Building and Publishing a Thesis

What a thesis is (and isn't)

A thesis is a view about where value will be created, specific enough to act on and to be wrong about. "AI will change everything" isn't one. "Mid-sized logistics companies will buy software that replaces manual dispatch within five years, because the underlying models just became cheap and reliable enough" is.

A useful thesis has five parts:

The last row is what separates a thesis from a sales pitch.

Firm theses and personal theses

Some firms publish a thesis for the whole firm. Union Square Ventures has done it publicly for years; its Thesis 3.0 describes backing trusted brands that broaden access to knowledge, capital and well-being by leveraging networks, platforms and protocols. Others work bottom-up: Bessemer's partners each develop themes, which the firm calls roadmaps, research them, and bring them back to the firm to invest against.

As a principal, you'll usually work inside the firm's thesis and own one or two of your own within it. That's what gives you a reason to call founders before they call anyone.

How to build one

1. Pick a space where you have an edge: prior work, a network, or simply more hours of reading than anyone else.
2. Map it. Every company, incumbent and recent round. The market map you'd have built as an analyst, now with a point of view attached.
3. Talk to the people who'd buy. Customers and operators tell you what's actually changing faster than founders or investors do.
4. Write down what has to be true, and what would change your mind.
5. Test it against companies. Meet founders in the space whether or not they're raising. The thesis should get sharper with every meeting.

Why publishing is the point

A thesis in a notebook helps you evaluate deals. A published one also finds them. Founders building in the space read it and get in touch, other investors send you deals that fit, and your partners and LPs see how you think. Writing in public also forces clarity: it's hard to hide a vague idea in a post other people will read.

Published doesn't have to mean polished. Short, frequent, specific posts compound. The resources below include investors who built reputations each way: short and frequent, or long and rare.

The failure modes

• Too broad to act on. If every company in the sector fits, it isn't a thesis.
• Blinders. Missing the best company in a market because it didn't match what you wrote.
• Never updated. A thesis should change when the evidence does. Revisit it at least once a year, and say publicly when you've changed your mind.

What this means for you

Start one now, before you need it for a promotion case. Pick a space, write the one-page version with all five parts, and share it with your partners. Then publish something from it. The first founders who get in touch because they read it will tell you whether it's specific enough.