Board

In the future, 80% of venture-funded software companies will be open core.

Open core is the medium between open source and closed-source proprietary software. It typically has more resources than open source software thanks to monetization strategies and it is more transparent than proprietary, closed-source alternatives. Security, longevity, modifiability, and R&D velocity are the main reasons the trend toward open core continues to rise.

With the open core model, there are four types of contributors: community, customer, user (non-paying), and even competitors. Anyone in the world can work on the code. The company can seed issues for the community to work on, and contributors can submit ideas. It’s the network effect at its best.

Risk is uncertainty about the future. High technical risk means not knowing if a technology will work. High market risk means not knowing if there will be a market for your product. These are the primary risks that the VC industry as a whole contemplates. (There are other risks extrinsic to individual companies, like regulatory risk, but these are less frequent.)

Each type of risk has a different effect on VC returns. Technical risk is horrible for returns, so VCs do not take technical risk. There are a handful of examples of high technical risk companies that had great returns—Genentech,43 for example—but they are few.44 Today, VCs wait until there is a working prototype before they fund, but successful VCs have always waited until the technical risk was mitigated. Apple Computer, for example, did not have technical risk: the technology worked before the company was funded.

Market risk, on the other hand, is directly correlated to VC returns. When Apple was funded no one had any way of knowing how many people would buy a personal computer; the ultimate size of the market was analytically unknowable. DEC, Intel, Google, etc. all went into markets that they helped create. High market risk is associated with the best VC investments of all time. In the late ’70s/early ’80s and again in the mid to late ’90s VCs were comfortable funding companies with mind-boggling market risk, and they got amazing returns in exchange. In the mid to late ’80s they were scared and funded companies with low market risk instead, and returns were horrible.

First entrants into a market can have a number of advantages if their product has a network effect or other lock-in mechanism.

Alternatively, if your market is undergoing consolidation there are times when being the last independent company standing can be to your benefit.  Acquirers often screw up mergers or destroy products.

More recently, the two big trends transforming the enterprise are (i) Nocode/Lowcode/RPA, and the (ii) Collaborative Enterprise[1]. Collaborative enterprise is an updated take on building collaboration and learnings from consumer products into SaaS.

Flight Simulator is a living, breathing thing that has never been possible before."

Microsoft created its own gaming operation more than 20 years ago in part because the company recognized that gaming presented an opportunity to bring together various esoteric technologies in a way that everyday people could understand.

Even now, most normal folks have no real idea what "cloud computing" or "machine learning" or "photogrammetry" mean. As they soar over the Kalahari, Hong Kong, Paris, Mumbai or Brooklyn (as they actually sit at home in quarantine), they won't have to. What they will understand is that Microsoft made it possible. The appeal of flying over Earth is so fundamentally human (I'm looking at you, Icarus) and extends so far beyond the stereotypical "gamer" that Flight also presents a powerful marketing opportunity for Microsoft.

There's just no way you're going to produce the best of everything without a little bit of everyone.

These trends have resulted in an overhang of companies that either (1) lived without product market fit and survived well past their natural expiration point, or (2) hired way ahead of progress and burned large sums with high valuations and now are stuck with little progress per dollar and a large preference stack.

Many companies are likely about to meet a hard reckoning. This is likely to start end of 2023 and accelerate through end of 2024 or so. The likely timing is +/- 6 months. It is based on when companies last fundraised at scale, 2021, and how much runway they raised.

The macro economy and the startup valuation reset are now decoupling. The reset in private tech will happen roughly no matter what happens in the macro economy.

Think through your investors individual incentives as you consider what to do and hear (and filter) their opinion. A given company in a VC portfolio is part of a portfolio and may function as an option - while for you the company or how you otherwise spend your time is your entire livelihood. And never forget how precious your time is. Life is short, and productive years in which you can take a lot of risk are quite limited[2].

Investor incentives differ early vs late (and based on their role in VC firm)

This suggests if you plan to sell your company, you should pursue an exit sooner than later. End of 2023/early 2024 will become a crowded market for companies trying to sell.

It is possible we will see a rise in M&A solely for a company’s cash, which is a thing in biotech, and may also become a (smaller) thing in tech. In this scenario a cash rich company (whose product is not working) gets bought by a cash poor company whose product is working and who just wants the cash. This sort of M&A is effectively raising a round from another company’s founders and backers.

A number of companies such as Stripe, Instacart (rumored to have repriced 74% from $39B to $10B), and Klarna (valuation drop of 85% from $45B to $6.7B) have proactively written down their companies’ market caps in order to reset prices for their employees and/or fundraises (either private rounds or IPOs). These companies should be applauded for trying to do the wholistic right thing for their shareholders given that comparable public market companies may be down 70-90%.

A founders big idea and great team, at some point, will largely be converted into numbers around revenue, burn, growth, margin and related by the public markets weighing machine.

Some companies may be less impacted by a recession, such as security products or ones that clearly decrease costs.

However, the biggest lever for many companies to cut costs is not software - it might be to lay people off or shut down marginal products. This means even software that increase efficiency gains may see a strong slow down in the short term as their enterprise customers focus on more pressing matters in the short run. Gains may still be had for many enterprises through reasonably brute force actions.

The cloud wars ended up with AWS, Azure, and GCP as three large scale, fierce competitors. This is an oligopoly market with no single winner. Based on what we currently know about the world, this seems the most likely near term market structure for foundation language models, but it is early days and the future is uncertain.

This may take some time and is in the extreme a race of asymptote versus AGI, as well as further technological breakthroughs that could accelerate progress indefinitely. Moore’s “law”, which was more an observation than an actual law, lasted longer than anyone originally expected.

One potential future world would be OpenAI/MSFT versus Anthropic/Google versus Stability/Amazon vs Cohere/Meta (these are all made up pairings!!!!!). In other words, each incumbent would chose a startup partner to take on the brand and safety risk while buying major ownership in said startup. In exchange, the startup would get access to data, distribution, and other resources from the incumbent. To some extent, this results in roughly the same market structure as (2).

Crypto (BTC and ETH) are the obvious OS counterexamples to this, although one could argue given monetization being built directly into the crypto protocol equated in these groups self-funding. In some sense, crypto is its own corporate sponsor.

These reviews can take many months to complete, so it is often easier for enterprises to buy a less-good bundled product from an existing vendor, then a better stand alone product from a new supplier.

Creator advantage. Sometimes the teams that implemented or created an open source software project are best positioned to commercialize said product, given their ability to drive and control contributions to said project + brand and relationships in that community. See e.g. dbt Labs.

New business models. Sometimes a startup can innovate on business model to create a higher leverage business or different incentive structure. For example, Anduril in defensetech has a traditional tech margin-based business, while all the incumbents sell “cost plus” (ie they charge 5-8% on top of what is cost them to build the product for the Department of Defense). The cost plus model creates a lot of incentives to act badly - for example since labor is charged as part of the cost plus, delays and doing things slowly means more revenue for the incumbents. Similarly, the reason some items have $100 screws is so they can charge 5% on top of it (instead of just using a 10 cent screw).

1. It is sometimes hard to know what is actually working, versus hype.

2. Founders have a lot of pride in what they build, and may not want to just copy and out-execute someone. Often when a startup copies another’s idea, they put a unique spin on that approach or product versus default blankly copying it. All these tweaks and changes tend to make the product worse.

Most good startup ideas are definitionally non-obvious (otherwise everyone would be doing them). Often, good startup ideas seem small, niche, or toy like. Only desperate people will go and build in these areas as they may seem too small a market or use case initially to large incumbents.

Finally, there is user centric focus. User-centric companies tend to have a better understanding of their customer and their ongoing needs leading to superior product, sales, customer success, and pricing approaches. By aligning against the customer, more things tend to go right. While being customer-centric helps enormously in most cases, occasionally an incumbent is just too dominant for a superior customer experience and product to win.

The takeaway is that serving a customer need well is often more important (and harder) to think about than defensibility. In many cases defensibility emerges over time - particularly if you build out a proprietary data set or become an ingrained workflow, or create defensibility via sales or other moats.

The less building and expansion of the product you do after launch, the more vulnerable you will be to other startups or incumbents eventually coming after and commoditizing you. Pace of execution and ongoing shipping post v1 matters a lot to building one forms of defensibility above. Obviously, if your company is defensible up-front it is better then if it isn’t.

Complexity is conserved, meaning complexity cannot be removed it can only be moved around. For example, you can pay a provider to handle a class of complexity for you or bear those yourself - someone is doing it.

The interesting areas are where the complexity-cost tradeoff gets bent by innovation or commoditization. For example, nearly free static website hosting removed the need to run an apache or nginx webserver for these usecases.