The Paragraph version of That Was The Week
Support That Was The Week
That Was The WeekAlgorithms and Venture Capital
Support That Was The Week
That Was The WeekAlgorithms and Venture Capital
The Paragraph version of That Was The Week
Subscribe to That Was The Week
Subscribe to That Was The Week
<100 subscribers
<100 subscribers
@googleventures, @GVteam abandons data science for investments? @bariweiss on DALL_E and Art, @peteflint and @NFX on real-estate 3.0, @bgurley on Startups in ’22, @SubstackInc launch, @Cloudflare launches and @gassee on Apple Car, and more
By Keith Teare • Issue #329
Algorithms and Venture Capital
Google Ventures Shelves its Algorithm
Inside Google’s Venture Capital Machine
Substack Reader for the Web
Real Estate 3.0 - The Ownership Revolution
DALL-E and Investors
There is no Such Thing as AI Art
Apple Car - Bad Idea After All
It Only Takes 5 Minutes to Know if a Startup Has a High Probability of Success
The Cheat Sheet for Venture Capital Metrics
Samir Kaji and Josh Berman on “Trapped Liquidity”
Elon Musk to be Deposed
CloudFlare’s $1.46bn Fund and VC Partnerships
Ark Launches a Venture Fund
Flatfile
Bill Gurley
How can you tell the difference between a good startup and one that is not so good? One of this week’s curated writers thinks he can figure it out in 5 minutes.
In just a few minutes, I can tell you if a start-up will make it or not.
All I need to do is to have an interview with its founder.
And ask him one question: “Why?”
And i keep asking him “why” again and again until I feel him.
Having done quite a few investments in my time and had all kinds of outcomes I can confirm that feeling. I can also confirm that it is a false signal.
The founders that end up winning can be very different from the ones you think will win.
By contrast, according to Axios, Google Ventures concluded this week that it is shuttering its data algorithms that were designed to green, amber, or red light investments.
Google Ventures has mothballed an algorithm that for years had served as a gatekeeper for new investments, Axios has learned from multiple sources.
Why it matters: This is a strategic sea change for one of venture capital’s most data-driven firms, and a Big Tech acknowledgement that human judgement shouldn’t always be automated away.
The ability of AI, machine learning, and data-driven predictive analytics is not questioned in most business segments. But in venture capital, where stock picking is king (rarely queen) it seems that the assumption that machine intelligence can beat humans is a step too far.
On the face of it that is a bit rich. If data can be used for its predictive qualities then venture capital should be able to benefit from it too.
I was drawn to these stories because of my work at SignalRank. We are an entirely data-driven recommendation machine. In our case, we recommend B rounds. The companies we recommend are what we call internally The SignalRank Index. This Index has the goal of having the B Rounds that go on to become unicorns, or better. To recommend a company at the B Round, with limited information, is hard. But we do.
Our index today has 863 companies in it, picked by our algorithms. These companies include almost 200 unicorns. The average multiple from the B Round for these companies is over 10x with the top decile at over 50x in a 10-year window. The chance we will recommend a future unicorn at the B round is well over 30% of the time.
Compared to the average B round (chosen by humans) our algorithms can do over 300% better at unicorn picking and deliver over 300% better multiples.
So you would think that I, of all people, agree with those who champion machines over humans for venture investing. but that would be wrong.
The SignalRank algorithms focus on learning from human decisions. It is our trade secret how we do that. But trust that in weighting key performance indicators, those gleaned from learning from the behavior of the best human investors play a highly significant role in whether a B Round is recommended or not.
What this means is that machine learning and human decisions are intimately joined in how we predict outcomes. Either one alone would fail.
We have backtested the algorithm to 2012 with pure non-cheating data (that is to say only using data that would have been known at the time of a recommendation and not benefitting from any future knowledge).
Looking at the period 2011-2019 our recommended B rounds look like this.
356 companies joined the SignalRank Index during that time. That is from 12,478 B rounds. The average multiple of invested capital produced by a B round is 3.35% and 9% become unicorns. Of the 356, 113 are unicorns (31.7%). The multiples created from the 356 are already at 14x the B investment.
Since 2019 a further 520 B round companies have joined the Index. 179 of them this year, 2022. It is a reasonable prediction that the performance of these companies will, compared to the B rounds we did not recommend, outperform their cohort handily.
So, I do believe machines can help make decisions. And if it is based on learning from the human behavior that venture capital investors exhibit, it can outperform humans alone. The SignalRank Index is a proprietary output of our algorithms and is available for our investors to see. We are currently tracking over 1200 A-round companies that may qualify as index entrants when they do their B Round.
The article about DALL_E and whether AI can produce art makes some similar points this week. A highly recommended read.
And Bill Gurley, in this week’s Tweet of the Week, screams out that this year may be the best time to start a company. I agree with him.
The Video and Podcast with @kteare and @ajkeen accompanying That Was The Week is recorded separately and delivered to paying subscribers via email on Friday or Saturday each week. To subscribe, go to our home at Substack. This week there will be no video as Andrew Keen is traveling and unavailable.
Read That Was The Week in the Substack app
Available for iOS and Android
Google Ventures shelves its algorithm
Google Ventures has mothballed an algorithm that for years had served as a gatekeeper for new investments, Axios has learned from multiple sources.
Why it matters: This is a strategic sea change for one of venture capital’s most data-driven firms, and a Big Tech acknowledgement that human judgement shouldn’t always be automated away.
Backstory: Axios first reported on GV’s algorithm in 2018, explaining how it had begun as a due diligence tool to aid a nascent team that had more experience in engineering than in investing.
It later evolved into a “stoplight system” that could effectively halt deals in their tracks.
GV investors sometimes tried to game the algorithm by manipulating the inputs. In general, however, the firm abided by the machine’s red lights (plus greens and yellows). As we wrote previously, it became GV’s de facto investment committee.
State of play: There doesn’t appear to be a single incident that killed off the algo.
Instead, it appears to have been a gradual process born of growing self-confidence (GV now has nearly 40 investors managing around $8 billion in AUM) and growing frustration (particularly when the algo would rule against follow-on investments for existing portfolio companies, as the deal market deteriorated).
The bottom line: GV still relies heavily on data. After all, this is the corporate venture arm of Google. But data has been relegated to its original role as aide, rather than arbiter.
Scoop: Inside Google's Venture Capital "Machine"
When most venture capitalists want approval to make a new investment, they go to their partners. When venture capitalists at GV do it, they go to something called “The Machine.”
What we’re hearing: Axios has learned that the firm, formerly known as Google Ventures, for years has used an algorithm that effectively permits or prohibits both new and follow-on investments.
Staffers plug in all sorts of deal details into “The Machine” — which is programmed with all sorts of market data, and returns traffic signal-like outputs. Green means go. Red means stop. Yellow means proceed with caution, but sources say it’s usually the practical equivalent of red.
It was initially designed and used as a due diligence assistant that could be overruled but, according to three sources, it has evolved into a de facto investment committee.
The backdrop: GV was formed in 2009 as one of the first venture firms to employ engineers whose primary job was to work with portfolio companies on technical challenges. But, in the early days, there weren’t too many portfolio companies yet, so the engineers were tasked first with building a dealflow management tool dubbed “Vortex,” and then with what would become “The Machine.”
Another impetus was that few of the early GV investors had much, if any, investing experience. So “The Machine” would leverage the firm’s strengths (engineering) as a bulwark against its weakness (proven VC chops).
The engineers were also asked to have “The Machine” help source deal opportunities, but that wasn’t viewed as a terribly successful effort.
The first hints of this came in 2013, when then-GV CEO Bill Maris told the NY Times: “We have access to the world’s largest data sets you can imagine, our cloud computing infrastructure is the biggest ever. It would be foolish to just go out and make gut investments.”
What Maris didn’t say in that piece, in part because it wasn’t quite so codified yet, was the color-coding system that virtually took “gut” out of it entirely.
Inputs into “The Machine” include round size, syndicate partners, past investors, industry sector and the delta between prior valuation and current valuation. The algorithm then ranks deals on a 10-point scale, with green said to represent 8 or above.
Announcing the all-new Substack Reader for web
There’s a new reading experience waiting for you at Substack.com. Now you can read all your Substack subscriptions—and more—in a clean, simple, and fast web reader. Everything stays in-sync with your Substack app for iOS.
Want to add a publication from outside Substack? No problem—just select “Add RSS feed” from the left sidebar.
And for old-school online readers who like to navigate by hot-key, we’ve got you covered. Just use J/K to hop between posts without touching your trackpad.
Other noteworthy features:
Press E to archive a post, S to save it for later, or L to leave a like.
Access your Profile, your Library, and Discover from the left sidebar.
Use Search at the top of the screen to find new Substacks.
See recommendations from the writers you subscribe to right beside your inbox. Posts that generate more subscriptions float to the top of the list, which we think is a reasonable indicator of quality.
Last but not least…dark mode!
Real Estate 3.0 – The Ownership Revolution
Real estate has always been more than just the largest asset class in the world. It is the embodiment of home and work, family and business – the opportunity of generational wealth, writ large. The American dream.
One that has become increasingly out of reach to most Americans.
But, obscured by news of 14-year-high rate hikes, outsized mortgages, rising rents, and company layoffs, there is something important and largely positive happening at the edges of the real estate industry:
Paths to ownership of real estate are expanding. What “ownership” even means is also expanding.
It’s easy to see more constriction than expansion, and for good reason. Affordability and access to traditional home ownership has only eroded since the housing bubble in 2008, as lending standards tightened and home prices have soared in recent years.
And yet, restriction breeds innovation and market changes create opportunity. From the front lines with proptech Founders, we’re seeing early signs of a real estate revolution.
The sky’s the limit for AI tools like DALL-E, but investors have a rough road ahead
It’s a famous startup saying that the next big thing will start out looking like a toy. And there’s no toy that VCs have been more excited about playing around with recently than DALL-E and other generative AI image tools.
Put a few key words into a tool like Midjourney, Stable Diffusion, or DALL-E and it’s easy to see why the whimsical (and often wacky) images have captured investors’ imagination. An AI-generated artwork even recently won an art competition at the Colorado State Fair, a result that didn’t go over well among more traditional artists. It’s become disruptive enough that this week Getty announced a ban of AI-generated images on its platform, following similar moves by some online art communities.
What looks like an interesting art tool has become a prime feeding ground for investors. Investor interest has been nearly overwhelming for Poly’s Abhay Agarwal, who is building a “DALL-E for design assets” company. “It has literally been like dropping yourself into the Ganga River and fully being bathed in it,” Agarwal said of the interest. He’s already had over 80 meetings with VCs and is only halfway done following YC’s Demo Day.
The challenge now for investors is finding the business case in AI-generated imagery. Already, some companies like Stitch Fix have been experimenting with the technology, but with mixed success. “I feel quite strongly that these technologies are quite world-changing,” Khosla Ventures partner Kanu Gulati told me. “They’re still early. A lot of their shortcomings are known, but the community is super, super active and trying to resolve them.”
Perhaps unsurprisingly, the initial startup applications have been around design, marketing and e-commerce, like a company doing AI-generated stock imagery or a startup building AI models for fashion brands so they can skip photoshoots. Gulati has invested in startups like Rosebud, which is doing AI-generated photos and videos (including NFTs), while Khosla Ventures has directly backed research lab OpenAI, the creator of DALL-E. Poly is pitching itself as a way for designers to use AI to generate textures.
Already looking ahead, Gulati thinks AI imagery will be used with other forms of generative AI-like text, and that’s where more value can be created. “There will be huge industries out there giving Adobe a run for their money because of using these latest technologies,” Gulati said. “And these will be built on a new stack of AI-first companies.”
A version of this story appeared in Protocol’s Pipeline newsletter. Sign up here to get it in your inbox every Saturday.
There Is No Such Thing as A.I. Art
I’ve always had problems envisioning the underworld. Sulfurous flames belching up from gloomy caverns don’t trigger existential terror in me. This may be because I grew up in Minnesota, where, for over half the year, fire is inviting, cozy, not forbidding.
But even detailed scenes of suffering in hell have always fallen short, for me, of their awful equivalents on Earth: Real war and real famine horrify me more than paintings of the damned devouring their own arms. Literary evocations of hell, which focus on its prisoners’ inner states—I’m thinking here of Virgil’s Aeneid and Dante’s Inferno—affect me more deeply, but once again the miseries they speak of are also available in life. The only distinctively hellish thing about these torments is that they are said to persist for all eternity. Eternity, which, perhaps you won’t be surprised to learn, I also have trouble imagining.
All of this changed for me the other day when I came across a brief animated video. It struck me, at last, with authentic spiritual dread.
The video was a creation of DALL-E, a new artificial intelligence app from the wizards at OpenAI, which is said to represent a breakthrough in the production of machine-made art. You type in a verbal description of an image—“a tarantula wearing a green scarf,” say—and out of the digital void arrives a picture which reflects your specifications. If you’d like, you can tinker with the image the way you might customize a frozen pizza: You can tell the A.I. to render the tarantula in the style of a cubist drawing or a vintage photograph or a Soviet propaganda poster. (How all this works at a computing level I’ll explain in a moment, or I’ll try.) But when I saw the 30-second video, all I knew was foreboding.
After arguing that Apple’s EV project could be a big win that’s well within the company’s reach, today we turn to the other side of the bet.
Who wouldn’t want to drive a vehicle built by a company who’s sense of fit and finish, its attention to the user’s experience is second to none? A vehicle we could facilely call The iPhone of EVs — although “Apple Car” is a powerful enough monicker.
That was my conclusion in the August 21 Monday Note, Apple Car: Software and Money. But there’s another side to the story. As the sages insist, we don’t understand a problem, an idea, a case unless we’re able to see, to plead both sides. So, I’ll attempt to argue that the Apple Car is a bad idea.
In that Monday Note, I asked the money question: Why would Apple, with its 54% Gross Margin (more for services, a little less for hardware), wade into an auto industry swap that has notoriously low Gross Margins, around 7% worldwide with a little more for premium brands?
However, when we took a closer look at Tesla’s financial statements, we found that the preeminent EV company’s Gross Margin has fluctuated between 28% and 33% for the past five quarters. Surely, an “even more organized” company such as Apple could do better and achieve its customary Gross Margin level in a $3T (as in trillion) industry. Furthermore, whereas Tesla has to build its own factories, Apple could operate in its usual Asset Light (and software-heavy) fashion. As the company does for all its products, from iPods to iPhones and Macs, subcontractors managed by Apple would build Apple Cars.
But while the Asset Light business model (“where the company focuses on reducing the amount of capital that is invested in assets”) allows financial flexibility, it’s not free — there’s no magic. Apple must provide financial support for the contractors who build their devices. If you have the time and inclination, take a look at Apple’s Q2 FY (Fiscal Year) 2022 quarterly statement. In the always instructive Management’s Discussion and Analysis of Financial Condition and Results of Operations section, there is a subsection titled Manufacturing Purchase Obligations that details advance payments made to its manufacturing contractors. In Q2 FY 2022, the amount was a respectable $40.6B — soberly labeled as “primarily non-cancelable”. Apple would surely do the same, at the appropriate scale, for a car manufacturing contractor.
Then there’s the price challenge. Today, EVs cost $40K and up, where “up” means $150K for a Porsche Taycan, or $110K for a Mercedes EQS. The entry-level Tesla Model 3 starts at $46K for the two-wheel drive model and more than $60K for the four-wheel drive version — to say nothing of the infamous “Full Self-Driving” package, an additional $15K. (I’ve often wondered if Elon Musk’s regrettable (and privately admitted) exaggerations will ever finally catch up with him.)
But is this a challenge or an opportunity? Would Apple try to undercut existing EV makers by selling an Apple Car for significantly less than $50K? No, history tells us that Apple would vie for a premium spot by trading on its reputation and top-grade UI. I’ve tried several EVs (besides our own Tesla) and have found their UI lacking. I even watched from the rear seat of a German EV as a salesperson “unsold” my spouse, confusing her with the many ways to accomplish a simple task.
It Only Takes 5-minutes To Know If a Startup Has a High Probability of Success Or Not
In just a few minutes, I can tell you if a start-up will make it or not.
All I need to do is to have an interview with its founder.
And ask him one question: “Why?”
And i keep asking him “why” again and again until I feel him.
I need to feel the person I have in front of me. I need to know if he or she has the right profile to carry out the project to its term. Entrepreneurship is not a walk in the park. You can have a great idea, but you will fail if you don’t have the skills.
We know the qualities of a good entrepreneur: He is a leader, determined, and able to get up at the slightest setback.
For me, they can be split in two categories.
The first entrepreneur wants to do something with his life and aims to give it meaning. The second has no other aspiration than to make as much money as possible. The second is generally a good talker. He only has in mind the desire to impress the gallery. What motivates him is to see himself one day maneuvering his boat with coconut trees and sandy beaches in the background.
What drives the first entrepreneur is immediately apparent in the way he talks, in the expression on his face, and the tone of his voice. He will not speak about his product. For him, it is (almost) secondary. His pitch is not a sales pitch. He does not sell for the sake of selling something. What he wants is to achieve his purpose. The product is only the fuel that will get him to his destination.
The Cheat Sheet for Venture Capital Metrics | Diligent Equity
Venture capital (VC) is a type of private equity financing many companies use to scale and grow. It requires investors to take educated risks and make calculations on companies with the potential for high growth.
Because of the risk involved, if you want to get deeper into the world of VC, it’s necessary to learn how to do important calculations to determine the state of your VC fund.
In this VC fund metrics cheat sheet, you’ll learn how to:
Determine the performance and health of your fund using nine crucial venture capital fund metrics.
Perform calculations for multiple and internal rates of return (IRR) calculation — and how to distinguish them from one another.
The Goal of Reporting Fund Performance
Before diving right into calculations, it’s essential to understand what you’re looking to learn. Intended for beginners and pros alike, this cheat sheet will remind you of the key metrics you should track to measure a fund’s performance. This will help you understand how well your fund and company are doing relative to others.
Specifically, the cheat sheet will help you remember and address the following:
When should you use multiple calculations versus IRR calculations? What are the differences?
How is your particular investment performing right now?
How has the company performed historically?
Which metrics are more vital to limited partners (LPs)? Which metrics are more important to general partners (GPs)?
How is your overall fund performing?
How is your fund doing compared to similar VC funds with the same vintage year?
How do these returns compare to market averages?
Remember, if you need a trusty cheat sheet you can download all the material in this blog post here!
Listen now | Episode 92
@samirkaji
This week we are joined by Josh Berman, Co-Founder and Managing Partner of private lending firm Quid, an active funding platform that provides liquidity to shareholders of top private companies. Quid has raised $420M across two funds. Josh has been in technology for over two decades, co-founding MySpace in 2003, after which he went on to start BeachMint before moving to the investing side and starting both Troy Capital Partners in 2016 and private sharing financing company Quid in 2018. During the show, we talked about the difference between secondary selling and borrowing, the issue of trapped liquidity at funds, and the learning he took away from his Myspace experience.
ventureunlocked.substack.com • Share
Elon Musk to face deposition by Twitter lawyers ahead of trial
Elon Musk is scheduled to spend the next few days with lawyers for Twitter, answering questions ahead of an October trial that will determine whether he must follow through on his $44bn agreement to acquire the social platform after attempting to back out of the deal.
The deposition, planned for Monday, Tuesday and a possible extension on Wednesday, will not be public. As of Sunday evening, it was not clear whether Musk would appear in person or by video. Reuters reported the deposition did not happen Monday nor was a reason given for the delay, citing sources with knowledge of the situation.
Twitter’s attorneys are expected to use the interview to try to show that Musk abandoned the deal due to falling financial markets and not because the company misled him about the real number of users or hid security flaws, as he alleged.
Musk, the world’s richest man, agreed in April to buy Twitter and take it private, offering $54.20 a share and vowing to loosen the company’s policing of content and to root out fake accounts.
Musk indicated in July that he wanted to back away from the deal, prompting Twitter to file a lawsuit to force him to carry through with the acquisition for $54.20 per share.
Musk wants a judge to allow him to walk away from the agreement without penalty.
Cloudflare unveils $1.25 bn fund to help startups, partners 26 VC firms - The Siasat Daily
Digital infrastructure services provider Cloudflare on Tuesday announced a $1.25 billion ‘Workers Launchpad Funding’ programme to help startups grow their businesses.
The company partnered 26 leading venture capital firms to help startups building applications on Cloudflare Workers, a highly-scalable serverless computing platform that allows developers to build or augment apps without configuring or maintaining infrastructure.
“While we can provide the technology, we’re thrilled to partner with some of the leading venture capital firms on the Workers Launchpad Funding Program, who will potentially invest more than a billion dollars in funding towards great startups built on Cloudflare Workers as they scale,” said Matthew Prince, Co-founder and CEO of Cloudflare.
The Siasat Daily
Cathie Wood's ARK Invest unveils new actively managed Venture Fund - Seeking Alpha
Cathie Wood and brokerage platform Titan partnered together to launch the actively managed ARK Venture Fund (MUTF:XARKX).
The new ARK Venture Fund aims to democratize venture capital by providing all market participants access to innovative organizations throughout the private and public markets with a minimum investment of $500.
Additionally, unlike traditional venture capital funds which lock capital up for years, ARK’s evergreen public-private crossover fund offers clients partial liquidity through quarterly redemptions. Between the private and public sector, the fund intends to invest 70% in private companies and 30% in public firms.
XARKX will traditionally hold between 25-50 holdings, and is attached with a total 4.22% expense ratio that includes a 2.75% management fee, a 0.65% distribution/services fee, and 0.82% in other expenses. Moreover, the fund is also listed as a closed end interval fund.
With regards to the launch Wood stated: “By launching the ARK Venture Fund, we seek to augment venture capital, offering all investors access to what we believe are the most innovative companies throughout their private and public market life cycles.” Seeking Alpha
Flatfile Closes $50M Tiger-Led Series B
Denver-based Flatfile, a company that manages partner data exchange, has raised a $50 million Series B led by Tiger Global. This funding closed 18 months after its $35 million Series A led by Scale Venture Partners.
Other investors in this round include Gradient Ventures, Scale Ventures and Workday Ventures.
We spoke with co-founder and CEO David Boskovic about the fast-growing company that has raised close to $100 million in three years.
Flatfile has “expanded across enterprises, Fortune 500s, and our objective is to be part of every data exchange in the world,” said Boskovic.
This may be a great time to launch a start-up, according to venture capitalist Bill Gurley @bgurley. Here's why ➡️ mck.co/3xTcVuC @benchmark
That Was The Week is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.
@googleventures, @GVteam abandons data science for investments? @bariweiss on DALL_E and Art, @peteflint and @NFX on real-estate 3.0, @bgurley on Startups in ’22, @SubstackInc launch, @Cloudflare launches and @gassee on Apple Car, and more
By Keith Teare • Issue #329
Algorithms and Venture Capital
Google Ventures Shelves its Algorithm
Inside Google’s Venture Capital Machine
Substack Reader for the Web
Real Estate 3.0 - The Ownership Revolution
DALL-E and Investors
There is no Such Thing as AI Art
Apple Car - Bad Idea After All
It Only Takes 5 Minutes to Know if a Startup Has a High Probability of Success
The Cheat Sheet for Venture Capital Metrics
Samir Kaji and Josh Berman on “Trapped Liquidity”
Elon Musk to be Deposed
CloudFlare’s $1.46bn Fund and VC Partnerships
Ark Launches a Venture Fund
Flatfile
Bill Gurley
How can you tell the difference between a good startup and one that is not so good? One of this week’s curated writers thinks he can figure it out in 5 minutes.
In just a few minutes, I can tell you if a start-up will make it or not.
All I need to do is to have an interview with its founder.
And ask him one question: “Why?”
And i keep asking him “why” again and again until I feel him.
Having done quite a few investments in my time and had all kinds of outcomes I can confirm that feeling. I can also confirm that it is a false signal.
The founders that end up winning can be very different from the ones you think will win.
By contrast, according to Axios, Google Ventures concluded this week that it is shuttering its data algorithms that were designed to green, amber, or red light investments.
Google Ventures has mothballed an algorithm that for years had served as a gatekeeper for new investments, Axios has learned from multiple sources.
Why it matters: This is a strategic sea change for one of venture capital’s most data-driven firms, and a Big Tech acknowledgement that human judgement shouldn’t always be automated away.
The ability of AI, machine learning, and data-driven predictive analytics is not questioned in most business segments. But in venture capital, where stock picking is king (rarely queen) it seems that the assumption that machine intelligence can beat humans is a step too far.
On the face of it that is a bit rich. If data can be used for its predictive qualities then venture capital should be able to benefit from it too.
I was drawn to these stories because of my work at SignalRank. We are an entirely data-driven recommendation machine. In our case, we recommend B rounds. The companies we recommend are what we call internally The SignalRank Index. This Index has the goal of having the B Rounds that go on to become unicorns, or better. To recommend a company at the B Round, with limited information, is hard. But we do.
Our index today has 863 companies in it, picked by our algorithms. These companies include almost 200 unicorns. The average multiple from the B Round for these companies is over 10x with the top decile at over 50x in a 10-year window. The chance we will recommend a future unicorn at the B round is well over 30% of the time.
Compared to the average B round (chosen by humans) our algorithms can do over 300% better at unicorn picking and deliver over 300% better multiples.
So you would think that I, of all people, agree with those who champion machines over humans for venture investing. but that would be wrong.
The SignalRank algorithms focus on learning from human decisions. It is our trade secret how we do that. But trust that in weighting key performance indicators, those gleaned from learning from the behavior of the best human investors play a highly significant role in whether a B Round is recommended or not.
What this means is that machine learning and human decisions are intimately joined in how we predict outcomes. Either one alone would fail.
We have backtested the algorithm to 2012 with pure non-cheating data (that is to say only using data that would have been known at the time of a recommendation and not benefitting from any future knowledge).
Looking at the period 2011-2019 our recommended B rounds look like this.
356 companies joined the SignalRank Index during that time. That is from 12,478 B rounds. The average multiple of invested capital produced by a B round is 3.35% and 9% become unicorns. Of the 356, 113 are unicorns (31.7%). The multiples created from the 356 are already at 14x the B investment.
Since 2019 a further 520 B round companies have joined the Index. 179 of them this year, 2022. It is a reasonable prediction that the performance of these companies will, compared to the B rounds we did not recommend, outperform their cohort handily.
So, I do believe machines can help make decisions. And if it is based on learning from the human behavior that venture capital investors exhibit, it can outperform humans alone. The SignalRank Index is a proprietary output of our algorithms and is available for our investors to see. We are currently tracking over 1200 A-round companies that may qualify as index entrants when they do their B Round.
The article about DALL_E and whether AI can produce art makes some similar points this week. A highly recommended read.
And Bill Gurley, in this week’s Tweet of the Week, screams out that this year may be the best time to start a company. I agree with him.
The Video and Podcast with @kteare and @ajkeen accompanying That Was The Week is recorded separately and delivered to paying subscribers via email on Friday or Saturday each week. To subscribe, go to our home at Substack. This week there will be no video as Andrew Keen is traveling and unavailable.
Read That Was The Week in the Substack app
Available for iOS and Android
Google Ventures shelves its algorithm
Google Ventures has mothballed an algorithm that for years had served as a gatekeeper for new investments, Axios has learned from multiple sources.
Why it matters: This is a strategic sea change for one of venture capital’s most data-driven firms, and a Big Tech acknowledgement that human judgement shouldn’t always be automated away.
Backstory: Axios first reported on GV’s algorithm in 2018, explaining how it had begun as a due diligence tool to aid a nascent team that had more experience in engineering than in investing.
It later evolved into a “stoplight system” that could effectively halt deals in their tracks.
GV investors sometimes tried to game the algorithm by manipulating the inputs. In general, however, the firm abided by the machine’s red lights (plus greens and yellows). As we wrote previously, it became GV’s de facto investment committee.
State of play: There doesn’t appear to be a single incident that killed off the algo.
Instead, it appears to have been a gradual process born of growing self-confidence (GV now has nearly 40 investors managing around $8 billion in AUM) and growing frustration (particularly when the algo would rule against follow-on investments for existing portfolio companies, as the deal market deteriorated).
The bottom line: GV still relies heavily on data. After all, this is the corporate venture arm of Google. But data has been relegated to its original role as aide, rather than arbiter.
Scoop: Inside Google's Venture Capital "Machine"
When most venture capitalists want approval to make a new investment, they go to their partners. When venture capitalists at GV do it, they go to something called “The Machine.”
What we’re hearing: Axios has learned that the firm, formerly known as Google Ventures, for years has used an algorithm that effectively permits or prohibits both new and follow-on investments.
Staffers plug in all sorts of deal details into “The Machine” — which is programmed with all sorts of market data, and returns traffic signal-like outputs. Green means go. Red means stop. Yellow means proceed with caution, but sources say it’s usually the practical equivalent of red.
It was initially designed and used as a due diligence assistant that could be overruled but, according to three sources, it has evolved into a de facto investment committee.
The backdrop: GV was formed in 2009 as one of the first venture firms to employ engineers whose primary job was to work with portfolio companies on technical challenges. But, in the early days, there weren’t too many portfolio companies yet, so the engineers were tasked first with building a dealflow management tool dubbed “Vortex,” and then with what would become “The Machine.”
Another impetus was that few of the early GV investors had much, if any, investing experience. So “The Machine” would leverage the firm’s strengths (engineering) as a bulwark against its weakness (proven VC chops).
The engineers were also asked to have “The Machine” help source deal opportunities, but that wasn’t viewed as a terribly successful effort.
The first hints of this came in 2013, when then-GV CEO Bill Maris told the NY Times: “We have access to the world’s largest data sets you can imagine, our cloud computing infrastructure is the biggest ever. It would be foolish to just go out and make gut investments.”
What Maris didn’t say in that piece, in part because it wasn’t quite so codified yet, was the color-coding system that virtually took “gut” out of it entirely.
Inputs into “The Machine” include round size, syndicate partners, past investors, industry sector and the delta between prior valuation and current valuation. The algorithm then ranks deals on a 10-point scale, with green said to represent 8 or above.
Announcing the all-new Substack Reader for web
There’s a new reading experience waiting for you at Substack.com. Now you can read all your Substack subscriptions—and more—in a clean, simple, and fast web reader. Everything stays in-sync with your Substack app for iOS.
Want to add a publication from outside Substack? No problem—just select “Add RSS feed” from the left sidebar.
And for old-school online readers who like to navigate by hot-key, we’ve got you covered. Just use J/K to hop between posts without touching your trackpad.
Other noteworthy features:
Press E to archive a post, S to save it for later, or L to leave a like.
Access your Profile, your Library, and Discover from the left sidebar.
Use Search at the top of the screen to find new Substacks.
See recommendations from the writers you subscribe to right beside your inbox. Posts that generate more subscriptions float to the top of the list, which we think is a reasonable indicator of quality.
Last but not least…dark mode!
Real Estate 3.0 – The Ownership Revolution
Real estate has always been more than just the largest asset class in the world. It is the embodiment of home and work, family and business – the opportunity of generational wealth, writ large. The American dream.
One that has become increasingly out of reach to most Americans.
But, obscured by news of 14-year-high rate hikes, outsized mortgages, rising rents, and company layoffs, there is something important and largely positive happening at the edges of the real estate industry:
Paths to ownership of real estate are expanding. What “ownership” even means is also expanding.
It’s easy to see more constriction than expansion, and for good reason. Affordability and access to traditional home ownership has only eroded since the housing bubble in 2008, as lending standards tightened and home prices have soared in recent years.
And yet, restriction breeds innovation and market changes create opportunity. From the front lines with proptech Founders, we’re seeing early signs of a real estate revolution.
The sky’s the limit for AI tools like DALL-E, but investors have a rough road ahead
It’s a famous startup saying that the next big thing will start out looking like a toy. And there’s no toy that VCs have been more excited about playing around with recently than DALL-E and other generative AI image tools.
Put a few key words into a tool like Midjourney, Stable Diffusion, or DALL-E and it’s easy to see why the whimsical (and often wacky) images have captured investors’ imagination. An AI-generated artwork even recently won an art competition at the Colorado State Fair, a result that didn’t go over well among more traditional artists. It’s become disruptive enough that this week Getty announced a ban of AI-generated images on its platform, following similar moves by some online art communities.
What looks like an interesting art tool has become a prime feeding ground for investors. Investor interest has been nearly overwhelming for Poly’s Abhay Agarwal, who is building a “DALL-E for design assets” company. “It has literally been like dropping yourself into the Ganga River and fully being bathed in it,” Agarwal said of the interest. He’s already had over 80 meetings with VCs and is only halfway done following YC’s Demo Day.
The challenge now for investors is finding the business case in AI-generated imagery. Already, some companies like Stitch Fix have been experimenting with the technology, but with mixed success. “I feel quite strongly that these technologies are quite world-changing,” Khosla Ventures partner Kanu Gulati told me. “They’re still early. A lot of their shortcomings are known, but the community is super, super active and trying to resolve them.”
Perhaps unsurprisingly, the initial startup applications have been around design, marketing and e-commerce, like a company doing AI-generated stock imagery or a startup building AI models for fashion brands so they can skip photoshoots. Gulati has invested in startups like Rosebud, which is doing AI-generated photos and videos (including NFTs), while Khosla Ventures has directly backed research lab OpenAI, the creator of DALL-E. Poly is pitching itself as a way for designers to use AI to generate textures.
Already looking ahead, Gulati thinks AI imagery will be used with other forms of generative AI-like text, and that’s where more value can be created. “There will be huge industries out there giving Adobe a run for their money because of using these latest technologies,” Gulati said. “And these will be built on a new stack of AI-first companies.”
A version of this story appeared in Protocol’s Pipeline newsletter. Sign up here to get it in your inbox every Saturday.
There Is No Such Thing as A.I. Art
I’ve always had problems envisioning the underworld. Sulfurous flames belching up from gloomy caverns don’t trigger existential terror in me. This may be because I grew up in Minnesota, where, for over half the year, fire is inviting, cozy, not forbidding.
But even detailed scenes of suffering in hell have always fallen short, for me, of their awful equivalents on Earth: Real war and real famine horrify me more than paintings of the damned devouring their own arms. Literary evocations of hell, which focus on its prisoners’ inner states—I’m thinking here of Virgil’s Aeneid and Dante’s Inferno—affect me more deeply, but once again the miseries they speak of are also available in life. The only distinctively hellish thing about these torments is that they are said to persist for all eternity. Eternity, which, perhaps you won’t be surprised to learn, I also have trouble imagining.
All of this changed for me the other day when I came across a brief animated video. It struck me, at last, with authentic spiritual dread.
The video was a creation of DALL-E, a new artificial intelligence app from the wizards at OpenAI, which is said to represent a breakthrough in the production of machine-made art. You type in a verbal description of an image—“a tarantula wearing a green scarf,” say—and out of the digital void arrives a picture which reflects your specifications. If you’d like, you can tinker with the image the way you might customize a frozen pizza: You can tell the A.I. to render the tarantula in the style of a cubist drawing or a vintage photograph or a Soviet propaganda poster. (How all this works at a computing level I’ll explain in a moment, or I’ll try.) But when I saw the 30-second video, all I knew was foreboding.
After arguing that Apple’s EV project could be a big win that’s well within the company’s reach, today we turn to the other side of the bet.
Who wouldn’t want to drive a vehicle built by a company who’s sense of fit and finish, its attention to the user’s experience is second to none? A vehicle we could facilely call The iPhone of EVs — although “Apple Car” is a powerful enough monicker.
That was my conclusion in the August 21 Monday Note, Apple Car: Software and Money. But there’s another side to the story. As the sages insist, we don’t understand a problem, an idea, a case unless we’re able to see, to plead both sides. So, I’ll attempt to argue that the Apple Car is a bad idea.
In that Monday Note, I asked the money question: Why would Apple, with its 54% Gross Margin (more for services, a little less for hardware), wade into an auto industry swap that has notoriously low Gross Margins, around 7% worldwide with a little more for premium brands?
However, when we took a closer look at Tesla’s financial statements, we found that the preeminent EV company’s Gross Margin has fluctuated between 28% and 33% for the past five quarters. Surely, an “even more organized” company such as Apple could do better and achieve its customary Gross Margin level in a $3T (as in trillion) industry. Furthermore, whereas Tesla has to build its own factories, Apple could operate in its usual Asset Light (and software-heavy) fashion. As the company does for all its products, from iPods to iPhones and Macs, subcontractors managed by Apple would build Apple Cars.
But while the Asset Light business model (“where the company focuses on reducing the amount of capital that is invested in assets”) allows financial flexibility, it’s not free — there’s no magic. Apple must provide financial support for the contractors who build their devices. If you have the time and inclination, take a look at Apple’s Q2 FY (Fiscal Year) 2022 quarterly statement. In the always instructive Management’s Discussion and Analysis of Financial Condition and Results of Operations section, there is a subsection titled Manufacturing Purchase Obligations that details advance payments made to its manufacturing contractors. In Q2 FY 2022, the amount was a respectable $40.6B — soberly labeled as “primarily non-cancelable”. Apple would surely do the same, at the appropriate scale, for a car manufacturing contractor.
Then there’s the price challenge. Today, EVs cost $40K and up, where “up” means $150K for a Porsche Taycan, or $110K for a Mercedes EQS. The entry-level Tesla Model 3 starts at $46K for the two-wheel drive model and more than $60K for the four-wheel drive version — to say nothing of the infamous “Full Self-Driving” package, an additional $15K. (I’ve often wondered if Elon Musk’s regrettable (and privately admitted) exaggerations will ever finally catch up with him.)
But is this a challenge or an opportunity? Would Apple try to undercut existing EV makers by selling an Apple Car for significantly less than $50K? No, history tells us that Apple would vie for a premium spot by trading on its reputation and top-grade UI. I’ve tried several EVs (besides our own Tesla) and have found their UI lacking. I even watched from the rear seat of a German EV as a salesperson “unsold” my spouse, confusing her with the many ways to accomplish a simple task.
It Only Takes 5-minutes To Know If a Startup Has a High Probability of Success Or Not
In just a few minutes, I can tell you if a start-up will make it or not.
All I need to do is to have an interview with its founder.
And ask him one question: “Why?”
And i keep asking him “why” again and again until I feel him.
I need to feel the person I have in front of me. I need to know if he or she has the right profile to carry out the project to its term. Entrepreneurship is not a walk in the park. You can have a great idea, but you will fail if you don’t have the skills.
We know the qualities of a good entrepreneur: He is a leader, determined, and able to get up at the slightest setback.
For me, they can be split in two categories.
The first entrepreneur wants to do something with his life and aims to give it meaning. The second has no other aspiration than to make as much money as possible. The second is generally a good talker. He only has in mind the desire to impress the gallery. What motivates him is to see himself one day maneuvering his boat with coconut trees and sandy beaches in the background.
What drives the first entrepreneur is immediately apparent in the way he talks, in the expression on his face, and the tone of his voice. He will not speak about his product. For him, it is (almost) secondary. His pitch is not a sales pitch. He does not sell for the sake of selling something. What he wants is to achieve his purpose. The product is only the fuel that will get him to his destination.
The Cheat Sheet for Venture Capital Metrics | Diligent Equity
Venture capital (VC) is a type of private equity financing many companies use to scale and grow. It requires investors to take educated risks and make calculations on companies with the potential for high growth.
Because of the risk involved, if you want to get deeper into the world of VC, it’s necessary to learn how to do important calculations to determine the state of your VC fund.
In this VC fund metrics cheat sheet, you’ll learn how to:
Determine the performance and health of your fund using nine crucial venture capital fund metrics.
Perform calculations for multiple and internal rates of return (IRR) calculation — and how to distinguish them from one another.
The Goal of Reporting Fund Performance
Before diving right into calculations, it’s essential to understand what you’re looking to learn. Intended for beginners and pros alike, this cheat sheet will remind you of the key metrics you should track to measure a fund’s performance. This will help you understand how well your fund and company are doing relative to others.
Specifically, the cheat sheet will help you remember and address the following:
When should you use multiple calculations versus IRR calculations? What are the differences?
How is your particular investment performing right now?
How has the company performed historically?
Which metrics are more vital to limited partners (LPs)? Which metrics are more important to general partners (GPs)?
How is your overall fund performing?
How is your fund doing compared to similar VC funds with the same vintage year?
How do these returns compare to market averages?
Remember, if you need a trusty cheat sheet you can download all the material in this blog post here!
Listen now | Episode 92
@samirkaji
This week we are joined by Josh Berman, Co-Founder and Managing Partner of private lending firm Quid, an active funding platform that provides liquidity to shareholders of top private companies. Quid has raised $420M across two funds. Josh has been in technology for over two decades, co-founding MySpace in 2003, after which he went on to start BeachMint before moving to the investing side and starting both Troy Capital Partners in 2016 and private sharing financing company Quid in 2018. During the show, we talked about the difference between secondary selling and borrowing, the issue of trapped liquidity at funds, and the learning he took away from his Myspace experience.
ventureunlocked.substack.com • Share
Elon Musk to face deposition by Twitter lawyers ahead of trial
Elon Musk is scheduled to spend the next few days with lawyers for Twitter, answering questions ahead of an October trial that will determine whether he must follow through on his $44bn agreement to acquire the social platform after attempting to back out of the deal.
The deposition, planned for Monday, Tuesday and a possible extension on Wednesday, will not be public. As of Sunday evening, it was not clear whether Musk would appear in person or by video. Reuters reported the deposition did not happen Monday nor was a reason given for the delay, citing sources with knowledge of the situation.
Twitter’s attorneys are expected to use the interview to try to show that Musk abandoned the deal due to falling financial markets and not because the company misled him about the real number of users or hid security flaws, as he alleged.
Musk, the world’s richest man, agreed in April to buy Twitter and take it private, offering $54.20 a share and vowing to loosen the company’s policing of content and to root out fake accounts.
Musk indicated in July that he wanted to back away from the deal, prompting Twitter to file a lawsuit to force him to carry through with the acquisition for $54.20 per share.
Musk wants a judge to allow him to walk away from the agreement without penalty.
Cloudflare unveils $1.25 bn fund to help startups, partners 26 VC firms - The Siasat Daily
Digital infrastructure services provider Cloudflare on Tuesday announced a $1.25 billion ‘Workers Launchpad Funding’ programme to help startups grow their businesses.
The company partnered 26 leading venture capital firms to help startups building applications on Cloudflare Workers, a highly-scalable serverless computing platform that allows developers to build or augment apps without configuring or maintaining infrastructure.
“While we can provide the technology, we’re thrilled to partner with some of the leading venture capital firms on the Workers Launchpad Funding Program, who will potentially invest more than a billion dollars in funding towards great startups built on Cloudflare Workers as they scale,” said Matthew Prince, Co-founder and CEO of Cloudflare.
The Siasat Daily
Cathie Wood's ARK Invest unveils new actively managed Venture Fund - Seeking Alpha
Cathie Wood and brokerage platform Titan partnered together to launch the actively managed ARK Venture Fund (MUTF:XARKX).
The new ARK Venture Fund aims to democratize venture capital by providing all market participants access to innovative organizations throughout the private and public markets with a minimum investment of $500.
Additionally, unlike traditional venture capital funds which lock capital up for years, ARK’s evergreen public-private crossover fund offers clients partial liquidity through quarterly redemptions. Between the private and public sector, the fund intends to invest 70% in private companies and 30% in public firms.
XARKX will traditionally hold between 25-50 holdings, and is attached with a total 4.22% expense ratio that includes a 2.75% management fee, a 0.65% distribution/services fee, and 0.82% in other expenses. Moreover, the fund is also listed as a closed end interval fund.
With regards to the launch Wood stated: “By launching the ARK Venture Fund, we seek to augment venture capital, offering all investors access to what we believe are the most innovative companies throughout their private and public market life cycles.” Seeking Alpha
Flatfile Closes $50M Tiger-Led Series B
Denver-based Flatfile, a company that manages partner data exchange, has raised a $50 million Series B led by Tiger Global. This funding closed 18 months after its $35 million Series A led by Scale Venture Partners.
Other investors in this round include Gradient Ventures, Scale Ventures and Workday Ventures.
We spoke with co-founder and CEO David Boskovic about the fast-growing company that has raised close to $100 million in three years.
Flatfile has “expanded across enterprises, Fortune 500s, and our objective is to be part of every data exchange in the world,” said Boskovic.
This may be a great time to launch a start-up, according to venture capitalist Bill Gurley @bgurley. Here's why ➡️ mck.co/3xTcVuC @benchmark
That Was The Week is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.
Share Dialog
Share Dialog
No activity yet