Spreadsheet, Software, or Counsel-Managed Cap Table?
A spreadsheet is adequate while the structure is simple and one person maintains it. Move to dedicated software when convertible instruments, an option pool with ongoing grants, and a priced round coexist, because that is when modeling errors become expensive. Whichever you choose, the signed documents remain authoritative and the tool is a representation of them.
What actually breaks, in order
Cap table problems rarely start as arithmetic problems.
Unrecorded promises. An offer letter mentioning equity that was never granted, an advisor promised a percentage in a conversation, a co-founder's arrangement agreed verbally. None of this appears in any tool, and all of it surfaces during diligence when it is expensive.
Convertible instruments. SAFEs and notes are not ownership until they convert, and their effect depends on the priced round that has not happened yet. Modeling that properly means handling caps, discounts, whether conversion is pre-money or post-money, and how the option pool interacts. This is where spreadsheets produce confidently wrong answers, because each assumption is a formula someone wrote once and nobody re-derived.
Option pool mechanics. Authorized versus granted versus outstanding versus available, plus expirations, forfeitures, and early exercises. Pools drift out of date quietly, and the number founders quote in a pitch is often the authorized figure rather than the available one.
Multiple people answering the same question. Once a founder, a lawyer, and an investor each maintain a version, they diverge. That is a coordination failure rather than a tooling failure, and it is the most common reason companies move to shared software.
Notice that only the second one is a modeling problem. The rest are record-keeping discipline, and no software fixes discipline.
The rule that prevents most of it
Nothing exists until it is documented and signed, and the cap table reflects documents rather than intentions. A promise made in an offer conversation is a liability that has not been recorded yet. Treating the executed stack as authoritative, and the tool as a representation of it, resolves most disputes before they become disputes.
When to move off a spreadsheet
A spreadsheet is genuinely fine for a small number of holders, a single class of shares, and no convertible instruments outstanding. Building the model yourself is also the best way to understand your own structure, which is an argument for starting there deliberately rather than by default.
The signals that it is time to move are specific.
You have convertible instruments outstanding and are approaching a priced round, so conversion modeling now determines real ownership outcomes.
Option grants happen on an ongoing basis rather than in occasional batches, which means the pool changes continuously and someone has to keep it current.
More than one person needs an authoritative answer, particularly under time pressure during a financing.
You are being asked for a cap table by an investor and the version you would send requires an explanation to interpret.
Or you are approaching an event with formal requirements: a priced round, a valuation exercise, an audit, or an acquisition.
One caution about migration timing: moving during a live financing is how errors get frozen into a new system. Move before the round, when nothing is in flight and there is time to reconcile against the documents.
How to choose, and what none of them fix
If you move to software, evaluate against the things that go wrong rather than the feature list.
Does it model convertible instruments the way your instruments actually work, including post-money mechanics and pool interactions? Can it produce a scenario analysis you would send to an investor without editing? Does it maintain an audit trail showing who changed what? Does it export cleanly, so leaving is possible? And does it accommodate your counsel working in it, since the people preparing your documents should be able to see the same record.
Counsel-managed records sit at the other end: strongest on correctness and document alignment, weakest on self-service modeling, and dependent on someone else's availability when you want to run a scenario at eleven at night before a meeting.
What none of these approaches fix is the discipline problem. Every tool records what you tell it. The failures that actually damage companies are grants nobody entered, promises nobody documented, and a pool figure that stopped being accurate two quarters ago.
This is general information rather than legal advice. Cap table structure has real legal and tax consequences, and both the instruments and the records are worth reviewing with counsel before a financing rather than during one.
| Approach | Best for | Real cost | Where it breaks |
|---|---|---|---|
| Spreadsheet | Few holders, one share class, no converts | Your time, and single-maintainer risk | Convertible conversion modeling and option pool drift |
| Cap table software | Ongoing grants, converts outstanding, several stakeholders | Subscription plus migration and reconciliation effort | Garbage in: it records what you enter, correctly |
| Counsel-managed | Correctness and document alignment | Turnaround time on every question | Self-service modeling at short notice |
| Mixed, counsel plus software | Companies approaching a priced round | Coordination between two records | Divergence if neither is declared authoritative |
Frequently asked questions
- When should a startup stop using a spreadsheet for its cap table?
- When convertible instruments are outstanding and a priced round is approaching, when option grants happen continuously rather than in batches, or when more than one person needs an authoritative answer under time pressure. Those are the conditions where modeling errors stop being cosmetic and start determining real ownership.
- What is the most common cap table error?
- Not arithmetic. Unrecorded grants and undocumented promises: equity mentioned in an offer conversation but never granted, an advisor promised a percentage verbally, a co-founder arrangement agreed but not papered. None of it appears in any tool, and all of it surfaces during diligence.
- Does cap table software prevent mistakes?
- It prevents modeling mistakes and it records exactly what you enter. It cannot know about a grant nobody recorded or a promise nobody documented, which is the category that actually damages companies. The discipline of documenting before promising is what prevents those, not the tool.
- What is the source of truth for ownership?
- The executed documents. Any tool is a representation of them, and where the two disagree the documents control. Treating the signed stack as authoritative and reconciling the tool against it resolves most cap table disputes before they escalate, and it is the check to run before any financing.