Tag: michael mauboussin

Posted on

The Science of Selling – Early DPI Benchmarks

The snapshot

Some of you reading here are busy, so we’ll keep this top part brief, as an abstract sharing our top three observations of leading fund managers.

Generally speaking, don’t sell your fast growing winners early.

Except when…

Selling on your way up may not be a crazy idea.

  1. You might sell when you want to lock in DPI. Don’t sell more than 20% of your fund’s positions unless you are locking in meaningful DPI for your fund. For instance, at each point in time, something that’s greater than 0.5X, 1X, 2X, or 3X of your fund size.
  2. You might consider selling when you’ve lost conviction. Consider selling a position when you feel the market has over-priced the actual value, or even up to 100% if you’ve lost conviction.
  3. You might consider selling when one is growing slower than your target IRR. If companies are growing slower and even only as fast as your target IRR, consider selling if not at too much of a discount (Note: there may be some political and/or signaling issues to consider here as well. But will save the topic of signaling for another blog post).

Do note that the above are not hard and fast rules. Every decision should be made in context to other moving variables. And that the numbers below are tailored to early-stage funds.

Net TVPI Benchmarks from Years 5-15
Net DPI Benchmarks from Years 5-15

Let’s go deeper…

On a cloudless Friday morning, basking in the morning glory of Los Altos, between lattes and croissants, between two nerds (or one of whom might identify as a geek more than a nerd), we pondered one question:

How much of selling is art? How much is science?

Between USV selling 30% of their Twitter stake, Menlo selling half of their Uber, Benchmark only selling 15% of their Uber pre-IPO shares, and Blackbird recently selling 20% of its Canva stake, it feels more like the former than the latter. Then when Howard Marks says selling is all about relative selection and the opportunity cost of not doing so, it seems to reinforce the artistic form of getting “moolah in da coolah” to borrow a Chris Douvos trademark.

Everyone seems to have a financial model for when and how to invest, but part of being a fiduciary of capital is also knowing when to distribute – when to sell. When RVPI turns into DPI. And we haven’t seen many models for selling yet. At least none have surfaced publicly or privately for us.
The best thought piece we’ve seen in the space has been Fred Wilson’s Taking Money “Off the Table”. At USV, they “typically seek to liquidate somewhere between 10% and 30% of our position in these pre-IPO liquidity transactions. Doing so allows us to hold onto the balance while de-risking the entire investment.”

Source: Fred Wilson’s Taking Money “Off The Table”

In aggregate, we’ve seen venture fund distributions follow very much of the power law – whether you’re looking at Correlation’s recent findings…

Source: Correlation Ventures

Or what James Heath has found across 1000+ firms’ data on Pitchbook.

Source: James Heath

As such, it gave birth to a thought… What if selling was more of a science?

What would that look like?

Between two Daves, it was not the Dave with sneakers and a baseball cap and with the profound disregard to healthy diets, given the fat slab of bacon in his croissan’wich, who had the answer there.

“To start off, in a concentrated portfolio of 30 investments, a fund returner is a 30x investment. For a 50-investment fund, it’s 50x. And while hitting the 0.5x DPI milestone by years 5-8, and a 2x DPI milestone by years 8-12, is the sign of a great fund, you shouldn’t think about selling much of your TVPI for DPI unless or until your TVPI is starting to exceed 2-3x.” Which seems to corroborate quite well with Chamath Palihapitiya’s findings that funds between 2010 and 2020 convert have, on average, converted about 25% of their TVPI to DPI.

“Moreover, usually you shouldn’t be selling more than 20% of the portfolio at one time (unless you’re locking in / have already locked in 3X or more DPI). You should be dollar-cost averaging – ensuring time diversity – on the way out as well. AND usually only if a company that’s UNDER-growing or OVER-valued compared to the rest of your portfolio. Say your portfolio is growing at 30% year-over-year, but an individual asset is growing slower at only 10-20% OR you believe it is overvalued, that’s when you think about taking cash off the table. Sell part (or even all) of your stake, if selling returns a meaningful DPI for the fund, and if you’re not capping too upside in exchange for locking in a floor.”

Meaningful DPI, admittedly, does mean different benchmarks for different kinds of LPs. For some, that may mean 0.25X. For others that may mean north of 0.5X or 1X.

“On the other hand, if a company is outperforming / outgrowing the rest of the portfolio, generally hold on to it and don’t sell more than 10-20% (again, unless you’re locking in meaningful DPI, or perhaps if it’s so large that it has become a concentration risk).”

I will caveat that there is great merit in its counterpart as well. Selling early is by definition capping your upside. If you believe an asset is reaching its terminal value, that’s fine, but do be aware of signaling risk as well. The latter may end up being an unintended, but self-fulfilling prophecy.

So, it begged the question: Under the assumption that funds are 15-year funds, what is meaningful DPI? TVPI? At the 5-year mark? 7.5 years in? 10 years? And 12.5 years?

The truth is the only opportunities to sell come from the best companies in your portfolio. And probably the companies, if anything, you should be holding on to. By selling early, you are capping your downside, but at the same time capping your upside on the entire portfolio. When the opportunity arises to lock in some DPI, it’s worth considering the top 3-5 positions in your fund. For instance, if your #2 company is growing quickly, you may not be capping the upside as much.

Do keep in mind that sometimes it’s hard to fully conceptualize the value of compounding. As one of my favorite LPs reminded me, if an asset is growing 35% year-over-year, the last 20% of the time produces 56% of the return. Or if an asset is growing 25% YoY, if you sell 20% earlier (assuming 12 year time horizons), you’re missing out on 45% of the upside.

As a GP, you need to figure out if you’re IRR or multiple focused. Locking in early DPI means your IRR will look great, but your overall fund multiple may suffer.

As an LP, that also means if the gains are taxable (meaning they don’t qualify for QSBS or are sold before QSBS kick in), you need to pay taxes AND find another asset that’s compounding at a similar or better rate. As Howard Marks puts it, you need to find another investment with “superior risk-adjusted prospective returns.”

And so began the search for not just moolah in da coolah, but how much moolah in da coolah is good moolah in da coolah? And how much is great?

Net TVPI Benchmarks from Years 5-15
Net DPI Benchmarks from Years 5-15

Some caveats

Of course, if you’ve been around the block for a minute, you know that no numbers can be held in isolation to others. No facts, no data points alienated from the rest.

Some reasons why early DPI may not hold as much weight:

  • Early acqui-hires. Usually not a meaningful DPI and a small, small fraction of the fund.
    • There’s a possibility this may be the case for some 2020-2021 vintages, as a meaningful proportion of their portfolio companies exit small but early.
    • In other words, DPI is constructed of small, but many exits, rather than a meaningful few exits.
  • TVPI is less than 2-3x of DPI, only a few years into the fund. In other words, their overall portfolio may not be doing too hot. Obviously, the later the fund is to its term, the more TVPI and DPI are alike.
  • As a believer in the power law, if on average it takes an outlier 8 years to emerge AND the small percentage of winners in the portfolio drive your return, your DPI will look dramatically different in year 5 versus 10. For pre-seed and seed funds, it’s fair to assume half (or more) companies go to zero within the first 3-5 years. And in 10 years, more than 80% of your portfolio value comes from less than 20% of your companies. Hell, it might even be 90% of your portfolio value comes from 10% of your companies. In other words, the power law.
  • GPs invested in good quality businesses. Some businesses may not receive markups, but may be profitable already, or growing consistently year-over-year that they don’t need to raise another round any time soon.
  • Additionally, if you haven’t been in the investing game for long, persistence of track record, duration, and TVPI may matter more in your pitch. If you’ve been around the block, IRR and DPI will matter more.
  • As the great Charlie Munger once said, “selling for market-timing purposes actually gives an investor two ways to be wrong: the decline may or may not occur, and if it does, you’ll have to figure out when the time is right to go back in.” For private market investors, unless you can buy secondaries, you’ll never have a time to go back in until the public offering. As such, it is a one-way door decision.

Some LPs are going to boast better portfolios, and we do admit there will be a few with portfolios better than the above “benchmarks.” And if so, that’s a reason to be proud. In terms of weighting, as a proponent of the power law, there is a high likelihood that we’ve underestimated the percent of crap and meh investments, and overestimated the percent of great investments in an LP’s portfolio. That said, that does leave room for epic fund investments that are outliers by definition. 

We do admit that, really, any attempt to create a reference point for fund data before results speak for themselves is going to be met with disagreement. But we also understand that it is in the discourse, will we find ourselves inching closer to something that will help us sleep better at night.

One more caveat for angels… The truth is as an angel, none of the above really matter all that much. You’re not a fiduciary of anyone else’s capital. And your time horizons most likely look different than a fund’s. It’s all yours. So it’s not about capping your downside, but more so about capping your regret. In other words, a regret minimization framework (aka, “spouse regret/yelling minimization insurance”). 

That will be so unique to you that there is no amount of cajoling that we could do here to tell you otherwise. And that your liquidity timelines are only really constrained by your own liquidity demands.. For instance, buying a new home, sending kids to college, or taking care of your parents (or YOU!) in their old age.

But I do think the above is a useful exercise to think through selling if you had a fund. You would probably break it down more from a bottoms up perspective. What is your average check size? Do you plan to have a concentrated portfolio of sub-30 investments? Or more? Do you plan to follow on? How much if so? And that is your fund size.

In closing

Returning above a 3x DPI is tough. Don’t take our words for it. Even looking at the data, only 12.5% of funds return over a 3x DPI. And only 2.5% return three times their capital back on more than 2 separate funds.

In the power law game we play, as Michael Mauboussin once said, “A lesson inherent in any probabilistic exercise: the frequency of correctness does not matter; it is the magnitude of correctness that matters.” Most will return zero, or as Jake Kupperman points out: More than 50%.

Source: Jake Kupperman’s The Time Has Come to Modernize the Venture Capital Fund of Funds

But it’s in the outliers that return meaningful DPI, not the rest. Not the acqui-hire nor really that liquidation preference on that small acquisition.

At the end of the day, the goal isn’t for any of the above to be anyone’s Bible, but that it’d start a conversation about how people look at early returns. If there is any new data points that are brought up as a result of this blogpost, I’ll do my best to update this thread post-publication.

Big thank you to Dave McClure for inspiring and collaborating on this piece, and to Eric Woo and all our LP friends who’ve helped with the many revisions, sharing data, edits, language and more. Note: Many of our LP friends chose to stay anonymous but have been super helpful in putting this together.

Footnotes

For the purpose of this piece, we know that “good” and “great”, in fact all of the superlative adjectives, are amorphous goalposts. And those words may mean different things to different people. This blogpost isn’t meant to establish a universal truth, but rather serve as a useful reference point for both LPs, looking for “benchmarking” data, and GPs to know where they stand. For the latter, if your metrics do fall in the “good” to “great” range, they’re definitely worth bragging about.

And so with that long preamble, in the piece above, we defined “good” as top quartile, and “great” as top decile. “Good” as a number on its own, enough for an LP to engage in a conversation with you. And “great” as a number that’ll make LPs running to your doorstep. Or at least to the best of our portfolios, leveraging both publicly reported and polled numbers as well as our own.

Our numbers above are also our best attempt in predicting steady state returns, divorcing ourselves from the bull rush of the last 3-5 vintage years. As such, we understand there are some LPs that prefer to do vintage benchmarking, as opposed to steady state benchmarking. And this blogpost, while it has touched on it, did not focus on the former’s numbers.

EDIT (Aug 18, 2023): Have gotten a few questions about where’s the data coming from. The above numbers in the Net DPI and Net TVPI charts are benchmarks the LPs and I agreed on after looking into our own anecdotal portfolios (some spanning 20+ years of data), as well as referencing Cambridge data. These numbers are not the end-all-be-all, and your mileage as an LP may very much vary depending on your portfolio construction. But rather than be the Bible of DPI/TVPI metrics, the purpose of the above is give rough reference points (in reference to our own portfolios + public data) for those who don’t have any reference points.

Cover Photo by Renate Vanaga on Unsplash

Stay up to date with the weekly cup of cognitive adventures inside venture capital and startups, as well as cataloging the history of tomorrow through the bookmarks of yesterday!

The views expressed on this blogpost are for informational purposes only. None of the views expressed herein constitute legal, investment, business, or tax advice. Any allusions or references to funds or companies are for illustrative purposes only, and should not be relied upon as investment recommendations. Consult a professional investment advisor prior to making any investment decisions.

Posted on

Chasing Revenue Multiples and Revenue

unicorn, sunset

On Wednesday this week, I hosted an intimate dinner with founders in the windy backdrop of San Francisco. And I’m writing this piece, I can’t help but recall one founder from that evening asking us all to play a little game she built. A mini mobile test to see if we could tell the difference between real headshot portraits and AI-generated ones based on the former. There were 15 picture. Each where we had to pick one of two choices: real or AI.

10/15. 6/15. 9/15. 11/15. 8/15… By the time it was my turn, having seen the looks of confusion of my predecessors, I wasn’t confident in my own ability to spot the difference. Then again, I was neither the best nor the worst when it came to games of Where’s Waldo? 90 quick seconds later, a score popped up. 10/15. Something slightly better than chance.

Naturally, we asked the person who got 11/15 if he knew something we didn’t. To which, he shared his hypothesis. A seemingly sound and quite intellectual conjecture. So, we asked him to try again to see if his odds would improve. 90 seconds later, 6/15.

Despite the variance in scores, none were the wiser.

Michael Mauboussin shared a great line recently. “Intuition is a situation where you’ve trained your system one in a particular domain to be very effective. For that to work, I would argue that you need to have a system, so this is the system level, that it’s fairly linear and stable. So linear in that sense, I mean really the cause and effect are pretty clear. And stable means the basic rules of the game don’t change all that much.”

For our real-or-AI game, we lacked that clear cause and effect. If we received individual question scores of right or wrong, we’d probably have ended up building intuition more quickly.

Venture is unfortunately an industry that is stable, but not very linear. In many ways, you can do everything right and still not have things work out. That same premise led to another interesting thread I saw on Twitter this week by Harry Stebbings.

In a bull market, and I was guilty of this myself, the most predictable trait came in two parts: (a) mark-ups (and graduation rates to the next round), and (b) unicorn status. In 2020 and 2021, growth equity moved upstream to win allocation when they needed it with their core check and stage. But that also meant they were less price-sensitive and disciplined in the stages preceding their core check.

The velocity of rounds coming together due to a combination of FOMO and cheap cash empowered founders to raise quickly and often. Sometimes, in half the funding window during a disciplined market. In other words, from 18 months to 9 months. Subsequently, investors found themselves with 70+% IRR and deploying capital twice or thrice as fast as they had promised their LPs. In attempts to keep up and not get priced out of deals. Many of whom believed that to be the new norm.

While the true determinant of success as an investor is how much money you actually return to your investors, or as Chris Douvos calls it moolah in da coolah, the truth is all startup investors play the long game. Games that last at least a decade. Games that are stable, but not linear. The nonlinearity, in large part, due to the sheer number of confounding variables and the weight distribution changing in different economic environments. A single fund often goes through at least one bull run and one bear run. So, because of the insanely long feedback loops and venture’s J-curve, it’s often hard to tell.

Source: Crunchbase

In fact, in recent news, Business Insider reported half of Sequoia’s funds since 2018 posted “losses” for the University of California endowment. We’re in the beginning of 2023. In other words, we’re at most five years out. While I don’t have any insider information, time will tell how much capital Sequoia will return. For now, it’s too early to pass any judgment.

The truth is most venture funds have yet to return one times their capital to their investors within five years. Funds with early exits and have a need to prove themselves to LPs to raise a subsequent fund are likely to see early DPI, but many established funds hold and/or recycle carry. Sequoia being one of the latter. After all, typical recycling periods are 3-4 years. In other words, a fund can reinvest their early moolah in da coolah in the first 3-4 years back into the fund to make new investments. There is a dark side to recycling, but a story for another time. Or a read of Chris Neumann’s piece will satiate any current surplus of curiosity.

But I digress.

In the insane bull run of 2020 and 2021, the startup world became a competition of who could best sell their company’s future as a function of their — the founders’ — past. It became a world where people chased signal and logos. A charismatic way to weave a strong narrative behind logos on a resume seemed to be the primary predictors of founder “success.” And in a market with a surplus of deployable capital and heightened expectations (i.e. 50x or higher valuation multiples on revenue), unicorn status had never been easier to reach.

As of January of this year — 2023, if you’re a time traveler from the future, there are over 1,200 unicorns in the world. 200 more than the beginning of 2022. Many who have yet to go back to market for cash, and will likely need a haircut. Yet for so many funds, the unicorn rate is one of the risks they underwrite.

I was talking with an LP recently where he pointed out the potential fallacy of a fund strategy predicated on unicorn exits. There have only been 118 companies that have historically acquired unicorns. And only four of the 118 have acquired more than four venture-backed unicorns. Microsoft sitting at 12. Google at 8. And Meta and Amazon at 5 each. Given that a meaningful percentage of the 1200 unicorns will need a haircut in their next fundraise, like Stripe and Instacart, we’re likely going to see a slowdown of unicorns in the foreseeable future. And for those on the cusp to slip below the unicorn threshold. Some investors have preemptively marked down their assets by 25-30%. Others waiting to see the ball drop.

The impending future is one not on multiples but one of business quality, namely revenue and revenue growth. All that to say, unless you’re growing the business, exit opportunities are slim if you’re just betting on having unicorn acquisitions in your portfolio.

So while many investors will claim unicorn rate as their metric for success, it’s two degrees of freedom off of the true North.

In the bear market we are in today, the world is now a competition of the quality of business, rather than the quality of words. At the pre-seed stage, companies who are generating revenue have no trouble raising, but companies who don’t are struggling more.

As Andy Rachleff recently pointed out, “Valuations are not the way you judge a venture capitalist, or multiples of their fund. […] The way that I judge a venture capitalist is by how many companies did they back that grew into $100M revenue businesses.” If you bring in good money, whether an exit to the public market or to a partner, you’re a business worth acquiring. A brand and hardly any revenue, if acquired, is hardly going to fetch a good price. And I’ve heard from many LPs and longtime GPs that we’re in for a mass extinction if businesses don’t pivot back to fundamentals quickly. What are fundamentals? Non-dilutive cash in the bank. In other words, paying customers.

Bull markets welcome an age of chasing revenue multiples (expectation and sentiment). Bear markets welcome an age of chasing revenue.

The latter are a lot more linear and predictable than the former.

Photo by Paul Bill on Unsplash

Stay up to date with the weekly cup of cognitive adventures inside venture capital and startups, as well as cataloging the history of tomorrow through the bookmarks of yesterday!

The views expressed on this blogpost are for informational purposes only. None of the views expressed herein constitute legal, investment, business, or tax advice. Any allusions or references to funds or companies are for illustrative purposes only, and should not be relied upon as investment recommendations. Consult a professional investment advisor prior to making any investment decisions.

Proudly powered by WordPress