Beyond the dashboardAI & Measurement · Independent guides
TrackingByAI Journal
← All guidesAI and model evaluation / Guide 13

Training data and live data serve different roles

Performance on familiar examples does not guarantee performance on new cases. A useful evaluation keeps those roles distinct.

A female scientist conducting research in a well-equipped laboratory, focusing on chemical analysis.

First, define the question

Training data helps fit a model. Validation and test sets probe choices made during development, while live data shows behavior after deployment. If the same person, account, or time period leaks across these sets, the evaluation can look better than real use.

Model performance is meaningful only in relation to a task and a population. Start by writing the decision the output will inform, the examples used for evaluation, and the kinds of error that matter. A single headline score rarely describes a complete workflow. A reviewer needs to see where the model performs well and where its assumptions may fail.

The boundary of the measure

Development data and future cases have different roles. A model may learn patterns in a training set and still struggle when the service, language, or user group changes. Keep evaluation examples separate and record when they were collected. Test periods after a major change can be more revealing than another random slice of familiar material.

Split data by the unit that can leak information. If many rows come from the same person or account, a random row split may place nearly identical cases in training and test sets. A later time period or group-level holdout can give a more realistic view of future use.

Inspect examples near the edge of the rule, not only the easiest successes. Ambiguous inputs, missing fields, unusual formats, and conflicting evidence expose how a system handles uncertainty. If the workflow requires a person to decide, show that person the source information and a reason to pause. A confident-looking number should not hide a weak underlying case.

Try a small example

A model trained on last year's support tickets may perform well on a random sample from that year but poorly after the service changes. A later holdout period can reveal the gap.

An automated label is a claim made by a system under a particular version. Record the model, threshold, input definition, and date. If any of those changes, compare like with like before describing a trend. A rise in alerts may reflect a new threshold rather than a rise in the underlying behavior.

Before drawing a conclusion

Document how examples were collected, cleaned, and split. A high score without that history leaves readers unable to judge what the model actually learned.

Evaluation should reflect the live mix, including cases the system declines to process. A model may be accurate on clean, fully labeled inputs while the actual workflow contains partial records. Report the coverage as well as the score on the cases that were scored.

Good evaluation asks about consequences as well as scores. A false alarm can consume time or unfairly interrupt a person; a missed case can allow a problem to continue. The balance depends on the use. Explain the trade-off plainly and give a route for correction when an automated result has a material effect.

A review path for this question

Human review needs real capacity. A reviewer who sees hundreds of cases without context cannot reliably challenge an output. Define which cases are escalated, what records are available, and how disagreements are recorded. Review overrides and appeals as evidence about both the model and the process surrounding it.

Three things to record

  • The exact definition or decision behind “Training data and live data serve different roles”.
  • The source and date for the information being compared.
  • At least one limitation or alternative explanation.

Keep the explanation proportional to the evidence. Say which finding came from a source record, which came from a model, and which is an analyst's judgment. If the answer may change with new data, name the trigger for another review. That makes uncertainty part of responsible operation rather than an afterthought.

Use the guide as a starting point for reviewing a real service. Check the live configuration, current provider documentation, and applicable requirements before making a consequential change. If a page here is unclear or outdated, send its URL and the disputed passage through our contact route.

Continue reading

More perspectives