Skip to content
BENIBIX Learn

Data Freshness

Also known as Timeliness (Data Quality), Data Timeliness, Data Latency

How up to date data is, measured as the time since its last load.

Think of it like

It is like the date on a milk carton. Old milk may look fine, but it is not what you want.

Example

A dashboard shows yesterday's sales as zero. The real problem is that the table is 30 hours old, because last night's load failed.

Why it matters

Stale data does not look broken. People act on old numbers without knowing it.

Definition

Data freshness is how up to date the data is. It is measured as the time since the newest record arrived. A table loaded each hour should never be more than about an hour old. When a load silently stops, every number still looks normal, just old. That is why freshness is one of the first things teams watch.

Example code

sql
SELECT MAX(loaded_at) AS last_load,
       CURRENT_TIMESTAMP - MAX(loaded_at) AS age
FROM orders;
The age of the newest row. If it is more than a few hours on an hourly table, a load has stopped.
Freshness is the fourth tile. Nothing in the table is wrong, it is just three days old.

Common mistake

Showing no "last updated" time on a dashboard. Readers then cannot tell stale from real.

Key takeaways

  1. 1
    Time since last load. The age of the newest row.
  2. 2
    Set an expectation. Know how old is too old.
  3. 3
    Fails quietly. Old numbers still look normal.