Skip to content

What Arri actually does with your ads.

Six kinds of change, each one diagnosed, previewed, approved by you and measured afterwards. This page is the full list — nothing here is a roadmap.

  1. An ad or ad set keeps spending below its break-even return with enough data to be sure.

See how it works

Meta will never tell you to spend on Google.

Each platform only optimises inside its own walls. Arri compares both every night and moves budget to the one earning more, with your approval.

Example

Both budgets change as one. If Google's fails, Meta's is put back.

  • Compared every night

    Thirty days of return on each platform, side by side, in every workspace with Meta and Google connected.

  • Moved in steps

    When both platforms make money, 5% of the weaker one's daily budget moves across. When one is losing money, 20%.

  • Both sides or neither

    The two budget changes run as one action. If the second fails, the first is rolled back before you hear about it.

Every recommendation shows its work.

Before you approve anything you see the entity, the diagnosis, the evidence behind it, the expected outcome and how to undo it. Arri never asks for trust it has not earned on the page.

Illustrative photo: a café owner in an orange shirt holding a cup of coffee at a sunny window table.
ArriExample
Retargeting — WarmAd set · Meta Ads · Budget · High impact

Your best ad set is underspending — it uses 40% of its daily budget because the campaign-level cap holds it back.

Example — not a live recommendation

  1. Proposed change

    Raise the ad set's daily cap so it can spend its share.

  2. Evidence

    'Retargeting — Warm' hit the cap on 6 of the last 7 days at 4.6× ROAS.

    The 2 lowest-ROAS ad sets in the same campaign spent €410 for 3 orders.

  3. Outlook

    Expected: +25–35% revenue within a week as spend shifts to what's working.

  4. If it does not work

    One click restores the previous budget. Arri keeps measuring for 72 hours either way.

What Arri looks for.

Every campaign, ad set and ad is classified nightly against seven days of data, with a three-day window for early movement. These are the diagnoses it can reach.

Creative fatigue
Frequency climbs while click-through and conversion rates fall.
Audience saturation
Reach flattens and cost per result rises inside the same audience.
Delivery issue
Spend stalls or swings for reasons outside performance.
Conversion issue
Clicks hold but purchases or leads drop — usually the page or tracking.
Budget constraint
A profitable entity hits its cap while weaker siblings keep spending.
Learning phase
Too recent or too changed to judge. Arri waits instead of guessing.
Account structure
Overlapping audiences or too many ad sets for the budget.
Scaling success
More spend, same or better efficiency. Worth another step.
Efficiency gain
Cost per result improved without a volume trade-off.

Then it measures itself.

An approved change is not the end of the job. Arri snapshots the entity after 24 and 72 hours and reconciles the measured result against what it promised in the preview.

  1. 04:00

    Nightly sync

    Meta and Google data lands, seven-day and three-day windows recomputed.

  2. Morning

    Brief and diagnosis

    Each entity answers five questions: what happened, what it means, why, what to do, what to expect.

  3. You

    Preview and approve

    Nothing changes on the platform until you say so. Dismissals are remembered.

  4. Applied

    Hands-off cooldown

    The changed entity is left alone while its learning phase resettles — whether you or Arri changed it.

  5. +24h · +72h

    Impact snapshots

    Spend, results, cost per result and return captured before and after, per day.

  6. Closed

    Measured against expected

    The gap between promise and result is written to the action's record and to your savings ledger.

What Arri will not do.

Every write goes through the same gate. These are the lines Arri does not cross, and the mechanism that enforces each one.

  • Change anything without your approval

    Connecting gives Arri permission to make changes, but it makes none until you approve one. Every change is a single previewed action tied to your account.

  • Scale a loser

    Break-even return and contribution margin gate every scale-up. Negative contribution never gets more budget.

  • Touch something that just changed

    A hands-off cooldown follows any edit, yours or Arri's, so the learning phase can settle.

  • Break a campaign budget

    Shared-budget ad sets are steered with caps and floors, never converted to fixed budgets.

  • Leave you with half a change

    Cross-platform shifts run as two legs. If the second fails, the first is rolled back before you hear about it.

Ask, and get the evidence.

Questions are answered from your synced data, and every answer names the entities and dates it used. Follow-ups keep the context.

Example conversation

Connected where your ads live.

Connect in two minutes. Arri gets permission to make changes and uses it only for the ones you approve.

  • Meta Ads

    Live

    Campaigns, ad sets, ads and creatives, with changelog reconciliation so manual edits are seen.

  • Google Ads

    Live

    Campaigns and bidding strategies through the same brief and the same approval flow.

  • Workspaces

    Included

    One brand, both platforms. A single budget shift can move spend from one to the other.

  • Since your last visit

    Included

    Arri remembers when you were last here and leads with what changed — new issues, resolved ones and metric moves.

Get your first brief tomorrow morning.

Start free

Connect one account tonight. Free while we're in early access — no card, disconnect anytime.

What Arri actually does with your ads. | Arri