Budget Simulator
See next quarter's mix before you fund it.
Shift budget between channels and read modeled revenue off the response curves your model fit. Bayesin proposes the split. You decide.
Same budget, a better mix
Share of a fixed weekly budget
| Paid search | 34% to 42% |
|---|---|
| Paid social | 30% to 21% |
| Microsoft Ads | 24% to 27% |
| 12% to 10% |
Modeled weekly revenue, total spend unchanged
With its 80% credible range
| Current mix | $401k ($392k to $410k) |
|---|---|
| Proposed mix | $418k ($407k to $429k) |
How it works
Start from the fit. Move the share. Read the gap.
01 Start from the fit
The simulator opens on today's allocation and the response curves from your latest fit.
02 Move the share
Drag a channel up or down, or lock it. The total holds unless you change it.
03 Read the gap
Modeled revenue under your proposal, against the mix you run now, with its range.
How to read it
A fair comparison, every time.
The simulator is your model with a handle on it. It adds no assumptions of its own.
Fixed total
Every proposal is judged at the same spend, so a bigger budget is never mistaken for a better mix.
The model's own curves
Readouts come off the response curves the model fit. If a channel saturates, the simulator shows it flattening.
A proposed split
Optimize ranks channels by the return on their next dollar and proposes a split within the bounds and locks you set.
The edge of the data
Push past the spend you have tried and the range widens to say so. That is where the next test belongs.
Scenarios
Bring the plan to the room.
Save an allocation, compare it with the others, and send the link to whoever owns the budget. Export it when finance wants the rows.
Modeled revenue
| Current mix | $401k |
|---|---|
| Proposed by Bayesin | $418k |
| Search capped at 40% | $414k |
Also in the product
One model behind every plan.
What each channel returned, and what its next dollar will.
See media mixNext quarter's revenue from the same model, with the range in plain sight.
See forecastPlan a funnel from your own assumptions before a dollar goes out.
See the campaign simulator
Fund the mix your model can explain.
Early access is by application. We set up each team ourselves.