking falsification. I have yet to come across a blog that actively seeks falsification, reader comments can provide counter-examples but often degenerate into trolling.
- Base rate fallacy: far more common than realized, base rate is the frequency of a generic type of event. This is significantly driven by the fact that blogs are topical and hence writers are exposed to specific rather than representative samples from any given population. The best way to avoid this fallacy is to always keep in mind (or look for) the base rate (of event under consideration) for the whole population.
- Self-serving bias: refers to tendency of individuals to attribute their success to themselves and failure to external factors. This shows up heavily in business and technology contexts as halo and network effects are systematically underestimated as drivers of success. For example, is a company successful because of the attributes (schooling etc.) of a CEO/ company (behind headlines such as: “How to Innovate like…” or “10 Takeaways from ….”) or because the starting conditions led to dominant positions?
- Survivorship bias: this is perhaps the most misleading bias. It refers to tendency to focus on companies (or individuals) that survived a process rather than the starting population. I claim this to be the most misleading because it give a false sense of understanding and causality. Just because some attribute (innovation, html5 ninjas, hoodies, short meetings) is present in successful companies doesn’t imply that it is causally linked to their success. An easy way to spot this one is to see the sample being used, did the writer/ blogger/ researcher talk to or survey the starting population or just those that succeeded? This bias is rampant in studies of entrepreneurship, technology and innovation. Discount the claims of any blog post/ research that is based ONLY on successful companies or entrepreneurs. The danger of such research has been highlighted by writers like Nassim N.Taleb and the medical researcher John P.A. Ioannidis.
I tend to keep this checklist in mind when reading and (hopefully) will be able to avoid falling into these traps while writing. Whenever it is not possible for me to avoid a bias (for example, when writing about social networks I will likely be indulging in survivorship bias by focusing on the top few companies in the industry) I will point out that shortcoming as a limitation of any inferences I make. As always, caveat lector – reader beware!
This blog is motivated by the following excerpt from Shakespeare:
There is a tide in the affairs of men.
Which, taken at the flood, leads on to fortune;
Omitted, all the voyage of their life
Is bound in shallows and in miseries.
On such a full sea are we now afloat,
And we must take the current when it serves,
Or lose our ventures.
Brutus in Julius Caesar Act 4, scene 3,
The common thread running through all posts will be the importance of time, randomness, path dependence and strategy in outcomes in business, technology and life.