Splitting the Pie: Founding Team Equity Splits

Let's kick off 2006 with an issue about which I get a lot of questions (and which StatCounter reports is one of the highest-ranking Google search terms that lead people to this blog). That issue is equity splits within the founding team.

In our first-year MBA Entrepreneurship course, the Zipcar case represents the typical approach to splitting the equity. The two founders, who barely knew each other before founding the company, went the "easy handshake" route and split the equity 50/50. (Commonly known as the "Rule of N" approach; with N founders, each founder gets 1/N of the equity.) One founder ended up joining Zipcar full-time as its CEO and went on to become the driving force (sorry for the pun :->) behind the company. The other founder kept her other job and contributed far less to the company, but still owned the same amount of equity in the company. When she later visited our MBA classroom to watch us teach the case, the founder-CEO said, “That was a really stupid handshake because who knows what skill sets and what milestones and what achievements are going to be valuable as you move ahead. That first handshake caused a huge amount of angst over the next year and a half.”

Question #1: How do your experiences with equal equity splits compare to this?

We use my “Ockham” case as a counter-case to Zipcar. First, the three founders (who had long been co-workers) split the initial equity unequally (50/30/20, based on the different amounts of capital that the 3 contributed; the founder who became CEO got the 50%). Second, they crafted a "dynamic" founder agreement where any founders who were not participating in the company full-time a year after founding would surrender their equity (in effect, imposing conditional vesting on themselves).

In my recent keynote at the Emerging Ventures conference, I also drew upon some ancient Microsoft history, wherein Microsoft's founders seem to have used parts of each approach. Their initial equity split was unequal (after a slight early adjustment, 64% for Bill Gates, 36% for Paul Allen), due to differences in what each founder was giving up to join Microsoft, in their anticipated roles (their titles would be "President" and "Vice President," respectively), and other factors.

However, as in the Zipcar model, the split seems to have been static. At the time of Microsoft's IPO in 1986, Gates owned 45% and Allen owned 25%, according to the company's filings at the time. Even though Allen had left Microsoft in late 1982 for health reasons, this was still the same ratio of ownership (45%/25% = 1.8 in 1986) as when they split the equity in 1977 (64%/36% = 1.8), suggesting that except for dilution from selling additional shares after the initial split, there was no equity adjustment between the two founders.

Question #2: Have you seen other constructive or creative solutions to solving the problems with the Zipcar approach?

The contrast between the Zipcar and Ockham cases also sparked my interest in how the past relationships of the founders (e.g., prior friends or co-workers, vs. strangers) might affect their approach to splitting the equity (both the equality of the split and the timing of the split) and how the split might affect both the founding-team’s stability and the company’s growth. Therefore, in my IT start-up survey, I have started collecting data on founder equity splits, and will be doing so again in this year's survey. As shown in the pie chart below, last year's survey showed that far more start-ups follow the Zipcar model than the Ockham model, despite the former model's apparent flaws. (To use the Microsoft example, the Gates-Allen gap would be 64%-36%=28%, which would put Microsoft's founders into the "11%-30% Gap" slice, along with 19% of the start-ups in my sample.)

From these cases and other observations, there would seem to be at least 4 core factors on which teams could (should?) base their equity splits:

1. Past contributions: Who came up with the Big Idea? Helped refine the idea? Put money into the company to help get it started? Helped find another co-founder or seed investor?
2. Future contributions: What role will each person play in the early months? Will that person still be playing a key role in a year or two (or more)? Still be working for the company at all?
3. Opportunity cost: Is one founder giving up a cushy job at a top company, while the other is not currently employed? Is one dropping out of a good school, the other otherwise unemployed?
4. Your relationship: Do you trust your co-founder to surrender equity to you later if you end up feeling like you're contributing more than he is? Are you willing to fight over the equity (e.g., sacrifice some of the relationship with your co-founders in order to get another 5%)?

Question #3: What other factors should founders consider in splitting the equity?

By far, the toughest to figure out -- and often most crucial -- is usually #2, "Future contribution." If we take it as a given that the venture will evolve in ways that can't be anticipated fully at the beginning, or that a founder might not be able to scale with the venture (or might change his mind about his involvement), then #2 would seem to suggest that a dynamic element, akin to Ockham's founder agreement, is important.

In the coming months, Matt Marx (a doctoral student here at HBS) and I will be using the data from my surveys to analyze several research questions about splitting the pie, and plan to write at least one paper about it. Our initial questions include:

• Does the prior relationship among the founders (prior friends, family, or co-workers, versus strangers) affect how (or when) they split the equity?
• If the founding team splits the pie equally, should the team be more stable or less stable than teams that split it unequally?
• Does the equality or timing of the split affect how long each founder keeps working for the company? Affect the company's growth or its eventual valuations?

Question #4: What impacts do equity splits have on individual founders, the founding team, and the company overall?

Finally, as I posted after my Emerging Ventures keynote, a GP at the conference argued that even if the founding team does a bad job of splitting the equity, it is actually very easy to adjust equity within the team, using additional equity grants. In response, an entrepreneur posted that no, it's not so easy in practice. (Zipcar's founder didn't find any easy solutions, even after Zipcar's A round.) Let's revisit this.

Question #5: How easy is it to adjust "bad" equity splits later on (e.g., when the company is raising its first VC round)?

Quick Follow-up (1/16/06): I just finished watching “Startup.com,” a documentary about the founders of govWorks.com. Regarding Question #5 above, the issue of co-founder hold-up (because of a bad initial equity split with no provision for later adjusting the split) dominates the “Pressure” segment of the film.

Founders Kaleil and Tom had a 3rd co-founder, Chieh, who put up $19,000, worked “after-hours” for 5 months (he had kept his day job and refused to join govWorks full-time), and then dropped out. When they’re about to close their big first round, Mayfield was not willing to close until the other 2 founders bought Chieh out. Correct me if I'm wrong, but from what I can tell from the film, the VCs were willing to do a $410,000 “sweetheart deal” to buy Chieh out. However, Chieh wanted $800,000. Amid the pressure to close the round and Chieh’s alleged threats about “all kinds of nasty things” (or at least the potential for him to sabotage the venture), Kaleil and Tom ended up caving in and settling with him for $700,000. (Kaleil: “I’m being extorted.”) They had to make up the difference (between the $410,000 and $700,000) out of their own pockets.

Updates (11/13/06 and 12/18/06): I just posted initial results from my quantitative analyses of the factors that affect whether teams split equity equally and whether equity splits affect founding-team stability, based on data collected earlier in Spring 2006 on more than 300 founding teams. Love to get your thoughts on the initial results there!