The Missing Record

The Strongest Case Against This, and Where It Lands

The obvious rebuttal is that this is early, and early looks like this. The 6% will grow as agents mature; the plumbing problem is exactly what agents are being built to solve; and the record is a feature the vendors will add. Arch already keeps the diligence file attached after close, Gridline already markets a persistent diligence record, and Bipsync's Toro already answers questions over a firm's own research history. On this view, "the decision record" is a roadmap item, not a missing layer, and the allocator's job is to wait twelve months and buy it.

Two parts of that hold and one doesn't. Agents will improve the plumbing; the data-layer consolidation makes extraction cheaper every quarter; and the vendors will keep moving upstream. What the data does not support, in our view, is the idea that time fixes the impact gap on its own. The 6% didn't move during a year of rapid model improvement, and McKinsey's high performers are distinguished by workflow redesign and measurement, which no vendor can ship to you. The deeper question is what "record" means. As we understand these products, a diligence file kept after close is primarily an archive of documents that tells you what the GP said. A decision record is a different object: it holds what the firm concluded and why, who dissented, and what happened next, across every mandate, in a form that can be queried and compared. The vendors are converging on the vocabulary. None of them hold the judgment, because the judgment isn't theirs to hold. Our own build has made the gap concrete, and the blockers have mostly not been technical. They come from the fact that a decision record has to reflect how one particular firm reasons:

  • Judgment is spread across people, and most of it is never written down. The partner who has watched a manager through three cycles has a view. So does the operating expert who knows which value-creation claims to discount, and the investor relations lead who remembers what a client asked last time. Today that perspective comes out in an IC conversation and leaves with the person. A record worth having gives experienced people somewhere to put that context, next to the numbers it qualifies, so the next decision can draw on it. No vendor can supply this, because it comes from a particular mix of people at a particular firm.
  • The nuance usually sits outside the number. A reported figure may be a target, a blend of several figures, or a point someone on the team disputes. A system that keeps only the value loses the reasoning that made it usable.
  • A view is not a snapshot. A firm's read on a manager changes as evidence arrives and as people disagree. If dissent is overwritten instead of kept, the record cannot show how a conclusion was reached.

None of this can be configured into a product, because it is a description of the firm itself. That is the part of stage four an allocator cannot buy, and our reading of the surveys is that few firms have started on it.

This content is for informational purposes only and does not constitute investment advice, an offer, or a solicitation to buy or sell any securities. Thirdpath, LLC is not acting as a fiduciary or financial advisor in providing this content. View our full disclosures.