Monday, December 11, 2017

Bitcoin Network Effects

One force people underestimate are network effects. Network effects allowed Twitter survive its fail-whale period when the site was constantly down. Facebook destroyed international competition, and Airbnb and Microsoft all succeeded due to network effects. Open source examples like Ethernet survive to this day due to network effects. 

In the crypto world, Bitcoin is perceived as slow to change, clunky technologically, and as having bad governance. While all these things may be true, Bitcoin has strong network effects that will maintain its status as the primary value store in the short to medium term. It is always possible Ethereum or another newer protocol will take over the value store use case in 5 to 10 years, but network effects decrease its likelihood in the short term. 

Bitcoin's network effects consist of: 

1. User, fund and investor adoption. Bitcoin has the largest user base adoption, as reflected in its market cap. With millions of people now holding bitcoin, there is incentive for a large sub population to spread adoption. Additionally, large institutional holders have started to buy bitcoin and stand alone funds like GBTC have reached $1 billion in market cap holding bitcoin alone.

2. Exchange adoption & network liquidity: Bitcoin as the "reserve currency". Bitcoin is listed by all the major exchanges, which in concert with its large userbase creates large scale market liquidity. This liquidity has caused additional large institutional investor to dip their toes into the crypto waters buy buying in and accelerating further liquidity.

Indeed, bitcoin along with fiat is the reserve currency of the crypto world. Prices on exchanges are often denominated in bitcoin. Bitcoin trades into other coin volume is the second highest after fiat. 

3. Derivatives and other financial infrastructure. Once liquidity starts to be achieved in an asset class entrepreneurs start to build infrastructure to support it. Bitcoin is the first cryptocurrency to have derivativescustody solutions, and vaults. These incremental solutions like derivatives will accelerate adoption by additional large scale financial buyers, pushing the network effect forward.

Obviously, other cryptocurrencies will (or already) have financial infrastructure extended from bitcoin to them. However, the first mover effect by bitcoin will increase its market cap and liquidity further.

4. Mainstream press and attention. Bitcoins large userbase, fund investment, ongoing market cap increase, and financial infrastructure build out all contribute to ongoing press. Press coverage in turn feeds back to increase the user base and fund participation of bitcoin. In popular culture, bitcoin is now more or less synonymous with cryptocurrencies.

Reflexive Assets and Network Effects
Fundamentally, value stores like bitcoin and gold are reflexive assets - assets whose value grows because others value them. Store of Value is the ultimate coordination game and for reflexive assets, price itself is a network effect. As the virtuous cycle outlined above kicks in, the reflexivity of bitcoin accelerates its value. See Soros on reflexivity here.

Forks as Dividends
Because of bitcoin's dominant position, over ten forks are planned next year for bitcoin. Barring a civil war in the bitcoin community, these are perceived by the investor community effectively as dividends. Bitcoin's market dominance makes it a juicy target for airdrops, reenforcing its dominance with investor.

Is The Slow Change Of Bitcoin A Feature or a Bug?
One knock on bitcoin is the slow pace of iteration and technology advancement in its community. The argument made is that new forms of innovation will cause migration off of bitcoin to another blockchain for the value store use case. For example, privacy features provided by zk-SNARKs or STARKs, increased network scalability, or lower transaction fees should all drive people off of bitcoin to ethereum or some still to be defined protocol. 

The central value store case (buy and hold) is a simple one. Privacy features are valuable for specific applications, and low fees make a big difference for specific use cases. For example, a payment network would need to be high scale and low cost. However, for the value store use case one can argue little additional innovation is needed for bitcoin outside of "don't fall over on performance and don't have a civil war". Stability is a good thing when a lot of your money is tied up in something. Bitcoin perhaps should be more like C++ avionics software than a javascript social media app. "Measure twice cut once" vs "move fast and break things"[2].

Bitcoin Failure Modes
Given the network effects above, bitcoin is likely to fail for one or more of the following reasons:
(i) Network degradation due to bad software performance or fees. Given the slow pace of innovation in bitcoin its performance may suffer as it scales and eventually become so bad another coin takes over. Similarly fees may grow too large.

(ii) Attacks via internal civil war or external government action. The much anticipated Segwit2X fork failed to materialize in November. However, the doomsday scenario in bitcoin is a war between developers and miners with coordinated attacks on bitcoin infrastructure on one side of a fork or another. Similarly, a government actor could try to attack bitcoin nodes rendering the network ineffective.

Government attacks could come as regulatory action, cyberattack, or adding sufficient hash power to take over the protocol (for example if China nationalized its bitcoin miners).

(iii) Dilution of what "Bitcoin" means. With each subsequent fork the branding of bitcoin becomes more confusing. In a world of Bitcoin Cash, Bitcoin Gold, Bitcoin Diamond, Bitcoin Zero which is the real bitcoin? What if hash power gets diluted down with each fork as a subset switches over to mine the new coin?

(iv) Mining power centralization. Mining power continues to be concentrated in the hands of a small number of actors. If mining blocks surpass 50% of the bitcoin network a miner can abuse its position in the network. The extreme example would be a state actor taking control of the bitcoin network and seizing or freezing assets. A less extreme example would be a large mining company abusing its position to censor or modify the blockchain for its own gain. This may actually be the biggest threat to bitcoin in the medium term.

(v) Unkowns. Given how early we are in the evolution of this technology and market, there is still a lot of that could go wrong that we have not anticipated.

Unless one of the above happen, bitcoin should continue to thrive in the short term as the primary store of value in the crypto world. Twitter survived literal years of fail whale, and bitcoin can survive some rough patches of its own. In the long run, it is possible technological innovation by a new chain may allow bitcoin to be surpassed, but networks effects are tough to overcome.

What Would A Post-Bitcoin World Look Like?
While I do not believe bitcoin will be replaced for value store in short run, some possible candidates to displace bitcoin would be:

(i) A derivative of bitcoin. The most likely replacement for bitcoin as value store would be a fork of bitcoin. By airdropping the new chain to all bitcoin holders, a bitcoin fork may fix performance and governance in bitcoin while keeping its network effects. The best way to participate in a future fork of bitcoin is to hold bitcoin today.

(ii) Ethereum. Ethereum has enormous developer interest and the second largest market cap. It also has characteristics that could make it a good payment network. At a minimum, ethereum is a likely "silver" to bitcoin's "gold" status. There are some scenarios where bitcoin collapses and ethereum takes over as primary value store. 

(iii) Privacy tokens. Zcash and Monero have additional features for privacy and anonymity that bitcoin lacks. For some use cases this is an attractive feature. Some approaches like Bulletproofs may allow for anonymous transactions that are still auditable for tax and other business purposes-which are likely needed for mainstream adoption. 

(iv) Stable coin. For payments you want stability in your currency. For value store, you want appreciation but may be happy with inflation-adjusted stability. Stable coin projects like Basecoin are working to create tokens whose price has minimized volatility.

(v) Parachain. Parachains such as Cosmos or Polkadot allow for cross chain transactions. If it is easy to move between bitcoin, ethereum, and other blockchains, each individual blockchain may cede value to the parachain.

(vi) Something that does not exist yet. The world changes fast.

Just as in the world of atoms there is gold, silver, and platinum - in the digital world there is likely to be more then one value store (although one will be dominant). As such, I am bullish on more then one token aggregating ongoing value in the coming years with bitcoin being the primary.

Thanks to Lucas RyanMatt HuangFred EhrsamAvichal Garg for comments on this post.

[1] Infrastructure extends beyond financial derivatives and other products I mentioned. Bitcoin miners have custom ASICs to mine bitcoin that will not be as performant on other hash functions. This creates some stickiness to bitcoin mining, although given hardware turnover cycles it is not a long term effect of importance.

[2] The last two lines were taken verbatim from Matt Huang. I did not think I could state it better.


Wednesday, October 11, 2017

Cryptocurrency Incentives and Corporate Structures

The cryptocurrency world faces a tragedy of the commons and free rider issue: there is little ongoing economic incentive to contribute to an existing, major crypto project versus to launch your own token [1]. Some projects are helping to rectify this by thinking about developer bounties for work done, a venture fund, and other incentive mechanisms. However, the corporate structures and token distributions used by many other crypto projects today tend to exacerbate a lack of ongoing incentives. In this post I review common crypto corporate structures and brainstorm an alternative approach to try.

Existing Project Structures (AKA Where Do All The Tokens Go?)
Crypto projects tend to have the following structures[2]:
  • Individual founders or early scientists or developers. They tend to get ~10% of the initial token on a vesting schedule. (In some cases this can be as high as 20%+)
  • A company, typically an LLC or C-corporation. The company raises money, hires the initial developer staff etc. The company often ends up with anywhere from 1-2% to 15% of the tokens post-ICO.   
  • A foundation, often set up in Zug Switzerland for tax reasons. The foundation mission is typically to watch out for the long term development of the protocol. In some cases, how it is to go about this mission or how it will divvy work with the corporation is uncertain. The foundation often receives 1-10% of the tokens and may use tokens to e.g. pay developers for bounties on projects. 
  • A venture fund. Some projects have raised enough money via ICO or pre-sale that they can set up a fiat currency fund to invest in companies that use their protocol. 
The primary challenges and nuances with the model above includes:
(1) People get rich fast and may drop out. The early distribution of a large amount of highly liquid tokens to early team members can cause people to drop out of the project once they are rich enough.
(2) Most of the value escapes from the company or foundation. By distributing tokens to early team members and founders, value escapes from the company/foundation and therefore the future of the protocol. By splitting tokens between the foundation and the company, there is further dilution of effort and focus. What should the company versus foundation do? And why create multiple separate pools of tokens when you need to have a concentrated pool to incentivize ongoing contributions to the protocol? (Besides the tax efficiency of a Swiss foundation).
(3) Most of the work is done by employees of the company.  If you look at the code commits to major crypto projects outside of Bitcoin, most of the work is done by the 10-20 people that largely work for the corporation that started the cryptocurrency project. Is it possible to create incentives for these individuals to stick around as well as for the teams to scale to hundreds of people ongoing[3]? This dynamic is nothing new in the open source world. The open source projects that have tended to do best recently have companies or foundations whose focus is to monetize the protocol or project, and therefore will help pay for ongoing development (Docker, Mozilla/Firefox, Android/Google etc.).

What Can We Learn From Technology Startups? 
The standard technology company model may hold some lessons. This system, while imperfect, evolved over the last 50 years in an attempt to balance incentives and liquidity[4].
(1) Founder equity is pooled into general corporate equity. By keeping value in the company, shares accrete aggregate value that can be distributed to others. Imagine for example if the Google founders directly got their share of 30% of the profits of Google, instead of it going to the Google bottom line. Google overall as a company would be much less valuable and less able to hire and incentivize employees.
(2) There is limited liquidity for founders and employees before the company has a public (and late) liquidity event. Secondary sales by employees and founders are quite limited and are coordinated by the company in a board approved manner.
(3) Venture capital is invested stepwise based on the company hitting milestones that increase company valuation. Instead of raising $200M up front, the company will raise a seed round of a few million, a series A of $10M, and then later rounds that are much larger if the company is working. If the company does not work, it can not raise more money. This Darwinian incentive can be brutal from a founder's perspective but creates incentives to show progress and keep working hard.

The venture model clearly does not solve all the problems mentioned - for example the early employees at Google eventually left the company. However, there is an equity base that can pay for ongoing work. The difference of course is the value of a technology stock is driven by future cash flows of the company, while the value of a crypto company is often solely the value of the token held by the company.

Is it possible to combine token economics with the useful parts of the venture model?

In the crypto world, the parallel brainstorm proposal would be:
1. Put all founder, early developer, company, and foundation equity into a single entity. Lets assume for now it is the corporation, although it could be the foundation. This means 10-30% of the token resides in a single entity that can hire and incentivize employees at scale. At the founding of the project, you can raise a traditional seed round or hybrid SAFE/SAFT (which is what most crypto projects today do).
2. Give mainly equity instead of tokens. Instead of receiving a large major token distribution, most of the compensation that is distributed to founders and employees is in the form of vesting equity in the corporation. As new people join, they are issued shares backed by tokens. If their work increase the value of the token, they directly benefit from it.
3. Raise money based on equity, not tokens. You can still do token sales if you want. The company now has a non-token basis to raise money. If for example NewCoinCo holds $500 million of its own token, it should be able to raise money as a stock sale from investors at that valuation. This could be done as a crowdsale or as a more traditional financing event.

As the token rises in value, the corporation would do another fundraise via an equity sale at a higher valuation and hire more people. In other words, there is now a recurrent, milestone-based funding model for a crypto project besides selling tokens.

The money it raises via equity sales would then go to pay for ongoing development and marketing and the token base is not diluted down. One common issue with some crypto projects is their lack of experienced managers and ability to scale their teams beyond 10-20 people. As part of the equity sales, founders could ensure a subset of investors buying stock would be on the hook to help the company to hire, with marketing and strategy, and other related aspects of the business like a tech investor would.

The big drawback of course is that companies continue to generate revenue, while most crypto projects often have value solely through token value. This means that at some eventual time point you will need to sell or distribute tokens and dilute value of the company.

4. An ICO would still be held to bring capital into the company as well as create a liquid token and community incentives. By raising money in advance of the ICO, it might help create additional clarity between the equity (a security) and the token (effectively an API key for the major projects).

While there are a number of items to be worked out above, this may be a mechanism to increase the number of engineers, community managers, and marketers focused on a protocol over a longer period of time.

The challenge to this hybrid token/equity model is that in a world of nigh-infinite capital, it is hard to defer fast and early liquidity for founders contributors and investors. In some sense it is a giant marshmallow test - how to trade off short term liquidity versus building a persistant corporate entity and scaled team to drive your protocol forward for decades? As a founder of multiple companies myself, I can relate viscerally to the challenges inherent in this question.

Other Ongoing Incentives
Fred Ehrsam has a good post on how to provide ongoing incentives for protocol development including inflation, developer bounties etc. He asks the question of how do you build an evolutionary advantage into the protocol itself, so it will have incentives for it to continue to evolve.

Thinking up front about token distributions and corporate structure may be a complimentary approach to also lay a solid early foundation for the long term contribution and success of a protocol.

Choosing Financing Structures
Dan Romero and Matt Huang laid out a set of interesting comments captured below:

The "joint stock company" concept is a few centuries old. Crypto is a structural shift in many ways, so an open question is whether it should/will also change corporate structures. There might be 3 common structures that will emerge:

A. Token-only funding. Projects that require censorship-resistance as a core feature might need to fund natively via a crowd sale of tokens (as institutional capital may be scared off and there may be value in maintaining anonymity of founders). Similarly, there may be crypto projects where the userbase is inherently skeptical of any institutional involvement - which would also default to token-sales only. Finally, teams may simply decide to forgo venture capital. In all cases, there may still be best practices that emerge that signal quality efforts such as multi-year vesting via smart contracts.

Another approach other then vesting would be to have a fixed budget of tokens per year that are rewarded to the team for their efforts. Existing token holders (outside of the core team) could vote on distribution of tokens/budget per year to the team. This would be something to experiment with and is similar to some of the inflation ideas that Fred Ehrsam mentions.

B. Venture Capital (VC) + token projects. The projects will not require censorship-resistance and may accumulate value via ongoing cash flow (e.g. REP/Augur) or via token value increases. Key questions to ask before going down the token route per Dan: (i) Why do you need your own token versus just using underlying ETH? (ii) Why can't someone just fork your protocol and remove the token?

Some tokens may actually be considered securities by the SEC. In this case, a hybrid VC + token model, or VC-only model is most likely to be used.

C. VC-only financing models. VC-driven financing could occur in a number of scenarios that include non-token business models. For example, if there is no pre-sale and tokens are only available via mining or another distribution mechanism. Obviously, more traditional business models like monetizing via ads would also avoid the need for token.

There is still a lot to sort out in terms of best corporate structure and rewards for crypto projects. The most important things to keep in mind are (1) how do you create sustainable rewards/value that can be used to hire or incentivize teams over decades (2) how to create transparency on economic rewards and incentives.

Thanks to Dan RomeroMatt HuangDavid SacksLucas Ryan, and Warren Buffetcoin for comments on this post.

[1] Obviously, there are a number of other reasons to contribute to a protocol. These include wanting to make an impact, believing in the mission of the protocol, working with other brilliant people, and peer admiration and feedback. This is true of any open source project.

[2] These numbers can mean radically different things due to potential inflation (or lack thereof) of the currency, the number of rounds they plan to do (e.g. Dfinity doing a second round), and the potential to use Dutch auctions or other mechanics (a la Gnosis).  This is also all further complicated by proof of work versus proof of stake systems, as well as miner rewards.

[3] The current way for people to address this is to raise a lot of money via a pre-sale and use it to salary people. They can also give employees tokens or sell more tokens to pay for employees. However, at very large scale this model breaks down - by ongoing token sales you keep diluting the value captured by the company/foundation. If this dilution is faster then the appreciation of the token than you can have a collapse in the ability to pay a great team.

Big Banks and Blockchains
Cryptocurrency's Netscape Moment
End of Cycle?
Its M&A Time!
Market Caps and the 2% Rule
Machine Learning Startups
3 Types Of Platform Companies
Defensibility and Lock-In: Uber and Lyft
Uber And Disruption
Who Cares If Its Been Tried Before?
The Road To $5 Billion Is A Long One
How To Win As Second Mover
End Of Silicon Valley
Social Products
Hot Markets For 2015
Waiting Too Long To Go Public
What Is Your Startup Acquisition Really Worth?
5 Reasons To Sell Your Startup
M&A Ladder

Lead VC Vetos

Friday, September 22, 2017

Twitter Features

Twitter is a product that is used and loved by hundreds of millions of users. The core product has needed the same basic features for the last 7 or 8 years. Here are the things I think Twitter should build. In this post, I stay away from areas I no longer understand for Twitter (e.g. zero rated Twitter / developing world products) or big picture ideas (e.g. how to remake the timeline). Even simple changes could go a long way for the product.

1. Tweet structure.
Long form content. A key rule in product design is if your users keep wanting to do something, let them do it. Tweet storms are clear examples of users wanting to write longer form content on Twitter.
The simplest approach to longer form content would be to allow the linking of tweets as a formal product feature and structure a la tweet storms.

It is clear users want standard longer form content as well. Twitter users will often screenshot a longer post written in e.g. iPhone notes app, and then post the image as their tweet. Obviously this content is now not searchable on Twitter. A simple, blog-like longer form editor may be a solid approach to capture this behavior on Twitter. This may include having the Tweet as the headline for a longer post, or a pull quote. Either way, being able to write a longer piece is one use case the Twitter userbase keeps trying to do.

The actual 140 character nature of Tweets does not need to change in your stream as part of this. E.g. everything could be a standard tweet with the ability to expand to see more inline.

A standard design approach is to get out of the way of your users. Just as Twitter eventually added @ mentions and replies, they should get out of their own way and add long form content. There is clear user demand.

Editable tweets. You should be able to go back and edit a tweet. A simple affordance would be the ability to switch to prior versions of the tweet / tweet history. Lots of wikis, Google docs, and other products have simple ways to reveal this.

2. Content controls.
Muting content areas. I don't care about certain sports but get deluged in my Twitter feed during game time. I should be able to keep following people but turn off specific topics. I would love to take a break from "political Twitter" and turn off politics for a few days. Twitter has the machine learning chops to build a feature like this.

Blocking people. There are classes of hyper aggressive people on Twitter who all attack someone in concert. There should be bulk controls for this. Twitter also has the odd feature that if you block someone, they are notified about it. It is unclear why you need to let someone know if they are blocked. It is the social networking equivalent of giving someone the finger. Twitter can make its platform a much nicer place to exist.

SPAM and fake users. It looks like Twitter is finally working on this feature in earnest. I have instances where I have been followed by 10 new people and 8 of them look like fake users to me. If I can tell a user is fake, a machine learning algorithm should do it even better. Fake users should be automatically removed from the platform.

3. Growth and re-engagement.
Email digest a la Nuzzel. Nuzzel is a news/social content digest that really works. Every day I read 80% of the links that Nuzzel sends me based on my Twitter feeed. Twitter's email digest is largely untargeted and therefore is largely ignored.

Twitter could re-engage me by upping the quality of their email digest, thereby pulling me back to the product daily. Alternatively, they could push me the same content via a notification, or into my Twitter feed. If a Nuzzel-like page were added within Twitter itself I would go there with some frequency. Being able to vote up and down links on such a page could create a distributed / personalized Reddit like experience. Maybe this could replace the "moments" tab?

Overall Twitter is a product whose useful core has evolved only slightly in the last few years. My hope is that some day the above features get added. This should increase utilization, engagement, and growth on a product percentages of humanity use and love daily.

Thanks to David King for comments on this post.

Cryptocurrency's Netscape Moment
End of Cycle?
Its M&A Time!
Market Caps and the 2% Rule
Machine Learning Startups
3 Types Of Platform Companies
Defensibility and Lock-In: Uber and Lyft
Uber And Disruption
Who Cares If Its Been Tried Before?
The Road To $5 Billion Is A Long One
How To Win As Second Mover
End Of Silicon Valley
Social Products
Hot Markets For 2015
Waiting Too Long To Go Public
What Is Your Startup Acquisition Really Worth?
5 Reasons To Sell Your Startup
M&A Ladder

Monday, September 11, 2017

VC Negotiation Tricks: Simplified Term Sheets & Post-Money Valuations

Every 5 years, the VC community comes up with new negotiation tactics that work well against first time entrepreneurs. Two such recent approaches in the VC community are (i) simplified term sheets that lack some of the details needed to assess all terms being offered and (ii) an emphasis on post-money valuations.

(i) Simplified Term Sheets
A few years ago, venture firms would send entrepreneurs a 3 to 5-page terms sheet that would spell out the explicit details of their offer. These term sheets would contain the main highlights entrepreneurs tend to focus on (mainly board and valuation), but also got into the details on terms that are key such as protective provisions (special rights for investors). This meant that signing a term sheet came with an explicit view of all the major items that are part of the round.

More recently, venture firms have been sending 1-page term sheets which have simplified out a number of important details. One term sheet I helped a founder with recently contained only 200 words (this is 10-20% the length of term sheets just a few years ago)! This means the founder was being asked to sign a term sheet without knowing a subset of key terms that would impact future financings, potential company product directions, the ability to exit or sell the company, and limits or rights on employee option pools etc[1].

Once you sign a term sheet as a founder, you lose most of the leverage in a fundraise with a VC. You are typically locked into a 30 day exclusivity period where you can not entertain other VC offers. Founder etiquette at this point is to tell the other funds you have been talking to that you have chosen your financing partner.

Before signing a term sheet, you should have your lawyer fill out the following items in a simplified term sheet:
1. Board structure.
How many common and preferred board seats are there? Is there a CEO-specific seat or would a new CEO seat get created if the founder/CEO steps down or is fired?

One common VC trick is to write the common stock (i.e. founder) board seats as
"The holders of Common Stock shall elect two directors, one of whom will be the then-serving Chief Executive Officer." While the later part of this sentence seems innocuous, it gets interpreted by overzealous lawyers to mean that if a founder ceases to be CEO they lose their board seat which goes to a new CEO (i.e. it is a "CEO" seat, rather then a founder/common board seat. If the board hires in a new external CEO, the founders may immediately lose control of the company.

2. Protective provisions.
Protective provisions are the set of special rights investors in a financing get to protect their minority investment. Protective provisions include[2]:

Future financings. Often VCs ask for the right to block any future preferred stock sales by the company. If you agree to this block, only have it apply to preferred financings with terms worse then the round you just did. In other words, if you get terms equal to, or better, then your current round you should be able to close future financings without approval.

M&A. Most investors ask to be able to block future exits by the company. If you agree to such a term, it should be voted on collectively by all future rounds of stock. I.e. the series A, B, C etc. should all vote together, in a manner representative of shareholdings in the company, on such rounds. Otherwise you can have multiple different VCs, each with their own incentive, block exits.

Company product lines. Some VCs insert language that they can block changes in major lines of business for a company. While this may make sense for a private equity buy out, it does not make sense for an early stage company and should be taken out.

Board. A commonly asked for provision is the ability to change the size or composition of the board. Talk to advisors or angels to get advice on this in the context of your own startup.

Any protective provisions that one round of investors gets (e.g. your Series A) will often carry over into future rounds (series B, C, D, etc.). Be careful in the precedents you set.

3. Liquidation Preference.
1X non-participating preferred is the most typical structure for a series A or B financing. If a VC asks for anything more complex it usually implies something weird about your company (strange re-cap?), your valuation (is it super high prematurely?) or your investors (are they taking advantage of you?).

Ask your lawyer for more details[3].

4. Option pool.
You should shoot for an option pool that allows you to hire the team needed to get you to the next fundraise or major company milestone. It is good to add a little padding to this. To see how option pool may chance effective valuation, see this venture hacks post.

(ii) Valuation: Pre-money versus post-money.
Both pre- and post-money negotiations have been around for decades. While there is not a single right way to do it, many first time entrepreneurs tend to be less savvy in their understanding of a post-money valuation negotiation.

Setting a company valuation by negotiating its pre-money first means that you and the VC agree on a price for the company before the financing occurs. You then add in SAFE and convertible note dilution and new money in, and end up with a total valuation for the company. Negotiating valuation based on post-money means that no matter how many SAFEs or convertible notes are outstanding, or even the total new money raised, the company ends up with the same post-money valuation[4].

A pre-money valuation negotiation tends to be more transparent to a first time entrepreneur. Your company is worth a certain amount before you raise money. Once the money comes in it should be worth more in a manner that scales with capital. E.g. a company that has raised $15M should be worth $5M more than an identical company that raise $10M. Negotiating valuations around post-money tend to distort these economics.

VCs will negotiate the round valuation around the post-money valuation in some cases as it leads to significant additional dilution (and a much lower pre-money) then most first time founders realize:

A. More money in the round can be incredibly dilutive.
B. SAFE Conversion can jack down your pre-money significantly.

Lets take a simple example for (A) above. Suppose you are raising $5M. If you set your post-money at $25M (for a $20M valuation before the round is funded) then you take 20% dilution for the $5M. Say that a few weeks after you sign the term sheet another investor wants to come in for an additional $5M. In a post-money situation, your valuation before the round drops to $15M and you just sold 40% of the company ($10M/$25M).

In contrast, in a pre-money situation the company itself is deemed worth $20M before any money comes in. If you raise $5M, you similarly dilute by 20% and end up with a $25M post. However, when the incremental $5M comes in, you still start off with a $20M pre-money valuation, and end up with a $30M post. In other words, you dilute by 33% instead of 40%. This 7% incremental dilution makes a huge different in ownership.

SAFE notes effectively act the same way the incremental new money coming in above does. They can jack down your effective valuation if you negotiate the valuation on a post-money basis. If you raise $5M in SAFE notes and they convert at a 20% discount, you may have just dropped your pre-money by $6M ($5M + 20% of $5M). Caveat emptor!

Post-money valuation based negotiations have become more common in the last ~2 years and I think is a way for VCs to take benefit from entrepreneurial ignorance around SAFE conversion mechanics and how financing rounds work. As an entrepreneur, you should understand these mechanics and understand the details of your prior financing docs.

[1] A good lawyer should spot and put these terms back in.
[2] There are a number of other protective provisions that I don't mention - e.g. the ability to issue new stock is often blocked by investors to prevent unlimited dilution etc.
[3] I can also understand that VCs probably simplified term sheets to make negotiation faster/simpler with founders and to emphasize the terms first time founders care the most about.
[4] There have always been pre-money and post-money driven term sheets. However, the market has largely been shifting to post-money term sheets more recently. I am guessing this is due to SAFEs. There is nothing "wrong" with a post-money term sheet a priori - although it does create inflexibility in adding money to the round over time and sets a cap on potential fundraise size. Worst case you can always amend your financing docs.

Related Posts
Lead VC Vetos

Wednesday, August 16, 2017

Big Banks And Blockchain

After the crypto run up of 2013, every major bank decided it needed to do something about cryptocurrencies and blockchain. The way many banks responded to this disruptive technology was:
Step 1. Set up a special internal "blockchain group". 
These groups were supposed to explore how banks could make use of blockchain. The cryptocurrencies themselves were largely viewed as speculative and the "interesting parts" were blockchains and distributed ledgers as technologies.

Step 2. Have the blockchain group focus on building private chains. 
Focus at most banks was on private chains and more recently some smart contract work. Nothing was adopted or launched, it was more R&D and prototyping.

Step 3. Have the CEO brag about blockchain group, then do nothing.
When asked by bank's Board of Directors about crypto, the CEO could point to the largely impotent blockchain group and say "we are all over this blockchain thing!!". The Board would break into polite applause and the topic would quickly move on to something more important, like FX risk being hedged and what fancy restaurant to convene to for martinis post-board meeting[1].

It's the Currencies Stupid!
In the last 6 months, large banks & brokerages have started to wake up to the other side of crypto - in particular the currencies themselves[3]. This is largely driven by high net worth clients, pension funds, and others, starting to ask the asset and wealth management divisions of banks how they can participate in BTC (bitcoin), ETH (ethereum), and other cryptocurrencies.

Banks now have an incentive to adopt cryptocurrencies - if it becomes part of the basket of assets they manage for a client, they can take an annual cut. For example, if a bank charges 0.5% of AUM (assets under management) and a client wants to put 5% of their assets into crypto, the bank has a strong incentive to manage it for their client. If the bank pushes the client out to third party sites or wallets, the bank is loosing the 0.5% a year they charge for assets directly under management.

In order to be able to meet the demands of their clients, banks are looking for solutions that allow their clients to participate in BTC et al. in a way that the bank directly manages and controls (or, at a minimum, that the bank can charge for).

The following products need to be built for financial institutions to fully adopt cryptocurrencies:

Mutual funds and ETFs: Financial products that allow people to easily buy a basked of crypto currencies.
Major banks would love to enable their interested clients to buy into a mutual fund or single currency tracking fund as part of their AUM basket alongside various baskets of stocks, bonds, and gold. This prevents the need for them or their client to think about the new crypto hotness and instead the mutual fund can add or drop positions over time. Recent funds for high net worths include Grayscale, ICONOMI, PRISM, the Token Fund, and others.

In parallel a number of efforts to create the first crypto ETF have been ongoing.

Crypto custody / wallet / cold storage.  When talking to private wealth managers about their crypto needs there is strong interest in a wallet where they can store their client's cryptocurrencies, track changes in value for clients, and of course charge their % of AUM. Obviously Coinbase has done amazing work in providing a wallet and cold storage for individuals and have made some interesting moves like their recent Fidelity integration. In parallel, companies like Xapo were founded with the original intention of provided a bitcoin vault. However, from discussions I have had with the wealth management community, there is still not a comprehensive solution in this area.

Hedge funds.
The pioneering early hedge fund of the industry Metastable, and more recently Polychain, have pioneered the crypto hedge fund market. There are a dozen new entrants coming, many of them likely to be crappy.

The banking and brokerage industry would benefit from additional high quality long funds as well as algorithmic trading funds that their clients can participate in, and can be sold to high net worth investors as part of a basket of goods.

Intriguingly a lot of the crypto trading at hedge funds started as young employees trading crypto for their own accounts. This will likely change soon with bank's internal hedge funds also participating directly in the crypto market.

Derivatives exchanges.
As people buy and sell different cryptocurrencies the ability to create complex hedges, options and derivatives becomes increasingly important. This can both dampen volatility in a position as well as create leverage on capital. For both market liquidity to accelerate, and hedge funds to thrive a strong derivatives broker and market needs to emerge. As an example, LedgerX recently received US CFTC (Commodity Future Trading Commision) approval to move ahead in this market. There will likely be more entrants here soon.

With the above product areas fleshed out, we will see an accelerated adoption of cryptocurrencies due to accelerated (and finally real) adoption in the banking and brokerage world.

[1] I tend to imagine board meetings at global banks always ending with martinis for some reason[2]. At least I hope this is what happens, in which case I would be interested in joining a top 5 bank's board. As an aside, New York tends to have much better martinis than San Francisco. I guess most people drink bourbon or rye whiskey these days so this point is moot. As an aside to this aside: Boston has a surprisingly good martini tastiness/per capita ratio.

[2] I also picture all the board members wearing pin striped suits and monocles, sort of like the Monopoly Man.

[3] Blockchains and smart contracts will still change our financial system outside of core cryptocurrencies, it will just take longer.

Cryptocurrency's Netscape Moment
End of Cycle?
Its M&A Time!
Market Caps and the 2% Rule
Machine Learning Startups
3 Types Of Platform Companies
Defensibility and Lock-In: Uber and Lyft
Uber And Disruption
Who Cares If Its Been Tried Before?
The Road To $5 Billion Is A Long One
How To Win As Second Mover
End Of Silicon Valley
Social Products
Hot Markets For 2015
Waiting Too Long To Go Public
What Is Your Startup Acquisition Really Worth?
5 Reasons To Sell Your Startup
M&A Ladder

Monday, August 14, 2017

Unequal Cofounders

One of the big myths in Silicon Valley is that co-founders should be equal. However, if you look at the most successful tech startups of the last 50 years, almost all of them had a dominant co-founder for most of the life of the company. This includes[1]:
  • Amazon. Jeff Bezos.
  • Apple. Steve Jobs famously split equity unequally versus Wozniak. 
  • Facebook. While Zuck had multiple co-founders, the website used to be called "A Mark Zuckerberg production" and he had multiple times the equity and power of his co-founders.
  • Instagram. Kevin Systrom as dominant founder.
  • Intel. Robert Noyce for 7 years and then Gordon Moore for 12 years[2].
  • Intuit. Scott Cook as dominant founder.
  • LinkedIn. Reid Hoffman had multiple co-founders but was really dominant in terms of equity and control (despite hiring a CEO to take over pre-Jeff Weiner).
  • Microsoft. Paul Allen stepped down after a few years leaving Bill Gates as dominant founder.
  • Oracle. Larry Ellison as founder.
  • Pinterest. Ben Silberman has driven the success of the company.
  • Salesforce. Marc Benniof.
  • Uber. Travis Kalanick as primary force until recently.
  • WhatsApp. Jan as dominant founder and equity holder.
Apple, Facebook, Intuit, LinkedIn, Oracle, etc. are all also examples where both power and equity splits between founders were unequal. While people tend to fixate on equity ownership and equality, what really matter is whether there is equal power sharing between founders. In general equal power sharing yields worse outcomes than having an (eventually) dominate cofounder.

The limited set of counterexamples with more equal co-founding partnerships includes Google (confounded a bit by Eric Schmidt being hired as CEO early in its life). Having an equal co-founding relationship is not impossible, just rare for the most successful companies.

Harj Taggar has a great point on this - "Another interesting way of thinking about this is reversion to the mean. Startups need to be outliers in many ways to be an outlying success, one such vector where I think this helps is in product decisions. Any time you involve multiple stakeholders in these decisions (whether it's cofounders/customers/anyone else) you risk having your product revert to the mean (i.e. no one particularly hates it but no one loves it either). Having a product dictator is probably actually optimal for chasing outlying success."

Early Days Are Different
In the early days of a company, it may only be two or three co-founders working together for an extended of period of time before you know what to build. If you have two co-founders and no employees, it may feel natural to always get to consensus and make decisions together. Its just the two of you, and each of you has to be fully bought in due to the giant leap of faith you are taking. While some founding teams always have a clear CEO and decision maker, some do not. As you hire people and raising money, you will need to shift how you think of decision making.

As you grow, you need someone in charge
The two biggest reasons startups fail is running out of money, and co-founder conflicts. Co-founder conflicts tend to arise when there is a lack of clarity on decision making, product vision, and overlapping founder roles. For example, if more then one founder wants to be CEO, or make final decisions on product or other areas, conflict will be inevitable. Alternatively, if founders are willing to truly share power a startup may not be as aggressive due to the need to find compromises versus charge ahead. In these situations you may end up with a product or strategy designed by committee instead of making a single choice on what to build. 

Unequal co-founder relationships are a way to dampen future co-founder issues. By making it clear how decisions are made and who is in charge early, you decrease the likelihood of a founder blow up. This is separate from how you divide equity - ultimately you want to compensate someone for the tough times and many years of pain ahead that comes with starting a company.

This does not mean you will always agree. It is constructive for co-founders (and executives once you have them) to challenge the CEO and each other. There are bound to be lots of disagreements, and the non-CEO may often be right. As the CEO co-founder you need to pick your battles on when to make the final call and override everyone else. All else being equal, most specific decisions are often less important than actually making a decision. As the non-CEO co-founder, you need to understand the importance of falling in line and backing whatever decision is ultimately made.

As an investor, a clear warning sign of future co-founder conflicts is when the founders say they are "equal but own different areas". E.g. "I make all the calls on business, and my cofounder makes all the calls on engineering." But what if a business issue and a product or engineering issue overlap? A company needs a single person who is clearly in charge and can make a decision across all areas. The worst version of this is co-CEOs, which suggests an almost inevitable blow up.

Cofounders may claim "we make all decisions together, and have not fought before, so no need for a single decision maker." This is a recipe for disaster. Startups are hard and the right strategy is usually ambiguous. You can not defer organizational and decision making clarity for later. It is important to decide up front which co-founder is in charge (i.e. the CEO) and that person needs to be able to make decisions without being second guessed.

Making decisions
This is not meant to argue against making important decisions with your co-founders. You are working with them because they are smart, capable people. Their opinions are important and they may be right when you are wrong. They should own areas of the company and be able to make decisions on those areas without micromanagement. However, at some point there needs to be a clear owner who can make a final call if there is an impasse or lack of agreement. Once a decision is made, the non-CEO co-founders need to get behind the decision and make it successful.

Equity splits
As mentioned above many of the most successful companies of the last 20 years had unequal equity splits among co-founders (Facebook, LinkedIn, Twitter, etc.) while some have had equal splits (e.g Google[3], SnapChat). Equal equity does not need to be the default and in some of the most successful companies is not.

Sometimes this is counterintuitive - for example the CEO co-founder does not always get the most equity (e.g. look at the Twitter S1). The key is to be pragmatic and to think through the long term value each person brings to the table, the relative leverage each has, and investment made in different ways.

Other perceived value
One common point of co-founder jealousy and conflict is who gets to meet investors or talk to press. Usually this is the role of the CEO, and early on having multiple people involved in every conversation is not time efficient. As the company scales, there will be plenty of opportunities for multiple co-founders to have external exposure. As CEO, you should also watch out for your co-founders and make sure they also get some exposure if this is important to them.

Thanks to Harj Taggar for feedback on this post.

[1] A number of private companies also do not have equal co-founding relationships or equity splits but since these companies are not yet public it is harder to talk openly about them.

[2] An underreported phenomenon is the number of times one co-founder is replaced by another as CEO. This happened at Intel, Logitech, and other companies.

[2] Google was in reality slightly unequal with Larry Page having slightly more equity when Google went public. It is unclear if this was due to share transfers by Sergey or if it was set up that way.