Nobody re-reads every campaign every morning, so ACoS creep gets discovered at month-end, a full report cycle too late. Performance rules read yesterday's settled numbers every morning and speak up, or act, the day something crosses your line.
The data was there every single morning: yesterday's ACoS, sitting on the dashboard. But it's nobody's job to re-read 40 campaigns before coffee, so the creep compounds quietly until the monthly export makes it official.
How long a creeping ACoS problem typically lives before a monthly report surfaces it. Every one of those mornings, the number was right there.
What a mid-size account runs. Nobody re-reads all of them every morning, so the campaign that slipped waits for whoever opens the export.
Intraday numbers whipsaw and Amazon's attribution lags. Reacting in real time punishes campaigns for sales that simply haven't attributed yet.
One OAuth handshake through the official Amazon Ads API. Off Hours reads your campaigns and their daily performance. Nothing is guessed.
Yesterday's ACoS, yesterday's spend, or spend measured against the campaign's own baseline. Your numbers, your line, not a preset someone else picked.
Recommendation, budget change, or pause: graduated, and entirely your choice. A rule can advise for weeks without ever touching the account.
Settled numbers, checked daily, acted on the morning after, with the enforcement engine running every 15 minutes underneath. No whipsaw, no phantom problems.
A week of settled numbers, read the way the rule reads them. Thursday crossed the line, and Thursday got flagged the morning after with a recommendation, not a kill.
A $40 morning with no sales looks catastrophic at 10am and fine by dinner. Amazon attributes sales to clicks hours or days later. Off Hours reads the settled numbers on purpose. "No real-time noise" is the feature.
Daily granularity is a decision, not a shortcut. Hourly "performance" automation fires on phantom problems: sales that haven't attributed yet. Off Hours waits for the numbers to settle, then acts.
Spend-spike rules compare a campaign against its own recent normal: "50% above baseline," not "over $200." Growth isn't punished for being growth, and drift isn't missed because a fixed number sat too high.
Recommendation first, budget cut second, pause last. You pick how much rope each rule gets. Recommendations land in the dashboard's Recommendations tab, and in Slack if you want them there.
Every pause and every change is logged and reversible. Turn a rule off and your campaigns go back exactly as it found them. Your manual changes always win.
Performance rules act on results. Dayparting, budget, and event rules act on schedule: when campaigns run, how much they spend, what's special this week. Same account, no conflicts.
Free spend alerts watch every connected account for spikes, drops, and drift. Zero setup, on from the moment you connect. Rules act on your thresholds; alerts just tell you.
How spend alerts work →Start with spend alerts, the zero-setup version that just tells you. Every connected account gets always-on watching for spikes, drops, and drift, free, before you write a single rule.
For most accounts, "performance monitoring" means an export, an average, and whoever remembers to read it. Averages hide creep, and discovery arrives after the budget's already spent.
Performance rules watch your campaigns' daily metrics (yesterday's ACoS, yesterday's spend, and spend spikes against the campaign's own baseline) and act when a threshold you set is crossed. The action is your choice: send a recommendation, adjust the budget, or pause the campaign. Every action is logged and reversible.
Because intraday numbers lie. A campaign that spends $40 by mid-morning with no sales looks catastrophic at 10am and fine by dinner, and Amazon's attribution lags, so sales credit clicks hours or days after they happen. Automation that reacts hourly ends up punishing campaigns for sales that haven't attributed yet. Off Hours reads settled daily numbers and acts the morning after: fast enough to matter, slow enough to be right.
Three, graduated: send a recommendation to your dashboard's Recommendations tab (and to Slack, if connected), adjust the campaign's budget, or pause the campaign. Recommendations are first-class. A rule can watch and advise without ever touching the account. Whatever it does is logged and reversible, and your manual changes always win.
A rule that compares a campaign's spend against its own recent baseline instead of a fixed dollar amount. "Spend is 50% above normal" catches drift on a $30-a-day campaign and a $500-a-day campaign with the same rule. There is no magic number to guess, and growth isn't punished for being growth.
Yes, that's the intended setup. Dayparting, budget, and event rules act on a schedule: they decide when campaigns run and how much they can spend. Performance rules watch the results and speak up when something crosses your line. They read the same account and don't conflict by design, and deactivating any rule restores what it changed.
Set your first threshold in about 90 seconds. 14 days free, no credit card required.
Start free trial→