The Systems CFO Newsletter

Every CFO I've worked with — at Atari, at BeyondID, at companies scaling from $9M to $180M — eventually hits the same wall. The forecast is off. Not by a little. By enough to change board conversations, delay hiring plans, and erode trust with investors.

The instinct is to fix the model. Add more variables. Increase granularity. Build in more scenarios. But the problem isn't the spreadsheet. It's the information architecture underneath it.

When sales, marketing, and finance each optimize their own numbers independently, the forecast becomes a negotiation artifact — not a signal. The sales team sandbagging to beat quota. Marketing inflating pipeline to justify spend. Finance discounting both to protect credibility with the board. Three rational actors, each optimizing locally, producing a system-wide forecast that nobody trusts.

The Systems CFO approach starts somewhere different: map the feedback loops between the teams that produce the forecast. Identify where information degrades as it moves between systems. Find the constraint — usually a single handoff point — where forecast accuracy breaks down. Then redesign that one junction.

I've seen this work at companies of every size. The fix is never more data. It's better architecture.