I found the most interesting part of the ensuing discussion to be the disagreement over whether this CEO’s act of digging in and coding the feature himself was a leadership success or failure. We didn’t do a formal survey, but the group appeared to be divided into two camps: one that felt he should have focused on solving the communication and process (and possibly staffing) issues that prevented his team from executing as he desired; and the other that saw value in the example he set by showing he was capable of and prepared to do what he asked of others.
Earlier this week I read Zach Bruhnke’s excellent post You’re not the CEO – you’re the Fucking Janitor, and it took my mind back to that discussion about what good leadership looks like in a startup. My answer: it depends. It seemed to me that the folks at the summit who felt this CEO failed by doing instead of managing were leaders of more mature companies, while the ones who admired his leadership by example tended to be running earlier stage startups. As someone running a company that had recently raised our Series A and was growing from a team of 5 in January to 14 today, I found myself agreeing with both sides of the debate.
For a boot-strapped or even seed-funded startup, I think Zach’s post is spot on. The “CEO” in Zach’s story is a total douche, and my business cards say “Co-founder” precisely because calling myself the Chief Executive over 4 of my friends made me think of Yertle the Turtle. My dad always told me “the fish stinks from the head”, which is just his graphic way of saying great leaders lead by example. In my relatively short leadership career thus far, I’ve taken this to heart and always jump at the opportunity to do things myself.
In addition to the mutual respect and motivation Zach mentions in his post, one of the greatest advantages I’ve found in this approach is the intimate understanding a leader attains of how things are done within their team. Across the many failures of leadership I’ve observed (I was at Yahoo! for 4 years
), there’s a recurring theme of the leader being too removed from the actual doing. Especially in the technology world, the means of production can be just as important as the output. I can’t tell you the number of product and business leaders I’ve dealt with who treat engineering like a commodity instead of a potential competitive advantage. You only need to look to the world’s most valuable company to see what great supply chain management (i.e. caring how the sausage gets made) can do for your business. And when you’re a software company, every architectural decision your team makes has a bearing on essential business considerations like performance, reliability, time-to-market, and agility in responding to new threats and opportunities. That’s why awe.sm is, above all else, an engineering-driven organization (and looking for even more great engineers
).
Never trust a "startup" CEO who doesn't know how their product is built.
— J-Strizzle (@jstrauss) March 10, 2012
But in a later stage company, the leadership challenge is greater because you need to figure out more scalable ways of achieving these same goals. There was one particular line of Zach’s post that stuck out for me in this regard:
If you want to be a CEO in the sense that you dream of then you should remember to be the Fucking Janitor too.
A couple months after raising our Series A, I was washing dishes in the office and caught myself feeling self-satisfied because here I was, CEO of a company that had just raised millions of dollars, doing the dishes. I thought about my dad’s smelly fish saying and how he’d be proud of me. Then I thought about our investors and what they’d think of this…and it struck me they’d be pissed. Here I was, CEO of a company in which they’d just invested millions of dollars, doing the dishes instead of the dozens of other things only I could be doing to make their investment successful.
In the few months since then, my leadership focus has shifted. I still do the dishes when it’s my turn; when AWS shits the bed at some inhuman hour, I’m in our IRC room doing what little I can to help; and I always want to understand the gory details about why we made one architectural decision over another even if I wouldn’t know how to implement either of them myself. I am proud to continue to be a colleague to my team above all else. But leadership in a larger organization requires more than that. Our goal is to achieve on a scale bigger than what one person can achieve alone, and that means the leader needs to lead not just do. Doing is good, but when it turns you into a micro-manager or takes you away from leading, it can be counter-productive.
Delegation is hard. I’m finding delegating well to be much more challenging than doing things myself. Leading purely by example just requires effort and a willingness to do things that aren’t fun or glamorous, and as the leader you’re usually the most incentivized to get those things done. But effective delegation requires much more than mere will, it is a skill set developed with patience and learning and painful trial and error. It requires finding great people, training them in the skills you need them to have, motivating them to share your goals, empowering them with the resources and information to be successful, trusting them to do their jobs, and then giving them feedback on how to improve. I have come to believe my primary job as a leader is to enable the members of our team to deliver what the company needs from them, and that’s a lot harder and even less glamorous than being the Fucking Janitor.
]]>For me, the most interesting ideas were on 'constructing value' for a previously unknown product. First, you need to have a compelling free offering so there is minimal risk to customers trying your new product. Secondly, you should take the lead in setting the perceived value of your new product by using an 'anchor' of premium pricing even if you end up selling at a discount to it.
"We really enjoy the work that we’re doing and we don’t want to cheapen it. Consciously or unconsciously — probably both — we’re trying to create a manageable way to earn a living and still maintain our sanity. We value time as much, if not more so, than money."
"The basic pattern of openness is that better access to information and better systems lead to better decisions and better living. This general principal is broadly accepted, but we're just now discovering that it also applies to the minutiae of our lives."


2009 is not 1949. Modern consumer electronics and the Internet have ushered in a new economy of plenty. Production costs have decreased at an astounding rate. Distribution has become ubiquitous. It is no longer necessary, for all but the most ambitious projects, to seek out a studio to bankroll your pilot and a network executive to green light you for a time slot. Your talent and work ethic are now greater predictors of your success than who you know."





While theoretically possible, I believe modeling the impact of these two factors is a practical impossibility because they are so intertwined — de-leveraging sparked the recession and the recession is driving further de-leveraging. You could also do a technical analysis where you try to match current market behavior to past patterns and extrapolate what happens next based on what happened before. But that method requires making a bet on which past patterns to match against, i.e. is our current situation more similar to the Great Depression or all the recessions since. And that’s a big bet.
So, I propose a different (and much simpler) approach: assume a realistically sustainable growth rate over a long enough period and figure out where we would be if the market had grown at that pace. I picked 20 years as the period and charted the monthly percent change of the Dow Jones Industrial Average (DJIA) from 2,342.32 at the end of January 1989 to 8,131.33 at Friday’s close. The actual percent change is the blue line, and I plotted 3 other lines against it: 10% annualized growth in green; 6% annualized growth in orange; and 2% annualized growth in red.
Here’s what that period looks like in annual percent change for the DJIA and S&P 500:
And here’s the annualized rate of return for both the S&P 500 and the DJIA since 1989:
Over the course of this 20 year period, there were only 6 years in which the market declined *at all* (one of which, 2005, was basically break-even) and there were 10 years in which the market gained *more than 10%* and in 7 of those it gained *more than 20%*. Even after the tech bubble “burst” taking the market from 11,497.12 at the end of 1999 to 8,341.63 at the end of 2002 (a 27.4% decline in 3 years), the market would still have delivered a *10.1%* annualized rate of return over the prior 12 years. From where we sit today, it’s no wonder the DJIA declined 33.85% in 2008 (taking us to a still very respectable annualized rate of return since 1989 of 7.24%). But that doesn’t answer the question of how much further down it needs to go before we can consider the value of the equity markets stable.
That’s where the annual growth rate analysis comes in. It is still highly subjective — depending on what one believes to be a representative sample period and sustainable annualized growth over that period. But I like it because it helps me think about the broader market in terms I feel more comfortable making assumptions about, like what do I think is a reasonable rate of value creation for the economy as a whole over a given period. In the case of the 20 years since 1989, do I believe there’s a reason that the equity markets should have averaged ~10% annual growth while our Real GNP achieved only 2.76% annual growth over the same period? No, and obviously neither does the market at this point.
So, what is a reasonable expectation for a bottom? Your guess on the underlying assumptions is as good as mine. But if one believes technical advancements over the last 20 year period enabled us to double efficiency (i.e. extract twice as much profit from the same revenues), then the markets *should* have grown at around twice the rate of Real GNP. In that case, we would be expecting 5.51% annualized growth in the markets since 1989 as of the end of 2008. Starting from 1989 closing prices of 2,753 on the DJIA and 353.4 on the S&P 500, 19 years of organic equity growth pegged at 2x GNP growth should have closed 2008 at 7,634.38 and 979.95, respectively (actuals were 8,766.39 and 903.25).
Update: Dad accurately points out that GNP is a trailing indicator and equity prices are leading indicators. So, this analysis shouldn’t be considered anything other than directional. I find it helpful as one factor in my overall assessment of the current situation, but as Howard reminds us no one really has all the answers.
This final chart shows the actual level of the DJIA (blue) compared to what it would be if it was pegged at 1x GNP Growth (red), 2x GNP Growth (orange), and 3x GNP Growth (green). It is essentially the same as the chart at the top but now instead of arbitrarily picking annual growth rates, we have pegged them as multiples of Real GNP (i.e. ratios of business efficiency). As you can see, for most of the last 20 years the markets were assuming >300% improvements in business efficiency.
All the above charts and underlying analysis can be found in this spreadsheet.
]]>