New

Peaky raised €1.1M, led by JK Invest

→

All articles

Company

/

Written by Peaky

The 4-day forecast rebuild: a finance hangover story.

29 days
Unnoticed drift
29 days
Unnoticed drift

For twenty-nine days, the drift went completely unnoticed.

Sales hired a fifth AE in the second week of January, three weeks ahead of plan. The CRO had cleared it with the CEO. Nobody told finance, because nobody thought to. The new hire was visible in the HR system from day one. The cost was visible in the next payroll run. Nobody had connected those two systems to the financial plan, because that was what the monthly close was for.

In the same window, marketing kicked off a campaign that doubled spend on paid social. CAC moved with it. The ad platform showed the new numbers from week one. The forecast still ran on the old number, because nobody had updated it. There was no formal moment to update it. The forecast was a quarterly artifact. The campaign was a Tuesday decision. They lived in different timeframes.

Pricing shipped a new enterprise tier on the third week. Stripe started picking up larger ARR per customer. The model was still calculating revenue with the old tier mix. The error was small at first. By the end of the month it was not.

Then the actuals hit. The CFO opened the file on the second of February. Burn was 18% over plan. Revenue mix was wrong. CAC was up 40%. The campaign had not been in the model. The hire had not been in the model. The tier had not been in the model. Each individual change had been visible somewhere in the operational stack for the entire month. None of them had been visible in the plan.

The next four days were the rebuild.

Day one

The first day was reconciliation. The CFO pulled the accounting export, the payroll run, the Stripe data, the ad spend report, the CRM export. None of them shared a structure. None of them used the same period boundaries. The accounting platform closed on the last calendar day of the month. Stripe ran on a different cycle. Payroll happened in the middle of the month. The CRO’s pipeline report ran on his own cadence and did not include any costs.

By the end of day one, the CFO had a spreadsheet with raw data in six tabs, no headers aligned, no totals reconciled, and a flag in column G next to seven line items that needed clarification.

The hires were not in HR’s standard report. Two of them had been booked against a budget code from last year. The senior PM who had started on January eighth was being paid as a contractor. He was not flagged as a contractor in the HR system. He was flagged that way in payroll. Nobody had noticed the inconsistency until the CFO went looking.

The campaign spend included a media buy that had been authorised but not yet booked. The accounting platform did not show it yet. The ad platform did. The reconciliation needed to include both, but the way the model was structured, only one would feed through. The CFO had to decide which version to trust for the forecast and add a manual override for the other.

By 7pm, the data was sort of reconciled. The variance was still wrong.

Day two

Day two was the model rebuild. The original forecast had assumed four AEs by end of Q1. The model had hardcoded the headcount in row 23. The CFO had to change the assumption, but changing the assumption broke the formula in seventeen downstream cells, because the original formula referenced the headcount as a fixed number instead of a driver.

This is the kind of work nobody talks about until they have done it. The model that was “right” in January was right because every assumption was hardcoded. Changing one assumption surfaced the structural debt across the entire spreadsheet. The CFO spent six hours pulling out hardcoded values, replacing them with driver references, fixing broken formulas, and validating that the new totals reconciled to the corrected actuals.

By the end of day two, the model worked. It also did not match the version anyone else in the company had a copy of. Every dashboard, every board pack, every operating review for the last six weeks had used the old version. The CFO sent a Slack to the CEO at 11pm saying the numbers we have been using are wrong. The CEO replied at 6am the next morning asking how wrong.

Day three

Day three was the impact analysis. How wrong, exactly.

Burn was 18% over plan. The campaign and the hires explained most of that. The pricing tier change moved revenue forward in the year but not enough to offset the cost. The cash runway model had been showing eighteen months. The new model showed thirteen.

Five months. Lost in twenty-nine days of unnoticed drift.

The CFO spent the day building scenarios. What if the campaign continues. What if it stops. What if the hires were the only acceleration. What if pricing offsets more than expected. By the end of the day, there were six versions of the model, each with a different set of assumptions, each with a different runway answer.

Then the CFO had to pick one to present.

Day four

Day four was the board pack rebuild. New numbers. New story. A version of the narrative that explained why the previous month’s pack had been wrong, what had changed, and what the team was going to do about it.

The pack went out on the morning of day five. The CFO had not had a full weekend in three weeks. The CEO had a tight meeting on the runway model the following Monday.

The board’s first question was: how did we not see this sooner.

Why this happened

The CFO did not do anything wrong. The CRO did not do anything wrong. The CMO did not do anything wrong. Each of them was running their function correctly inside their own cadence.

The problem was structural. The finance plan refreshed once a month. The business changed every day. The gap between those cadences was where the drift hid. By the time the close ran, the drift had compounded for twenty-nine days. The cost of that compounding was the four-day rebuild plus the runway model being wrong by five months.

This is not a horror story. It is a Tuesday for most growing companies. Every CFO at scale has lived some version of it. Most have lived several.

What would have prevented it

There is no version of this story where the underlying changes did not happen. The hires were the right hires. The campaign was the right campaign. The pricing tier was the right pricing tier. The decisions that drove the burn were not bad decisions.

What would have prevented the four-day rebuild is the financial plan being connected to the systems where those decisions showed up. The HR system. The ad platform. Stripe. If the plan had been pulling those signals continuously, the drift would have been visible by January tenth. The campaign decision would have included its true CAC impact. The hiring decision would have included its true burn impact. The pricing change would have flowed through the model the day it shipped.

There would have been no rebuild. There would have been continuous adjustment. The four days would have stayed where they belonged. With the CFO’s family, or sleep, or anything else that is not reconciling a spreadsheet.

This is what we built Peaky to make possible.

You can see what continuous financial planning looks like on your own data at peaky.ai.

The financial infrastructure for scaling without surprises.

Previous

How CFOs lose their Sundays (and how to get them back).

Next

Why your monthly close keeps surprising you.

Keep exploring

Getting started

From your Excel plan to a live plan in a day.

Integrations

The systems Peaky connects to, and what each one brings in.

The Forecast Audit

Nine questions to find out whether your forecast is a report or a tool.

Peaky app icon

Sync or sink.

Ask AI for a summary of this website