What is data freshness?
The simplest way to think about it is as a gap in time. If a campaign spent money at 9 a.m. and the dashboard shows it at 10 a.m., the data is an hour old. If the feed stopped updating on Friday, Monday's dashboard is showing Friday.
Freshness is a property of each source. Different platforms make data available on different schedules, and any connection can fall behind.
How do you measure data freshness?
| Measure | What it tells you | Example |
|---|---|---|
| Last successful update | When the source last delivered data | Meta feed updated 40 minutes ago |
| Lag behind activity | How far the data trails real events | Spend data runs about three hours behind |
| Expected schedule | Whether an update is overdue | Daily feed has not arrived by 9 a.m. |
| Completeness | Whether the latest period is fully loaded | Yesterday shows only half the usual rows |
What is an example of stale data?
A no-spend alert that wasn't about spend
A dashboard shows a campaign spent nothing yesterday. The media buyer is about to rebuild the campaign.
A quick freshness check shows the platform's feed stopped updating the day before. The campaign was spending normally. The data just hadn't arrived. Fixing the connection, not the campaign, was the right move.
Data freshness vs. data accuracy vs. data completeness
| Freshness | Accuracy | Completeness | |
|---|---|---|---|
| Question | Is it current? | Is it correct? | Is all of it there? |
| Typical failure | Feed stopped or delayed | Wrong mapping or definition | Missing rows, accounts, or fields |
| How it shows up | Numbers look normal but are old | Numbers disagree with other sources | Totals are lower than expected |
Why does data freshness matter?
Decisions made on stale data look reasonable at the time and can turn out wrong once the real numbers arrive. Freshness problems are especially risky because stale numbers rarely look broken. Freshness is set by the marketing ETL pipelines that load the data, and a feed that stops refreshing is exactly the kind of problem anomaly detection should flag.
- Better calls. Decisions reflect this week, not last month.
- Fewer false alarms. A missing day of spend reads as a feed delay, not a stopped campaign, so nobody rebuilds a campaign that was working.
- Fair comparisons. Cross-channel views are only valid when every source is equally current.
- Trustworthy alerts and AI. Monitoring and AI answers are only as current as the data they read.
How does Alli do this?
Alli by PMG's Datasources captures each data point as soon as the source makes it available, replacing delayed batch reporting with a steady, always-on flow. Connection status and feed health for every source are visible in one place, so teams can tell a stale feed from a real performance change.
Alli Alerts runs on that same data, with monitoring for feed failures and no-spend issues delivered into Slack, so a stalled source is caught before it misleads a report.
Explore Alli DatasourcesData freshness FAQ
What is stale data?
Data that hasn't been updated recently enough to reflect current activity, often because a feed was delayed, failed, or runs on a slow schedule.
Is data freshness the same as real-time data?
No. Real-time is one end of the freshness scale. Many decisions only need data that is a few hours or a day old, as long as everyone knows how current it is.
How do you monitor data freshness?
Track when each source last updated, compare it with the expected schedule, and alert when a feed falls behind or arrives incomplete.
Why do different marketing platforms have different freshness?
Each platform makes data available on its own schedule, and some revise recent data as conversions are attributed, so the latest period can change after it first loads.
Last updated ‹ Back to the glossary