Tacnode™
← The lineup

The Overnight Truth

The nightly load finished at six. Everyone has been reading it since.

Nightly load finishes. Dashboard shows the refund rate.

What’s true
1.2%
The warehouse
1.2%
reads
Decision
Hold the fraud threshold where it is?
waiting

Illustrative figures.

Where this shows up

Fraud and risk operationsRules get retuned from a picture that’s hours old.
Intraday treasuryCash positions change by the hour, and the report shows one of them.
Marketplace operationsSupply and demand read off last night’s load.
Trading desksYesterday’s exposure, during today’s session.

Why the warehouse gets it wrong

06:0012:0018:00what the report can see after 06:00: nothing newreport says 1.2%actual 5.1%14:40 decision

A warehouse is built for questions about periods that have closed, and it answers them well. Data arrives when the load runs, and between runs it gets older. For a quarterly review that hardly matters. For a threshold someone is setting this afternoon, it means working from this morning.

By the time the decision ran, its context was outdated. We call that a context gap.

Put this afternoon’s numbers in front of whoever sets the threshold.

06:0012:0018:0014:40 decision reads 5.1%current to the secondWAREHOUSEhistory via Iceberg

Tacnode keeps the refund rate current to the second, with its history right beside it for comparison. The same goes for exposure, balances and anything else your teams check during the day. Your warehouse and analysts read those tables through Iceberg, so there’s no second copy to keep in step.

What comes inSource data arrives from the systems you run today, via change streams.
What staysThe warehouse still does the long-range analysis.
What the decision getsWhoever sets the threshold sees refunds from the last few minutes.
If an engineer asks how

Operational aggregates are computed on demand against current data, or kept as views refreshed on a schedule you set. Iceberg lets dbt, Spark, Trino or Snowflake external tables read the same storage.

Questions to put to your team

  1. Which calls get made off a dashboard that refreshes once a day?
  2. If that number were current to the minute, would anyone decide differently?
  3. How many reports are built on a copy somebody has to reconcile?

Is your morning report telling this lie?

Bring it to a demo, and we’ll trace what it reads and how old that data is by the time it runs.

Book a demo →
The Averaged Lie →