Skip to content
← The intelligence layer

A forecast is a range from the budget engine, not a promise from a model.

The budget engine projects spend, customers, CAC and margin forward from measured response, current pacing and the present mix, and it returns ranges rather than single numbers. The model reads that output and explains which assumptions carry the range — saturation in a channel, the learning phase after a change, seasonality in a market. It writes what would have to be true for the high end and for the low end. Each forecast is stored and later scored against what your finance system actually records.

In practice

Demo data

If I spend €55,000 next month instead of €50,000, what happens to CAC?

Blended CAC rises to between €148 and €157, from €144.69 today. The budget engine puts the extra €5,000 where headroom remains — TikTok and Meta retargeting — because Meta prospecting flattens above about €22,000 a month and sits at €19,800 already. Evidence: 90 days of channel response from the metrics service, current pacing, and invoiced customer value from your finance system. Business impact is roughly 32 to 38 additional customers, with gross margin landing between 63.8% and 65.2%. Confidence is medium: the top of the range assumes TikTok holds €148 CAC through a creative refresh it has not had yet. The alternative explanation sits on the margin side — Germany is credited with the most revenue and the least margin at 61.1%, so extra spend landing there lowers margin without moving CAC. Recommended action is to stage the increase at €2,500 for two weeks and re-read before committing the rest. Constitution check: €55,000 exceeds your €50,000 monthly budget limit, so this stays advisory and no action can execute until a person raises the limit.

Ranges carry their assumptions

A customer range is shown with the response curve behind it: the monthly spend level at which a channel's prospecting flattens, and how close it already sits to that level. Change that assumption and the range moves with it.

Forecasts are scored later

Last month's projection is compared against collected revenue and customer counts once invoices settle. The error is kept per channel, so a channel that is consistently over-forecast is discounted the next time it asks for budget.

Learning phase is priced in

A budget change resets a network's optimisation and the forecast includes that dip rather than assuming smooth delivery. It is also why the 15% daily change cap exists: a forecast is worthless against a budget that moves every morning.

Demo data

Behind one answer

Answer
Meta. Its customers are credited with €14,800 more collected 90-day revenue than Google's on 15% less spend.
Evidence
311 customers, 90-day window, invoiced value from the finance system, last-non-direct attribution.
Business impact
Moving €1,200/day would add an estimated 31 customers this month.
Confidence
High — 14 days above the minimum evidence threshold.
Alternative explanation
Meta's audience skews to annual plans. Some of the gap is plan mix rather than channel quality.
Recommended action
Move €1,200/day from Google Prospecting to Meta Prospecting.
Constitution check
Passes — within the €1,500/day movement limit.

One goal. Every channel. Real business truth.

Connect your ad accounts, say what a good customer is worth, and let the Growth Brain work the objective inside limits you set.

No ad spend commitment · Connect one channel to start · [email protected]