Thesis-Driven Investing

What is thesis driven investing?

Thesis driven investing means deciding in advance where value will be created, writing that view down, and then sourcing against it. The fund picks a change it believes is happening, maps the companies that would win if it is right, and goes looking for them instead of waiting for introductions.

What does a thesis actually look like?

A usable thesis is a sentence about the world, not a list of sectors. Union Square Ventures publishes its own as an example of the form: USV invests at the edge of large markets being transformed by technological and societal pressures. The firm also publishes how the statement has changed, which is the more instructive part. In its Thesis 2.0 period it began to look for narrower vertical networks, for example in learning and health, new layers of enabling infrastructure such as developer tools and fintech, and decentralized computing systems. Under Thesis 3.0 it looked to leverage those networks to broaden access to key resources such as knowledge, capital and wellbeing, and looked for opportunities to increase trust in those systems. Approaching 2020, with the climate crisis accelerating, it sought to identify transformational startup opportunities to mitigate its impacts, accelerate the energy transition, and adapt to a changing world.

Note what those statements do. Each one names a mechanism, and each one is falsifiable within the life of a fund. Compare that with a stated focus on artificial intelligence, which excludes nothing and predicts nothing.

Why a thesis is an economic choice, not a marketing one

The case rests on how venture value is distributed. Cambridge Associates found that an average of 61 firms account for value creation in the top 100 investments in venture capital per year, and that after 1999, investments ranked 11 through 100 accounted for an average of 60 percent of the total gains generated by the top 100 investments per investment year, more than the top 10 contributed. A large share of the value sits outside the handful of firms everyone can name, and the firms capturing it change from year to year.

That has a direct consequence for a small fund. It cannot see every deal, and it does not have to. It has to be present early in a narrow set of places where value will concentrate. A thesis is the instrument that chooses those places and, more importantly, gives the team permission to say no to everything else.

The mechanics

• Write the claim. One or two sentences describing the change, the mechanism, and the time horizon over which it should be visible.
• Derive the implications. If the claim is true, which categories of company become viable that were not before, and what has to be true about cost, regulation, distribution or behavior for that to hold.
• Build the market map. Enumerate every company operating against the implication, including the ones that are too early, too small or funded by someone else.
• Define the disqualifiers in advance. Which observations would make the thesis false, and what the fund does if they appear.
• Source against the map rather than against the inbox. Outbound contact with mapped companies, plus relationships with the researchers, operators and buyers inside the category.
• Review on a schedule. A thesis reviewed only when a deal needs justifying is not a thesis.

Worked example

The figures below are hypothetical and chosen so the arithmetic can be checked.

A $60 million fund charges a 2 percent management fee over a ten-year term. Fees are $1.2 million a year, so $12 million across the term, leaving $48 million of investable capital. The construction is 20 initial positions of $1.8 million, which is $36 million, with $12 million of reserves.

The thesis: claims adjudication inside mid-sized insurers is being rebuilt because the adjusters who hold the domain knowledge are retiring faster than they are replaced. The implication: software that encodes adjudication judgment can be sold to carriers that have historically refused to buy workflow tools.

Sourcing against that map, over 18 months the team identifies 140 companies in the category, meets 42 of them, takes 9 into diligence, and invests in 3. The hit rate against the mapped universe is 3 divided by 140, which is 2.1 percent. The same team reviewed roughly 1,200 inbound decks in the same period and invested in none of them, a rate of 0 percent on a much larger denominator.

Ownership and the fund-returner test: a $1.8 million check at a $15 million post-money valuation buys 12 percent. Diluted to 7 percent by exit, returning the entire $60 million fund from that one position requires an exit value of $60 million divided by 0.07, which is about $857 million. The thesis is what makes a bet on an $857 million outcome in mid-market claims software an argued position rather than a hope.

Three investments in one category out of 20 positions is 15 percent of the portfolio by count and, at $1.8 million each plus a share of reserves, a concentration an LP will ask about. That question is the thesis's real test.

Where the thesis appears in the fund's paperwork

In the investment memo. Bessemer Venture Partners publishes its historical memos as the early analysis behind its investment decisions, and the documents read as thesis applications. The firm's Shopify memo, dated October 12, 2010, opens by seeking approval to invest up to $7 million in the Series A financing of Shopify, described as a provider of e-commerce software to small and medium businesses. The memo makes the category claim first and the company claim second, which is the structure a thesis imposes: the fund is arguing that this space will matter, then that this team wins it.

A thesis also shows up in the LPA's concentration limits, since a thesis-driven portfolio tests them; in the quarterly LP letter, where the thesis is restated and marked against reality; and in the reserve policy, because a fund concentrated in one category needs a view on funding the whole cohort rather than picking winners late.

Common mistakes

• Confusing a thesis with a sector label. Categories are containers; a thesis is a claim.
• Letting the thesis approve the deal. The thesis says a company is worth studying. Team, price and evidence still decide.
• Never marking the thesis to market. If nothing has been learned in two years, the thesis has become identity.
• Quiet drift. A fund that raised on vertical software and holds six infrastructure positions has changed strategy without telling anyone, which is a governance problem before it is an investing one.
• Mapping only the fundable. The interesting names in a new category are usually pre-institutional, and a map built from funding announcements finds the companies a fund is already too late for.

Related terms

A thesis is the claim itself; thesis-driven investing is the operating discipline built on it. Thesis drift is the failure mode, where the portfolio stops matching the stated claim. A market map is the artifact that turns a thesis into a sourcing list. Top-down investing is the broader family this belongs to, as against opportunistic sourcing, and a sector specialist fund is the extreme version, where the thesis and the fund's identity are the same thing.

Related tools and reading

• Carry Calculator
• Waterfall Calculator
• PitchBook Alternatives and Pricing
• VC Fund Pitch Deck Examples