The Science of Re-Upping

baseball, follow on

Soooooooo… (I know, what a great word to start a blogpost) I started this essay, with some familiarity on one subject. Little did I know I was going to learn about an entirely different industry, and be endlessly fascinated about that.

The analogy that kicked off this essay is that re-upping on a portfolio company is very much like re-signing a current player on a sports team. That was it. Simple as it was supposed to sound. The goal of any analogy was to frame a new or nuanced concept, in this case, the science of re-upping, under an umbrella of knowledge we were already familiar with.

But, I soon learned of the complexity behind re-upping players’ contracts, as one might assume. And while I will claim no authority over the knowledge and calculations that go into contracts in the sports arena, I want to thank Brian Anderson and everyone else who’s got more miles on their odometer in the world of professional sports for lending me their brains. Thank you!

As well as Arkady Kulik, Dave McClure, and all the LPs and GPs for their patience and willingness to go through all the revisions of this blogpost!

While this was a team effort here, many of this blogpost’s contributors chose to stay off the record.


The year was 1997.

Nomar Garciaparra was an instantaneous star, after batting an amazing .306/.342/.534. For the uninitiated, those are phenomenal stats. On top of batting 30 home runs and 11 triples – the latter of which was a cut above the rest of the league, it won him Rookie of the Year. And those numbers only trended upwards in the years after, especially in 1999 and 2000. Garciaparra became the hope for so many fans to end the curse of the Bambino – a curse that started when the Red Sox traded the legendary Babe Ruth to the Yankees in 1918.

Then 2001 hit. A wrist injury. An injured Achilles tendon. And the fact he needed to miss “significant time” earned him a prime spot to be traded. Garciaparra was still a phenomenal hitter when he was on, but there was one other variable that led to the Garciaparra trade. To Theo Epstein, above all else, that was his “fatal flaw.”

Someone that endlessly draws my fascination is Theo Epstein. Someone that comes from the world of baseball. A sport that venture draws a lot of inspiration, at least in analogy, like one of my fav sayings, Venture is one of the only types of investments where it’s not about the batting average but about the magnitude of the home runs you hit.

If you don’t follow baseball, Theo Epstein is the youngest general manager in the history of major league baseball at 26. But better known for ending the Curse of the Bambino, an 86-year curse that led the Red Sox down a championship drought that started when the Red Sox traded Babe Ruth to the Yankees. Theo as soon as he became general manager traded Nomar Garciaparra, a 5-time All-star shortstop, to the Cubs, and won key contracts with both third baseman Bill Mueller and pitcher Curt Schilling. All key decisions that led the Red Sox to eventually win the World Series 3 years later.

And when Theo left the Red Sox to join the Chicago Cubs, he also ended another curse – The Curse of the Billy Goat, ending with Theo leading them to a win in the 2016 World Series. You see, in baseball, they measure everything. From fly ball rates to hits per nine innings to pitches per plate appearance. Literally everything on the field.

But what made Theo different was that he looked at things off the field. It’s why he chose to bet on younger players than rely on the current all-stars. It’s why he measures how a teammate can help a team win in the dugout. And, it’s why he traded Nomar, a 5-time All Star, as soon as he joined, because Nomar’s “fatal flaw” was despite his prowess, held deep resentment to his own team, the Sox, when they tried to trade him just the year prior for Alex Rodriguez but failed to.

So, when Danny Meyer, best known for his success with Shake Shack, asked Theo what Danny called a “stupid question”, after the Cubs lost to the Dodgers in the playoffs, and right after Houston was hit by a massive hurricane, “Theo, who are you rooting for? The Dodgers so you can say you lost to the winning team, or Houston (Astros), because you want something good to happen to a city that was recently ravaged by a hurricane.”

Theo said, “Neither. But I’m rooting for the Dodgers because if they win, they’ll do whatever every championship team does and not work on the things they need to work on during the off season. And the good news is that we have to play them 8 times in the next season.”

You see, everyone in VC largely has access to the same data. The same Pitchbook and Crunchbase stat sheet. The same cap table. And the same financials. But as Howard Marks once said in response how you gain a knowledge advantage:

“You have to either:

  1. Somehow do a better job of massaging the current data, which is challenging; or you have to
  2. Be better at making qualitative judgments; or you have to
  3. Be better at figuring out what the future holds.”

For the purpose of this blogpost, we’re going to focus on the first one of the three.

To begin, we have to first define a term that’ll be booking its frequent flier miles for the rest of this piece – expected value.

Some defined it as the expectation of future worth. Others, a prediction of future utility. Investopedia defines it as the long-term average value of a variable. Merriam-Webster has the most rudimentary definition:

The sum of the values of a random variable with each value multiplied by its probability of occurrence

On the other hand, venture is an industry where the beta is arguably one of the highest. The risk associated with outperformance is massive as well. And the greatest returns, in following the power law, are unpredictable.

We’re often blessed with hindsight bias, but every early-stage investor in foresight struggles with predicting outlier performance. Any investor that says otherwise is either deluding you or themselves or both. At the same time, that’s what makes modeling exercises so difficult in venture, unlike our friends in hedge funds and private equity. Even the best severely underestimate the outcomes of their best performers. For instance, Bessemer thought the best possible outcome for Shopify was $400M with only a 3% chance of occurring.

Similarly, who would have thought that jumping in a stranger’s car or home, or live streaming gameplay would become as big as they are today. As Strauss Zelnick recently said, “The biggest hits are by their nature, unexpected, which means you can’t organize around them with AI.” Take the word AI out, and the sentence is equally as profound replaced with the word “model.” And it is equally echoed by others. Chris Paik at Pace has made it his mission to “invest in companies that can’t be described in a single sentence.”

But I digress.

Value itself is a huge topic – a juggernaut of a topic – and I, in no illusion, find myself explaining it in a short blogpost, but that of which I plan to spend the next couple of months, if not years, digging deeper into, including a couple more blogposts that are in the blast furnace right now. But for the purpose of this one, I’ll triangulate on one subset of it – future value as a function of probability and market benchmarks.

In other words, doubling down. Or re-upping.

For the world of startups, the best way to explain that is through a formula:

E(v) = (probability of outcome) X (outcome)

E(v) = (graduation rate) X (valuation step up from last round) X (dilution)

For the sake of this blogpost and model, let’s call E(v), appreciation value. So, let’s break down each of the variables.

What percent of your companies graduate to the next round? I shared general benchmarks in this blogpost, but the truth is it’s a bit more nuanced. Each vertical, each sub-vertical, each vintage – they all look different. Additionally, Sapphire’s Beezer recently said that it’s normal to expect a 20-30% loss ratio in the first five years of your fund. Not all your companies will make it, but that’s the game we play.

On a similar note, institutional LPs often plan to build a multi-fund, multi-decade relationship with their GPs. If they invest in a Fund I, they also expect to be there by Fund III.

How much greater is the next round’s valuation in comparison to the one in which you invested? Twice as high? Thrice? By definition, if you double down on the same company, rather than allocate to a net new company, you’re decreasing your TVPI. And as valuations grow, the cost of doubling down may be too much for your portfolio construction model to handle, especially if you’re a smaller sub-$100M fund.

It’s for the same reason that in the world of professional sports, there are salary caps. In fact, most leagues have them. And only the teams who:

  • Have a real chance at the championship title.
  • Have a lot in their coffers. This comes down to the composition of the ownership group, and their willingness to pay that tax.
  • And/or have a city who’s willing to pay the premium.

… can pay the luxury tax. Not to be too much of a homer, but the Golden State Warriors have a phenomenal team and are well-positioned to win again (at least at the time of this blogpost going out). So the Warriors can afford to pay the luxury tax, but smaller teams or teams focused on rebuilding can’t.

The Bulls didn’t re-sign the legendary Michael Jordan because they needed to rebuild. Indianapolis didn’t extend Peyton Manning’s contract ‘cause they didn’t have the team that would support Peyton’s talents. So, they needed to rebuild with a new cast of players.

Similarly, Sequoia and a16z might be able to afford to pay the “luxury tax” when betting on the world’s greatest AI talent and for them to acquire the best generative AI talent. Those who have a real chance to grow to $100M ARR, given adoption rates, retention rates, and customer demand. But as a smaller fund or a fund that has a new cast of GPs (where the old guard retired)… can you?

If a star player is prone to injury or can only play 60 minutes of a game (rather than 90 minutes), a team needs to re-evaluate the value of said player, no matter how talented they are. How much of a player’s health, motivation, and/or collaborativeness – harkening back to the anecdote of Nomar Garciaparra at the beginning – will affect their ability to perform in the coming season?

Take, for instance, the durability of a player. If there ‘s a 60% chance of a player getting injured if he/she plays longer than 60 minutes in a game and a 50% of tearing their ACL, while they may your highest scorer this season, they’re not very durable. If that player missed 25% of practices and 30% of games, they just don’t have it in them to see the season through. And you can also benchmark that player against the rest of the team. How’s that compared with the team’s average?

Of course, there’s a parallel here to also say, every decision you make should be relative to industry and portfolio benchmarks.

How great of a percentage are you getting diluted with the next round if you don’t maintain your ownership? This is the true value of your stake in the company as the company grows.

E(v) = (graduation rate) X (valuation step up from last round) X (dilution)

If the expected value is greater than one, the company is probably not worth re-upping. And that probably means the company is overhyped, or that that market is seeing extremely deflated loss ratios. In other words, more companies than should be, are graduating to the next stage; when in reality, the market is either a winner-take-all or a few-take-all market. If it is less than or equal to one, then it’s ripe to double down on. In other words, the company may be undervalued.

And to understand the above equation or for it to be actually useful (outside of an abstract concept), you need market data. Specifically, around valuation step ups as a function of industry and vertical.

If you happen to have internal data across decades and hundreds of companies, then it’s worth plugging in your own dataset as well. It’s the closest you can get to the efficient market frontier.

But if you lack a large enough sample size, I’d recommend the below model constructed from data pulled from Carta, Pitchbook, and Preqin and came from the minds of Arkady Kulik and Dave McClure.

The purpose of this model is to help your team filter what portfolio companies are worth diving deeper into and which ones you may not have to (because they didn’t pass the litmus test) BEFORE you evaluate additional growth metrics.

It is also important to note that the data we’ve used is bucketed by industry. And in doing so, assumptions were made in broad strokes. For example, deep tech is broad by design but includes niche-er markets that have their own fair share of pricing nuances in battery or longevity biotech or energy or AI/ML. Or B2B which include subsectors in cybersecurity or infrastructure or PLG growth.

Take for instance…

Energy sector appreciation values and follow-on recommendations

The energy sector sees a large drop in appreciation value at the seed stage, where all three factors contribute to such an output. Valuation step-up is just 1.71X, graduation rates are less than 50% and dilution is 38% on average.  

Second phase where re-upping might be a good idea is Series B. Main drivers as to such a decision are that dilution hovers around 35% and about 50% of companies graduate from Series A to Series B. Mark ups are less significant where we generally see only an increase in valuation at about 2.5X, which sits around the middle of the pack.

Biotech sector appreciation values and follow-on recommendations

The biotech sector sees a large drop in appreciation value at the Seed stage. This time, whereas dilution seems to match the pace of the rest of the pack (at an average of 25%), the two other factors shine greater in making a follow-on decision. Valuation step up are rather low, sitting at 1.5X. And less than 50% graduate to the next stage.

In the late 2023 market, one might also consider re-upping at the Series C round. Main driver is the unexpectedly low step-up function of 1.5X, which matches the slow pace of deployment for growth and late stage VCs. On the flip side, a dilution of 17% and graduation rate of 60% are quite the norm at this stage.

All in all, the same exercise is useful in evaluating two scenarios – either as an LP or as a GP:

  1. Is your entry point a good entry point?
  2. Between two stages, where should you deploy more capital?

For the former, too often, emerging GPs take the stance of the earlier, the better. Almost as if it’s a biblical line. It’s not. Or at least not always, as a blanket statement. The point of the above exercise is also to evaluate, what is the average value of a company if you were to jump in at the pre-seed? Do enough graduate and at a high enough price for it to make sense? While earlier may be true for many industries, it isn’t true for all, and the model above can serve as your litmus test for it. You may be better off entering at a stage with a higher scoring entry point.

For the latter, this is where the discussion of follow on strategies and if you should have reserves come into play. If you’re a seed stage firm, say for biotech, using the above example, by the A, your asset might have appreciated too much for you to double down. In that case, as a fund manager, you may not need to deploy reserves into the current market. Or you may not need as large of a reserve pool as you might suspect. It’s for this reason that many fund managers often underallocate because they overestimate how much in reserves they need.

If you’re curious to play around with the model yourself, ping Arkady at ak@rpv.global, and you can mention you found out about it through here. 😉

Photo by Gene Gallin 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.

How to Retain Talent When You Don’t Have the Cash

lightning in a bottle, spark, hold, light, jar

Earlier this week, I grabbed coffee with a founder. Let’s call him “Elijah.” He recently lost a key exec he’d been working with for two years to their incumbent competitor. The competitor’s offer happened to be too good to turn down. Triple the exec’s salary. As that exec had a family to feed and children’s education that didn’t come cheap, he made the hard decision to leave. Needless to say, Elijah was devastated. And he asked,

“David, what should I have done?”

I initially thought it was rhetorical. It seemed that way. But he paused, looked at me, and waited.

So I responded.

My response

I’ll preface by saying that the advice I shared with him was a collection of insights I learned from mentors over the years — some a lot more recently than others. I don’t hold all the keys to the castle. And every situation is, well, situational. So the last thing I wanted was for the founder to take my advice as the word of god (nor anyone reading this blogpost now). Merely a tool in the toolkit. At times, useful. Other times, just something that acts as décor in the shed.

“Elijah, it’s probably too late for that exec… for now. He’s made his decision and walked. That said, I think there are two things to be aware of here:

  1. The fact you didn’t know about this until it happened, and
  2. The decision itself.”

Pre-empting the ultimatum

For the former, here’s how I think about pre-empting your team’s career inflections.

  1. In their first week, have everyone put together their personal manifesto. What is their 6 month goal? 1 year? 5 years? 10 years? Lifetime goal? What motivates them? How do they like to give and receive feedback? Of course, it’s helpful to share your own first, so they have a reference point. Don’t expect anything you’re not willing to share first. So, naturally, this requires a level of transparency, and more importantly, vulnerability.
  2. Then within the first two weeks, you and their direct manager should review their manifesto with them for at least 30 minutes live. Really get to know them. Taking a page out of Steven Rosenblatt’s book, what drives them? What haven’t they achieved that they want to achieve? How do they do their best work? When do they feel the most motivated? Why did they want to work here? Why are they excited to do so? How does working at your company fit in their broader goal?
  3. Then every quarter, allow every team member one day of mindfulness away from their work to revisit their manifesto. I usually recommend a Friday. What’s changed? What’s stayed the same? Does their current role still fit in their broader goal? If not, why not?
  4. The week after, take time to sync again and be incredibly candid.

Of course, the above is easier to do if you have a company of less than 50. At some point, when your company scales past that, it’s at least helpful to do it with your direct reports and their direct reports.

Helping with the decision

For the latter, you can’t stop a river. Even if you build a dam, the flow will always find a way around. You can’t change what motivates someone else. But you can help them channel it. The best thing you can do is equip that person with the tools to make a decision they will not regret, and wish them the best.

I like to sit people down and first help them figure out why they’re considering a new role. People often conflate the three traits of a job — compensation, scope, and title — together when making a career move. But in truth, they’re similar, but all a bit different. And I want people to know that just because they’re getting paid more doesn’t necessarily mean an increase in responsibility. Just because they’re getting a new title doesn’t mean that they’ll get more money. Then I have them stack rank the three traits. From most to least important.

If they still rank compensation first, that’s fine. Maybe they’re saving to buy a new house or to pay for their child’s higher education. And there’s nothing you nor I can do there. But if it’s one of the other two that come out on top, there’s room to create a new position or set of responsibilities where the individual feels empowered. And if it’s not at your company, they’ll be equipped to think through it at their next company. If they don’t have one lined up yet, help them through your network find one that’ll fit the criteria.

The wonderful irony

The funny thing about helping people achieve their dreams — sometimes that’s actively helping them leave your company — is that the karma usually comes back in one way or another. In this case, and I’ve seen it and experienced it before, even if you lose this person at this time and place, they’ll remember the help you gave them. To which, one day, when they have an all-star friend looking for their next opportunity, they will think of you.

There’s a saying I love. ‘The best compliment an investor can get is to get deal flow from someone they passed on.’ And here, the best compliment you can get is to get talent from someone who left your team.

In closing

Shake Shack’s Danny Meyer recently said something that echoes this notion. While he uses the word “volunteering,” he defines “volunteering” as:

“I basically, to this day, treat all of our employees as if they are volunteers, which not in the real sense. You’re going to get paid. But if you’re working for me, it means you’re probably good enough to have gotten another 25 job offers at least. And so, as far as I’m concerned, you’re volunteering to share your gifts with us.”

He goes on to say, “I didn’t have any way to motivate them with money. I couldn’t give them a raise, couldn’t dock them their pay. So I learned such a crucial lesson, which is that, if someone’s volunteering, the only way to motivate them is to have a higher purpose.”

Of course, there’s more than one way to make a team member feel like they are valued and that they value their work here. Another way is to give your prospective team member a “love bomb”, as Pulley’s Yin Wu calls it.

Now I’m not saying that if Elijah did all the above, he’s guaranteed to retain the exec. Who knows? He might have. Might not. For a man with a family and financial needs, it’s a hard ask. But at the minimum, this career move wouldn’t have blind-sided him. And better, he could’ve supported that exec in making that career move.

Just like with your product, your goal with your team is also to catch lightning in a bottle. How do you attract the best talent to work with you? And then, once you are able to, how do you keep them?

With the latter, a big part of it is showing you care.

Photo by Diego PH 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.