Category: Uncategorized

Great 1.5 hour interview on product development, product strategy, AI, organizational structure/autonomy in decision making, the challenge with redesigning a popular feature used for a long-time and much more with Gustav Söderström on Lenny’s Podcast.

Two interesting, but not new, points:

1. Spotify went from 7 person development ‘squads’ to about 20 person teams per manager. The employees per manager went up significantly.
2. When you redesign a service or popular page (like Spotify did with the homepage), you will get a lot negative feedback. One reason is that you change the environment for old users, which means they don’t find things as easily. But you might also have broken things. One way to understand what is going on is to compare new user cohorts with old user cohorts.

Just a great listen for founders, management and employees.

Michael Moritz on what made Sequoia Capital successful: "We've always been afraid of going out of business."

"We've worked hard on trying to figure out how we make Sequoia endure. We've assumed that tomorrow isn't like yesterday. We can't afford to rest on our laurels." pic.twitter.com/ItATD2zESl

— Frederik Gieschen (@NeckarValue) May 20, 2023

The Verge: Big Tech is already warning us about AI privacy problems.

Not surprising that large companies that care about confidentiality and intellectual property don’t want to use “public” AI services like OpenAI and GitHub Copilot.

The development with privately run AI models/services for enterprises make a lot of sense from this point of view.

Acquired has a long discussion with Spotify founder Daniel Ek about Spotify, music, podcasting, business models, creators and more.

Super interesting chart from our friends at BofA regarding historical software multiples at IPO – some very good companies priced in the 5-8x forward revenue range prior to '20 and seems that's where we're headed again pic.twitter.com/cdPUREBECG

— Glenn Solomon (@glennsolomon) May 17, 2023

The average valuation of a software company at IPO informs all valuations from pre-seed to Series F. The reason being that “the prize” for companies reaching growth goals and getting to “the next level” is the valuation at the IPO.

At 7.4x revenue 2022 valuations are still a higher than the long-term trend around 5x. Exactly where the average valuation should be is as much driven by the companies going public as “what’s the historical”. If today’s companies are growing faster and are more profitable, they should have a premium valuation compared to historic numbers.

OnlyCFO: The Magical “10x Employee”. Some good thoughts on hiring great people and avoiding the wrong hires. Still this is worth remembering: “A 10x-er at one company won’t necessarily be a 10x-er at another company. A lot of hiring managers get this wrong.”

Raising money is hard. Raising money when things are not going well is even harder.

That doesn’t mean that every term sheet with a low valuation or tough demands is ‘dirty’ (as reading Breakit might make you believe).

There are dirty things like pushing the founder out and very short deadlines that in my mind always are dirty. In 99% of cases I’d put participating preferred liquidation preference as dirty, as in any normal fundraise more than 1x non-participating liquidation preference isn’t necessary.

But sometimes the situation is so bad that accepting terms that are normally considered dirty are better than going bankrupt. Because most often the alternative to accepting a ‘dirty’ term sheet is running out of money, not getting a ‘clean’ term sheet at an acceptable valuation.

One thing in company presentations that makes me write down a bunch of questions is seeing extremely high profit margins in the forecasted financials.

Among all the great software and Internet companies I don’t think I’ve seen a company, with the exception of Evolution in 2021, that has grown more than 80 % year-over-year at scale and had profit margins over 50 %.

Obviously that doesn’t mean it is not at all possible, but in most cases it is likely that the company is underestimating the costs of growing or overestimating how quickly it will grow. It is a type of ‘error’ that is called base case fallacy, a.k.a. the plan is not fully taking into account how rare something is compared to the standard outcome.

I believe Kevin Kelly (of Wired, 1000 True Fans and much more) is right when he says that the best thing for once’s work ethic is to have a rest ethic.

Raising venture capital for a startup is fundamentally a trade where entrepreneurs get cash and (hopefully) better odds to succeed, and in exchange they give up ownership and some control of the company.

Valuation, preference shares, and the shareholders’ agreement with all its terms are ways to define that trade.