Every channel counts differently
Meta counts conversions on its model, Google on its own, and the decision to move money is taken from two reports that never reconcile plus a hunch.
Set the total, the goal and the platforms, and the module lays out the current and the recommended split per platform, estimates the effect and explains the reasoning in words. The result is saved as a draft — nothing from it reaches your ad accounts.
Meta counts conversions on its model, Google on its own, and the decision to move money is taken from two reports that never reconcile plus a hunch.
The money is shifted, two weeks later results are worse, and there is nothing to roll back to: nobody wrote down the original split, let alone the reasoning.
Any optimisation easily becomes a neat table nobody reconciled with the ad account. Until you can see which part is measured and which is assumed, it is not trustworthy.
Enter the total, pick the goal — ROAS, CPA, conversions or clicks — and select the platforms: Meta, Google, TikTok, LinkedIn, Reddit, Telegram.
The module shows the current and the recommended amount per platform in money, an estimated effect, and a written argument tied to the goal you picked.
The result is stored as a plan marked draft and approval required. Pushing the change into the ad accounts is not performed by this module.
Meta, Google, TikTok, LinkedIn, Reddit and Telegram land in a single table instead of six tabs with different reports and currencies.
ROAS, CPA, conversions or clicks: the goal you choose decides which way the calculation pushes money and what counts as a win.
Each platform shows its current and recommended amount in money, so the price of the decision is visible without mental arithmetic.
The split comes with an explanation of why the money should move this way — something you can argue with rather than simply trust.
Recommended amounts and expected improvement carry an AI-estimate badge and a disclaimer, and that label is what blocks any automatic apply.
Drafts keep their budget and goal, the history tab shows budget changes as before and after, and plans stay attached to a project and a client.
No. The module runs in preview: you can calculate a recommendation and save it as a draft, but nothing from it is written to the ad accounts. Even a prepared apply requires a change set, an approval, a rollback plan and confirmation from the platform side.
Meta Ads, Google Ads, TikTok Ads, LinkedIn Ads, Reddit Ads and Telegram Ads — you tick the ones you need before running it. The plan is saved together with that platform set, the total and the goal, so a run can be repeated and compared.
The recommended split and the expected improvement are a model's estimate from the parameters you entered, and the interface labels them as an AI estimate with a disclaimer. Check them against real account data — that label is exactly what blocks automatic applying.
The plan sits as a draft with its budget, goal and platform list, tagged as approval required. You can come back to it, compare it with actual results and make the change in the ad account yourself: the module stays the calculation layer, not the executor.
Yes. Plans are attached to projects, so calculations for different clients never mix. A separate history tab shows, per platform and campaign, what the budget was and what it became — a convenient basis for a client report.
The calculation changes nothing in your ad accounts, and the plan stays a draft until you decide otherwise. The first run takes a couple of minutes.