Conviction

The stage

What the firm is selling

Judgment. A Conviction-stage firm has no operating rhythm to point at and no multi-generational structure to promise. What it has is a view, and a claim that the view is better than the market's.

Everything structural follows from that. If the advantage is judgment, then the configuration has to let judgment matter, which means fewer positions held at higher ownership in a scope narrow enough for the partners to actually have a view. A Conviction firm running a broad, thinly-held portfolio has diversified away the only thing it was selling.

Concentration is the point, not the risk

This is the stage where the model most visibly reads a configuration differently from how a naive risk view would. Sector concentration and single-company exposure that would be flagged at a later stage are read as structurally appropriate here.

The reasoning is that an emerging manager who spreads capital thinly across many sectors has not reduced risk in any way an LP cares about. They have taken on the market's return profile while charging fees for selection, and an LP wanting the market's profile has cheaper ways to get it. Concentration of conviction is what a Conviction-stage LP is buying.

The model does not treat this as licence for any concentration whatsoever. What it does is calibrate the reading, so that the same distribution which reads as a firm-level exposure at Continuity reads as a coherent early strategy here.

Succession is not expected yet

The one congruence check that reads lifecycle treats a Conviction-stage firm with no formal succession structure as aligned. This is worth stating because it is the opposite of what a governance-minded reader expects.

A firm on its first or second fund is its founding partners. Building the machinery to replace them is premature, and time spent on it is time not spent on the thing the firm is being paid for. The expectation arrives at Cadence and is enforced at Continuity.

The transition out is harder than the stage

The move from Conviction to Cadence is where most firms in this stage actually struggle, and the difficulty is structural rather than a matter of performance.

Hustle Fund lays it out plainly. Most firms begin raising Fund II something like two to four years after Fund I, and prospective LPs want to see what the first fund has produced. But early-stage companies take years to show anything. By the time Fund II is being raised, most of the portfolio will not have exited, few if any will have been acquired, and the returns that would settle the question do not exist yet.

So a General Partner is asked to prove a thesis on evidence the calendar has not produced. The mismatch is not a signal about the fund's quality, though it is easily mistaken for one, by LPs and by the General Partner. It is a timing problem built into the asset class: the fundraising cycle is shorter than the evidence cycle.

For the model this has a practical consequence. A firm raising Fund II is usually still declared at Conviction, and often should be. Fund sequence is only one of the transition triggers, and reaching Fund II is not one of them; the sequence trigger for Cadence is Fund III. A firm that reclassifies itself early, hoping the label helps the raise, will find every reading recalibrated against expectations it has not met.

What to watch

The characteristic Conviction-stage drift is scope widening while the portfolio does not. A firm that raised on a narrow thesis and began describing itself more broadly, without changing its portfolio count or its cheque, has kept its configuration and moved its story.

The second is the fund-size question, which is where the barbell argument becomes directly relevant: a Conviction firm scaling into the middle of the size distribution is walking into the part of the market several 2026 analyses argue is the hardest place to stand.

Sources

• Why is it so hard for VCs to raise Fund II?Hustle FundOn the structural mismatch between when Fund II fundraising starts and when Fund I can show anything an LP would call evidence.