The decision test for building an investor reporting package is simple: leave out anything that cannot change a decision in the next quarter, either yours or your investors'. Founders usually weigh two versions of this document, the everything-included deck and the one-page cash note. The right cut sits between them, and a criterion, not habit, should make it.
The one test: could this line change a decision?
The test is whether someone would act differently depending on the number. A decision means hiring or freezing a role, changing a price, cutting a program, drawing on a credit line, or an investor making an introduction you need. If no plausible action changes when the figure moves, the line is decoration, and decoration is exactly what you leave out.
Founders tend to draft one of two packages. The first includes every tab the finance system can export, as if the update were a defense file for an audit that never arrives. The second sends cash and a paragraph, treating the update as a chore between real work. Both fail the same test in opposite directions, and both train readers to skim.
Decisions run in both directions, which is the part founders miss. An investor reading your update might decide to double down, make a customer introduction, or flag a hiring risk before it costs you a quarter. You might decide to slow spend or fix churn this month instead of next. A metric that only you act on still belongs, because the package doubles as your own operating review.
Cadence shapes the size too. A monthly update can carry five numbers and three paragraphs; a quarterly one owes readers more depth, since more has happened and the next chance to correct course is further away. Decide the cadence first, then let it set how much each section can hold.
Picture a hypothetical seed-stage company with nine months of runway. Its useful update carries cash, burn, the hiring plan, the sales pipeline, and the two product milestones that matter this quarter. Its bloated version adds eleven spreadsheet tabs, a cohort analysis with no cohorts behind it, and a hiring funnel chart nobody has asked about. Same company, Cfo Service twice the work, half the readership.
A practical way to apply this: draft the memo first, in three parts. What changed since the last update, what we are doing about it, and what we need from readers. Then attach only the numbers that support those three answers. The useful question is not what investors might want to see. It is what would change a decision this quarter, and once you answer it, the cut list writes itself.
Cut first: numbers nobody acts on
The first cuts are easy to spot once you stop defending them: vanity metrics, cumulative totals, and breakdowns too small to matter. Registered users, lifetime sign-ups, follower counts, page views for a company that sells annual contracts. These numbers only move in one direction and no decision hangs on them.
Vanity metrics share a shape. They rise forever, they grow with activity that produces no revenue, and they never appear in an argument about what to do next. Try finishing the sentence "if this number falls next month, we will..." with a vanity metric in the blank. It cannot be done, and that failure is the cut.
Here is a hypothetical example with the arithmetic out in the open. A services firm has $2.4M in annual revenue: three clients at roughly $800K each, plus about $30K of incidental hardware resale. Giving that $30K its own chart, tab, and commentary spends attention on about 1% of revenue. Fold it into an "other" line and use the space for the renewal dates of the three clients who actually matter.
That folding rule is materiality in plain clothes. If a line is a small single-digit share of revenue and nothing triggers on it, aggregate it. The exception is small numbers that carry risk: a refund rate creeping up, one large customer paying late, churn ticking in a segment you just entered. Signal beats size, and risk lines stay in however small they look.
Prefer period figures over cumulative ones in every table. Customers added this quarter informs; customers since launch decorates, because the total hides churn and the trajectory disappears inside a number that can only go up. Anyone who wants the cumulative figure can add it, which is the reader's job, not the writer's.
Process narration deserves the same blade. Sprint boards, meeting recaps, conference attendance, tool migrations. If a reader cannot respond to it with money, a customer, or a warning, it is a diary entry, not reporting. One paragraph of operational highlights is plenty, and only when tied to something measurable.
Notice what is not on this list. Revenue, margins, cash, customer counts tied to revenue, pipeline with values attached: those stay. The goal is a package lean enough that every surviving number has a job to do.
Forecasts: leave the precision out, keep the assumptions
In a forecast, the thing to leave out is false precision. A twenty-four month projection with monthly figures modeled to the dollar, three quarters after it was built, signals effort rather than knowledge. Rounding is a form of honesty, because it tells the reader which parts of the plan are estimates and which are commitments.
Keep the drivers instead. New leads, conversion, average contract size, the headcount plan, the collection pattern of receivables. Those few lines carry the whole forecast, and cfo service a reader who disagrees with your plan will disagree with a driver, never with a decimal. Put the drivers in the update and leave the forty-tab engine in the model.
A forecast only earns its place if you keep revisiting the assumptions, since that is what makes it information rather than decoration. Each update should show what changed since the last version and why: the enterprise deal that slipped a quarter, the hire who started late, the payment terms a big customer demanded. Revisions build more trust than stability ever will, and a forecast that never moves is a forecast nobody believes.
Cut scenario overload at the same time. Base, downside, and upside are enough, each with a named trigger, for instance the downside arriving if pipeline coverage stays thin for two consecutive quarters. Twelve sensitivity tabs answer questions nobody asked and bury the two or three that matter.
If you are preparing to raise while you keep reporting, the forecast also connects the operating plan to the ask. Leaving out the burn plan and the milestones it funds makes a raise look like an event that happened to you rather than a plan you are executing. Keep the discipline, drop the promise: a forecast supports a raise, it does not guarantee one, and markets, timing, and luck all get a vote.
In practice this takes one small table: actual, prior forecast, and current forecast for cash and revenue, plus three bullets on what changed. That replaces pages of model printouts and survives being read on a phone.
Runway math earns its place in every version
Cash position, net burn, and months of runway are the last things a founder should cut. Whatever else changes with stage or audience, these three numbers anchor the package, because they set the clock on every other decision inside it.
Show the calculation, not just the result. In a hypothetical example, a company with $960K in the bank and an average net burn of $160K a month has six months of runway: 960 divided by 160.