Alerts and notifications
Twenty-six events, four groups, and a switch per channel deciding which ones reach you and where.
The account tells you when it changes
Twenty-six things worth knowing about, from a campaign spending its budget by noon to figures no longer arriving from Amazon. You pick which ones reach you, and where each one lands.
The event catalog
Events fall into four groups. Each row is a switch per channel, so you decide not just whether you hear about something but where.
Real-time
Hourly, from the Amazon Marketing Stream feed. These only appear once the feed is on and your profile's region supports it.
| Event | What it means |
|---|---|
| Budget exhausted | A campaign spent its full daily budget and stopped serving |
| Budget running out | A campaign crossed your warning threshold, with the projected exhaustion hour |
| Spend spike | Hourly spend above your multiplier against the trailing same-hour baseline |
| Sales drop | Hourly sales fell below your threshold against baseline, with spend still material |
| Spend with no sales | Spend passed your floor in an hour that returned no conversions |
| Feed problem | The real-time feed stopped arriving, or its subscription is no longer active |
| Feed incomplete | Real-time totals disagree with the daily report beyond tolerance |
| Campaign changes | An hourly digest of campaigns created, paused, or changed |
Daily checks
| Event | What it means |
|---|---|
| Buy box lost | An advertised ASIN lost the buy box on consecutive days while still taking ad spend |
| Buy box restored | It is winning the buy box again |
| Campaign went dark | A campaign that had been spending every day stopped delivering entirely |
| Campaign delivering again | It is spending again |
| Running out of stock | An ASIN that sells is out of stock at Amazon, or has under two weeks of cover with nothing on the way |
| Back in stock | It has three weeks of stock again |
Data pipeline
| Event | What it means |
|---|---|
| Data synced | The daily Amazon ingestion finished and your dashboard is current |
| Backfill finished | A historical backfill you started has finished loading |
| Retail data missing | A day of Amazon retail sales the warehouse should have is still missing after the nightly re-buy |
| Retail data complete again | Missing retail days have been re-ingested |
| Search-query data missing | A week of Amazon search-query data the warehouse should have is still missing |
| Search-query data complete again | Missing search-query weeks have arrived |
Account
| Event | What it means |
|---|---|
| Payment method declined | Amazon has no valid payment method, so campaigns have stopped serving |
| Payment method restored | Billing is valid again and campaigns can serve |
| Connection broken | Amazon rejected a request; data collection stops until you reconnect |
| Sync paused | Amazon access stayed revoked for a week, so syncing for that connection was paused until you reconnect. The account owner also gets an email |
| Reconnect needed soon | Your Amazon authorisation expires shortly |
| Goal breached | A KPI you set a target for moved outside its bound |
Channels
In app
The bell in the dashboard header. On by default for nearly every event, because it costs you nothing until you look at it.
Slack
Paste an incoming webhook URL and pick which events route to it. Slack defaults on only for the events worth interrupting someone: budget exhausted, spend spike, spend with no sales, buy box lost and restored, payment declined and restored, connection broken, sync paused, and the daily sync.
The thresholds are yours
Real-time alerts fire against numbers you set, not ours:
| Setting | Default | Range |
|---|---|---|
| Warn at budget used | 80% | above 0, up to 100 |
| Spend spike multiplier | 3× the trailing same-hour mean | above 1, up to 100 |
| Sales drop | 50% of normal | above 0, up to 100 |
| Zero-sales spend floor | Off | 0 disables it |
Three things worth knowing
Hourly, not overnight
The real-time group rides Amazon Marketing Stream, so budget and spend alerts fire inside the hour they happen rather than in tomorrow's report.
One problem is one message
Delivery is deduplicated on the event, so a campaign that stays dark all afternoon does not become six alerts.
Noisy events ship off
Campaign changes is the noisiest event here by an order of magnitude on a large account, so it ships off on every channel including in-app. Turn it on deliberately.
- ✓Route the critical and high events to Slack and leave the rest in the bell — a muted Slack channel hides the ones that mattered.
- ✓Set the zero-sales spend floor to a number that would actually annoy you if you spent it for nothing.
- ✓Turn the real-time feed on before expecting hourly alerts; without it those rows do not appear at all.
- ✗Turn everything on to Slack on day one — the failure modes are asymmetric, and someone buried in noise stops reading.
- ✗Treat 'Campaign went dark' as always bad; a campaign you paused yesterday will trigger it.