When to Start a Startup

Hey y’all — here’s today at a glance:

🔗 Houck’s Picks

My favorite finds of the week.

Fundraising

• An active early-stage fund looking to invest in startups (Link)

Growth

• Product launch checklist (Link)

• Mastering TikTok to get millions in ARR (Link)

• How to promote your startup with $0 (Link)

ICYMI

• Compare your DNA to 180,000+ founders worldwide and discover your strongest traits as a startup founder. Get your Entrepreneur DNA report here.*

• HubSpot’s co-founder’s biggest lessons (Link)

• Unexpected things this founder learned from the YC Alumni reunion (Link)

💡 Opportunity: Multiplayer AI App

It’s time for a new social app.

TikTok has matured, Meta’s products reek of last-gen energy, and no new breakouts have emerged yet despite AI’s emergence into the broader zeitgeist being almost 2 years old.

X is rising but is also fundamentally a last-gen product.

We’ve seen novel single-player AI products like character.ai, but no breakout multiplayer experiences.

What is the easy-to-digest implementation of AI that will create new relationships and social experiences online?

• Facebook connected people

• Instagram let them express themselves with photos

• TikTok made creating engaging videos possible for anyone

What’s next?

🧠 Framework: When to Start a Startup

No one wants to buy in at the top, but many founders actually do — or at least closer to the top than they believe to be the case.

It surprised me when I realized that, but it’s true.

Too often people who want to start a startup will look for and base their decisions about what to build on the wrong signals.

Good signals usually come from potential customers.

Bad signals usually come from people talking about “the industry.”

The further removed someone is from the potential customers for a product, the less weight you should give their opinion on whether an opportunity is viable or not.

Here are some examples of when the consensus view was completely wrong:

This doesn’t mean you shouldn’t seek advice about the things you yourself don’t understand. It just means you should be careful what parts and how much of that advice to accept as gospel.

Instead, operate from first principles.

📈 Trend: AI Service as a Software

This trend probably needs a grammatically-correct name, but it’s becoming hard to deny.

The benefit is clear — in the past, performing these tasks required expenses and headcount to scale linearly with revenue, since traditional SaaS wasn’t able to process the information fast or accurately enough.

But LLMs change that.

I imagine the biggest companies in this space have already been founded, and we’ll see them become giant in the next few years, but there are likely many, many niche opportunities where this playbook can be applied in B2B.

💬 Quote: The Collison Install

These days most software gets sold as a subscription or is priced based on usage. We all know that, and most founders like it because it means predictable revenue (and, often, higher LTV).

But it also means more work is required from you and your team.

Getting someone to purchase your product isn’t the only thing that matters — you need to make sure they actually set it up, implement it, and use it.

Otherwise they’ll simply stop paying you for it (or, in the case of usage-based pricing, won’t ever start). This reduces your acceptable CAC and makes it harder to invest in growth.

The Collison brothers, who founded Stripe, realized this early on and took a more hands-on approach to sales:

💡 How I Can Help

🚀 Advertise in this newsletter to get in front of 60,000+ founders.