A first fund runs on a spreadsheet, a shared drive, a signature tool and the manager's inbox. That stack is not wrong. It is the correct amount of infrastructure for a fund that does not exist yet.
The question is which parts of it become expensive, and when.
What you genuinely need at launch
- Somewhere investors sign in that is yours. On your domain, with your name on it. For an unproven manager, looking institutional is not vanity; it is a substitute for a track record.
- A subscription flow that does not require you to be online. Investors close at inconvenient times. A flow they can complete unattended is worth more than any feature.
- Signatures with a real audit trail. Cheap to have from the start, effectively impossible to reconstruct retroactively.
- One list of who has committed what. Not three lists that mostly agree.
- Wire instructions behind a login. The cost of getting this wrong once exceeds the cost of every other item combined.
What can wait
Plenty, honestly. Detailed performance reporting, capital-call automation, integrations with a fund administrator, granular role hierarchies, and investor-facing analytics are all reasonable to defer until you have investors who are asking for them.
A first fund with fifteen LPs does not need a permissions matrix. It needs the fifteen people to close without friction and to feel looked after afterwards.
The three shortcuts that get expensive
1. Documents that get overwritten
Uploading LPA_final_v3.pdf over LPA_final_v2.pdf costs nothing today. It costs a lot the first time someone asks which version an investor actually signed, because the answer no longer exists. Keep every revision, always.
2. Consent recorded in email
"They confirmed by email" is fine right up until the person who received that email has left, or the thread is in a mailbox nobody has access to. Consent belongs attached to the investor record.
3. Bank details in a document
Wire details inside a PDF that gets emailed around is the shortcut that funds actual fraud. It is also the hardest to walk back, because once investors have learned that details arrive by email, a forged email looks normal to them.
The fund-two test
The useful way to evaluate any decision at fund one: what does this cost when there is a fund two?
Most operational shortcuts are fine, because fund two starts fresh. The ones that are not fine are the ones that create records: signatures, consents, disclosures, identity evidence. Those follow the investors, and investors follow you from fund one to fund two.
Put differently: be as scrappy as you like about workflow, and rigorous about evidence. Workflow you can replace. Evidence you cannot recreate.
Where a portal earns its place
Not by having more features than the spreadsheet. By collapsing the number of places a fact can live. When the commitment, the executed document, the signature evidence, the identity determination and the wire reference are one record, the questions that used to take an afternoon take a click, and the fund starts looking like it has an operations function well before it can afford one.