Sang Ha Park

Recently Bessemer Venture Partners (BVP) released 13 of their internal investment memos. Sharing almost entirety of one’s internal memo is extremely rare in the venture capital scene (though we cannot confirm how much post-edits they have made if any). It’s all the more surprising considering that VC is still regarded as an exclusive club that remains enigmatic. In that I’m grateful that BVP has shared valuable windows into their decision making process and inner-workings. As the memos suggest, it’s more than just a luck to strike the right ones.

Below are some observations & questions that I had. I have bolded key points.

1. Already by 2012, BVP was practicing “data-driven sourcing efforts” (Twitch). To me, this seems extremely pioneering and I can only wonder how they did it then, and how massively they might have improved today.
2. BVP uses the word “comfortable” 5 times in their Twitch memo. All 5 are used to express their level of comfort with the uncertainty. Although this is an extremely minor observation, the word choice indicates that they are a) fully aware of the nature of the business—VC is a outsized bet agains almost infinite risks, b) proactively embracing and juggling the uncertainty, and c) clear delineate what is an acceptable risk and what is not. This is further strengthened by their Probability Table in their modelling.
3. In Yelp, BVP “pre-empts” a round anticipating possible competition and valuation rise due to Business Week press. A glimpse of how they can be aggressive when needed.
4. On a similar note, VC, especially early VC, is really a people’s game. It’s about who you know and how well you know them and vice versa. According to the memo, Jeremy Levine submitted “a termsheet last week and agreed to terms with the company over the weekend”(Yelp). Such pace and decisiveness is only possible because there are very few people involved in decisions—founders and the investor. Which tells us that the relationship that the investor has with founders is really really important, as obvious as it is. Another example is mentioned when they approach Pinterest’s founders leveraging their connection to Jimmy Wales, who apparently is one of the founder’s internet hero.
5. “Track progress” is probably one of the most important phrases in VC. As can be seen in DropCam, Yelp, Fiverr (8+ months), Pinterest (~2 years, thanks to the partner’s “dogged tracking”) and many others, being able to follow and track target’s progress ante investment is one of key traits that venture capitalists need to have. I cannot overstate its importance since unlike public markets, you cannot digest companies’ 3 year history in few hours with financial models and quarterly reports. You physically need to spend the corresponding time. 3 years only yield 3 years. How long you follow often becomes a key factor in turning the conviction switch. VC is a long game and that goes same for before investment as much as for post investment.
6. In MindBody, they compute LTV with “Contribution Margin” which they define as Gross Profit – Renewal Expense – R&D – G&A expenses. Using revenue to compute LTV / Acquisition Cost would have overestimated the multiple, company’s cashflow, and most importantly the underlying unit economics. Using the number that’s actually closer to the operations is a better indicator of whether the business actually works and how much cost it can save when it’s scaled. This was refreshing and comforting as lots of models (both from the company and investors) usually miss this when computing LTVs and BEPs.
7. Furthermore in MindBody, as world-class investors in SaaS businesses, they show amazing sense in estimating the abstract market into a quantifiable number. The following sentence succinctly describes that. “If MindBody can achieve 10% share in salon/spa (compared to its current 30% share in yoga/pilates), we estimate the company can reach \~$80m in ARR over the next \~5 years at its current $100/month average MRR per client.” This is especially inspiring because often times, I have conflated the market size of the product / services with the market size of the customer’s entire business when I’m doing back-of-the-envelope calculations. For instance, you might assume that the market size for MindBody in beauty salons is the revenue that all beauty salons generate rather than the expenses they spend on softwares and HR staffs.
8. One of the more surprising discoveries was that BVP was comfortable in acquiring founder shares (Fiverr, SendGrid, Shopify, and PagerDuty) even at an early stage. To be fair, some of the shares did belong to founders who have already left (Shopify). However, they generally seem to be okay with founders getting a bit of early liquidity. To my experience, it has been generally deemed not recommended, and when I was involved in one such transaction, it bugged me quite a bit (still does to this day). I wonder what criteria BVP has for early liquidity of existing founders. One criterion might be hitting the internal target shareholding as indicated in Fiverr. They seem to aim for 15 ~ 20% circa Series B.
9. As seen from LifeLock and Wix, BVP seems to take advantage of warrants pretty frequently. Warrants can be a very effective mean of increasing returns and shareholding with little downside and no upfront commitment. I’m not sure how popular warrants are now except for special situations since the ample liquidity and intense competition has driven the pendulum to companies’ favor.
10. As most firms now do, BVP also develops theses and lays out a product / services roadmap in a given space. For instance, they chose LifeLock as part of “a consumer security roadmap”. In my experience, such roadmap was crucial in identifying product vacuums and pre-emptively looking for companies to invest. In short, you know what you are looking for.
11. One of the most striking details is BVP’s use of probability table to estimate their exit. While we don’t get the full picture of how they arrive with their probability, the fact that they assign probability to different scenarios is very telling. First of all, it tells that even the best investors cannot predict company’s future and can only assign probabilities to it. Second, it shows that they have a way of quantifying the risk so that they can revisit, reflect, and revise their model. This is a prime example of what Annie Duke says in her book “Thinking in Bets”.
12. BVP value-adds even before investment! From LifeLock, “After AT&T exposed consumer data, we introduced Life Lock to their Chief Security Officer (thank you, Byron) and Life Lock is meeting with AT&T to discuss opportunities”
13. BVP is very happy to lead or take the entire round. They took $8 million for MindBody’s $11 million round; $19.6 million from $27.2 million round for PageDuty with an additional $6 million to buy founder shares; entire $12.5 million round for LinkedIn. From what I’ve heard, such level of conviction is pretty hard to find in Korean venture capital space. I hope we have more of that. I believe that each unique investor represents each unique opportunity and vision.

Finally, having looked at BVP’s successful investments, I can only wonder at how their failed investments (I’m sure there are) went and how their memos were crafted. As they say in the preface, “For VCs, learning from mistakes is easy. In fact, we make so many that we’re experts on what not to do.” It would be enlightening to see how they have learned from their failed investments and adjusted their mental model accordingly.