Most funds sign documents by sending investors out to a third-party e-signature tool and hoping they come back. It works. It also quietly costs more than it looks like it does.
The handoff is the expensive part
When an investor clicks "sign", they leave your portal, land on a page branded as somebody else, sign, and are then returned to a generic confirmation screen. Three things happen there.
- The brand breaks. An investor who has just been told this is your fund's portal is now looking at another company's product. For a first-time LP evaluating whether you are a serious operation, that is a real signal.
- The context is lost. The signing page knows about a document. It does not know about the investor's commitment, their entity, or the four other steps they are halfway through.
- The record splits. Your portal knows the subscription. The vendor knows the signature. Reconciling those is somebody's job forever.
What an audit trail actually has to contain
"We have an audit trail" is close to meaningless on its own. For a signature to be worth something in a dispute years later, the record has to answer all of these without reference to any other system:
- Who. The identity of the signer, and how that identity was established, not just the email address the link was sent to.
- What. The exact document revision, content-addressed. A hash of the bytes that were displayed.
- When. A timestamp that cannot be back-dated by anyone, including you.
- How. IP, user agent, and the sequence of events: opened, scrolled, consented, signed.
- Under what terms. The consent-to-electronic-signature disclosure the signer accepted.
The one that gets skipped is the document hash. Without it, the trail proves that someone signed something, at a time, from an address. It does not prove what they signed, which is the only fact that is ever actually in dispute.
Immutability is a design decision, not a feature
A signature ledger is only evidence if it cannot be edited afterwards, including by an administrator with database access. That means append-only records, no destructive migrations against the ledger, and executed PDFs stored as new objects rather than overwritten in place.
This is worth stating plainly because it constrains ordinary engineering. "Clean up the old rows" is a reasonable instinct on application data and a serious mistake on a signature ledger. Treating executed documents as legal records rather than app data is the whole discipline.
Countersigning and multi-party packages
Fund documents are rarely one signature. A subscription agreement usually needs the investor and then the GP. Trusts and joint accounts need two investors. Side letters need a negotiated round trip. Related: What Goes in a Fund Subscription Package and Who Signs What.
The practical requirement is that a package can sit in a partially executed state indefinitely, visible to both sides, with each party seeing exactly what remains. Tools built around a single "send for signature" action handle this by sending several separate envelopes, which is how a fund ends up with an investor who signed three of four documents and nobody noticed for a month.
What actually improves
The gain from bringing signing in house is not the signature. It is that the signature stops being a separate event. The investor never leaves, the document version is the one the portal served, the countersignature is a task in the same queue as everything else, and the completed record lands in the same place as the subscription it belongs to.
Which means the reconciliation job disappears, rather than getting faster.