“I Could Build This in a Spreadsheet.” Here's What That Version Actually Costs You.

Portfolio tracking spreadsheet vs software: the spreadsheet can do the maths. That is the honest answer, and anyone who tells you otherwise is selling you something. You can compute margin utilization, allocation drift and concentration in a sheet you build yourself in an afternoon. The cost of the spreadsheet version is not capability. It is that the same book, on the same day, can hand you three different answers depending on a choice nobody ever wrote down. If you already own a tool rather than a sheet, the dashboard version of the same objection is the sibling piece.

I built the spreadsheet first. I ran a leveraged, income-oriented book off a workbook with nine tabs for the better part of two years, and I want to be precise about what went wrong with it, because it was not the thing I expected. The formulas were fine. The arithmetic was fine. What failed was quieter: I could not reliably reproduce last month's number, I could not prove why I had trimmed a position in March, and if I had been hit by a bus, nobody in my household could have opened that file and known what to do on Monday.

Can you track your portfolio in a spreadsheet?

Yes. Genuinely, without hedging. I am not going to spend this post pretending the maths is hard, because it isn't, and because we have already published the opposite. Our fifteen-minute weekly review is a by-hand routine — pen, calculator, brokerage screen. The entire Governance Score method is published on this site, weights and bands and all, precisely so you can compute it yourself. And the piece on writing a policy before you buy any tool says plainly that you do not need software to start governing.

So let's set capability aside. It is a settled question and it is not the interesting one. The interesting question is what happens to a hand-built system over time — across several market regimes, a house move, a busy quarter, and one genuinely bad fortnight — and there the answer changes.

Your spreadsheet isn't wrong. It's unpinned. Nobody ever wrote down which of the defensible answers is yours.

The same book, three different answers

Here is the failure that convinced me, and you can reproduce it in ten minutes. Take one ordinary leveraged book, on one ordinary day. No ambiguity in the inputs at all:

• Holdings at market: $3,100,000
• Cash, held in another currency so it is not netted against the debit: $200,000
• Gross market value, holdings plus cash, before the loan is netted: $3,300,000
• Margin loan: $1,010,000

Now compute margin utilization. We read utilization against four zones — Clear under 30%, Harvest 30% to under 35%, Freeze 35% to under 40%, Forced 40% and above — and it is the zone, not the raw number, that tells you what to do. Hold that scale in mind, because the same book is about to land in three different places on it. There are three ways a careful person builds that formula, and every one of them is defensible enough to survive a conversation with your accountant.

One book. One day. One set of inputs. Clear, Harvest and Forced — three of the four zones, and three completely different instructions about what to do on Monday. Convention A is the one we use and the one I would argue for: gross securities value, holdings only with cash excluded, because that is the number that does not flatter you as leverage rises — and because a margin loan is itself a negative cash balance, so counting cash in the denominator nets the loan against itself. But the point is not that A is right and the others are wrong. The point is that your spreadsheet picked one of these when you wrote the formula, and you probably don't remember which.

That spread is not cosmetic, because the zone does not merely describe your leverage — it gates your governance. Our published scoring caps the Governance Score by margin zone no matter how clean everything else is: a book in Freeze cannot score above 65, and a book in Forced is capped at 34, which is the top of the bottom band — Exposed / At the Edge. Clear and Harvest carry no such ceiling. So read as C, this book's governance is uncapped and it competes on its merits. Read as B, it is hard-capped in the worst band on the board before a single other component is even considered. Same book, same day, same money — and the thing deciding which is a formula you wrote once and never wrote down.

A quick note on reading B, because this is where people scare themselves in one direction and reassure themselves in the other. Measured on the basis the regulator uses — equity as a share of the market value of long securities — this book sits at 73.9%, against FINRA's minimum maintenance requirement of 25%. On these numbers holdings would have to fall roughly 65% before that floor came into play. But do not take the wrong comfort from it: the FINRA minimum is a regulatory floor, not the thing that will actually liquidate you. Your broker sets its own house requirement, which FINRA notes can be higher than the regulatory requirement, runs higher still on concentrated or volatile positions, and — in FINRA's own words — firms "may increase the house requirements at any time and aren't required to provide you with advanced written notice," nor are they required to call you before selling.

Forced is not a proximity alarm, and the two numbers above are measured on different rulers. The 65% is the broker's ruler. Forced is measured on yours: it says you have spent most of the room your own policy allows, and a policy ceiling is set deliberately far tighter than anything a broker enforces, so that it bites while you still have choices rather than when someone else is making them for you. The real danger in reading B is that on a different day, with a different tab, you would have read A and done nothing at all.

You can't prove what you did

The second cost is the one nobody notices until they need it. A spreadsheet does not archive by default. You can beat that — save a dated copy every week, keep a one-line log tab, lean on version history — and a disciplined operator does exactly that; our own weekly routine ends with precisely such a log. But notice what you have just built: a second convention, unwritten, that has to be remembered on a week you are busy, moving house, or frightened. And a dated copy preserves the number without the policy version it was read under, so a reading taken before you nudged a threshold is quietly no longer comparable to today's.

That matters more than it sounds, because the rule that stops you overreacting depends on it. A single bad reading is noise; what warrants action is a breach that is still there a week later — the two-week persistence rule. Persistence is a comparison between two recorded readings taken on the same basis. Lose either the record or the basis and you are not measuring persistence, you are remembering it, and memory is exactly the faculty the whole discipline exists to replace.

A record that quietly rewrites itself isn't a record. It's a display.

It is worth seeing what the opposite looks like at scale. In February 2024 the Financial Times reported that Norges Bank Investment Management — manager of Norway's sovereign wealth fund, then around $1.5 trillion — had disclosed losing about $92 million (NKr980 million) through an incorrect date entered into the calculation of its mandated benchmark; the case is catalogued in the European Spreadsheet Risks Interest Group's register of documented spreadsheet failures. NBIM's global head of risk monitoring said he felt "physically ill." Let me be careful about what that proves, because it is tempting to over-read. A wrong date is a bad input, not an unpinned convention — a deterministic engine fed the same wrong date would have produced the same wrong answer very reliably indeed. What it demonstrates is the other half of the argument: NBIM was able to state publicly what went wrong, when, and precisely what it cost — down to the krone. That is what accountability looks like when a calculation can still be examined after the fact. The question worth sitting with is whether your workbook could do the same for a decision you made eighteen months ago.

There is a harder version of that question — whether you could reconstruct, from records rather than recollection, the reading that actually drove a decision. That question now has its own post: the five fields a decision record has to carry. For now it is enough to notice that a tool which overwrites by default cannot answer it, and that an investment process you cannot reconstruct is one you also cannot audit, cannot improve deliberately, and cannot defend to a spouse, a co-trustee, or your future self.

It can't outlive you

The third cost is the one I find hardest to argue people out of ignoring, because it is the least urgent and the most consequential.

Your workbook encodes a great deal that exists nowhere else: which tab is authoritative, why the third column has a manual override, that the crypto row is deliberately excluded from the drift calculation, that you always refresh prices before the loan figure and not after. None of that is written down. It lives in you.

The file is not the system. The file plus your memory is the system — and only one of those has a backup.

So the test is simple and slightly uncomfortable: could a competent person who is not you open that file on a Monday morning and produce this week's reading correctly, without calling you? For most hand-built systems the honest answer is no, and that means the governance of a seven-figure estate has a single point of failure, and it is a person. This is precisely the risk that portfolio governance — the operating layer that family offices run and solo investors skip — exists to remove. An institution's process survives the departure of any individual by design. That is not bureaucracy; it is what makes it a process rather than a habit.

When does a spreadsheet stop being enough?

Not always, and I want to be concrete rather than self-serving about this. If you hold a dozen positions in one currency with no leverage and no options, a spreadsheet is genuinely sufficient and you should not buy anything. The threshold is not portfolio size — it is the number of places a convention can silently diverge. In practice it is crossed when:

• You carry a margin loan. The reading gates everything else, it moves every day, and as the table above shows, it is the calculation most sensitive to an unwritten convention.
• You hold options. Premium versus notional, realised versus mark-to-market, assigned stock versus the original put — each is a judgment call, and a sheet makes each one invisibly.
• You hold more than one currency. Which rate, sourced when, applied to which leg.
• You run several strategy Layers — a Layer being one strategy in the Incomestead Stack, the tier an institution would call a "sleeve." Allocation drift is only meaningful against targets that are recorded and unchanged; a target you can edit in the same cell you measure against is not a target.
• Somebody other than you might need to run it. A partner, a co-trustee, an executor, or you in fifteen years.

Notice that four of those five are about consistency over time, not difficulty. That is the shape of the whole trade-off.

The Spreadsheet Stress Test

Here is the copy-paste artifact. Run it against your own workbook this week — not a hypothetical one. It takes about fifteen minutes and it is deliberately designed so that a good spreadsheet can pass. If yours passes all six, keep it, and take the honest win.

Every one of those six is a question about reproducibility, record or succession. None of them is a question about whether you are clever enough. That is the entire argument of this post in six lines.

What software doesn't buy you

I would rather you finish this page slightly sceptical than oversold, so let me mark the limits plainly. Software does not make better decisions than you do. It does not predict anything. It does not know whether your targets are wise — a badly chosen threshold applied perfectly is still a badly chosen threshold, and no engine will save you from that. It cannot tell you whether this is a good week to reduce leverage; it can only tell you, without flattery and without drift, what your leverage is and whether the breach has persisted. And it does not trade. That is not a missing feature, it is the posture: the instrument reads, ranks and recommends, and the decision stays with the person whose money it is.

There is one more limit, and it is the one a vendor is least likely to volunteer. A tool has key-person risk of its own. Software companies get acquired, change direction, or fold, and a governance record you cannot take with you is a hand-down problem wearing a different hat — the very thing this post has just spent a section warning you about. The only honest answers are portability and publication: your snapshots and your written policy should be yours in a format you can export and read without us, and the method itself should be public, which is why the whole scoring approach is on this site for free rather than behind a login. If we vanished tomorrow you should still hold your history and still be able to run the reading by hand. Any tool that cannot say that to you is asking for more trust than it has earned.

What determinism buys is narrower and duller than a sales page would like — the same answer twice, a record that does not quietly rewrite itself, and a process that does not evaporate when you do. For a hand-built workbook running a leveraged seven-figure book, those three are worth more than any feature.

Frequently asked questions

Can you track your portfolio in a spreadsheet?
Yes. Margin utilization, allocation drift and concentration are all straightforward arithmetic, and we publish a by-hand weekly routine and the full Governance Score method so you can do exactly that. Capability is not the constraint. The constraint is whether the same inputs reliably produce the same answer, whether last week's reading still exists on the same basis, and whether anyone but you could run it.

What's the difference between a portfolio tracking spreadsheet and portfolio governance software?
A tracker reports what you hold right now. Governance pins the calculation convention outside the formula, records each weekly reading immutably so a breach's persistence can be measured, applies written thresholds you can change without editing formulas, and produces a ranked action list. A spreadsheet can imitate any one of those; it struggles to hold all four consistently for years.

When does a portfolio spreadsheet stop being enough?
When conventions can silently diverge — typically once you carry a margin loan, hold options, hold more than one currency, run several strategy Layers (a Layer being one strategy in the Incomestead Stack) against recorded targets, or need someone other than you to be able to run it. A dozen unlevered positions in one currency does not need software.

What happens to my portfolio spreadsheet if I'm not around to run it?
Usually nothing good, because the file is only half the system — the other half is your memory of which tab is authoritative and why the overrides exist. Test it rather than assume: hand the workbook to a competent person and see whether they can produce this week's reading without calling you.

Written by Stefano Starkel, founder of Incomestead, from lived experience running a leveraged, income-oriented portfolio — as a practitioner, not a licensed adviser. This is education, not personalized investment advice; the figures above are illustrative. Margin amplifies losses as well as gains, a broker can liquidate positions without contacting you, and house margin requirements can be raised without notice. Consider your own situation and, where appropriate, a professional who knows it.

Incomestead recommends. You decide.