Why rules?
When tagging scales or even when it doesn't - Quality can drift
In the early stages of a tagging workflow, things feel controlled. Whether it's a single analyst or a small team, the logic is clear and the volume is manageable. But over time - as more videos are tagged, or even within a long tagging session - subtle inconsistencies begin to appear.
A single user may rush through segments and miss edge cases. In a team, different taggers may interpret the same situation differently. Timing errors, unusual durations, or inconsistent tagging patterns start creeping in. These are not major mistakes individually, but they accumulate quietly. The real challenge is not that errors happen - it's that they are easy to overlook during the tagging process.
A single user may rush through segments and miss edge cases. In a team, different taggers may interpret the same situation differently. Timing errors, unusual durations, or inconsistent tagging patterns start creeping in. These are not major mistakes individually, but they accumulate quietly. The real challenge is not that errors happen - it's that they are easy to overlook during the tagging process.
The downstream impact is bigger than it looks
These small inconsistencies rarely stay small. They begin to affect everything that depends on the data. Analytics start showing irregular patterns. Reports become harder to trust. If the data is used for AI workflows, the model learns from incorrect signals.
By the time these issues are noticed, the cost of fixing them is significantly higher. It often requires going back, rechecking events, and correcting data manually. This shifts the team’s effort from generating insights to cleaning up mistakes - which is not where the value should be.
By the time these issues are noticed, the cost of fixing them is significantly higher. It often requires going back, rechecking events, and correcting data manually. This shifts the team’s effort from generating insights to cleaning up mistakes - which is not where the value should be.
Why manual checking isn't enough - even for one person
A common assumption is that careful tagging or manual review will solve this. In reality, even a single experienced user cannot consistently catch every issue, especially during long or repetitive sessions. Fatigue, speed, and attention shifts all play a role.
In multi-tagger workflows, the problem becomes even more pronounced due to variation in judgment. In both cases, relying only on human checking introduces inconsistency and makes the process harder to scale. What is needed is a system that can apply the same logic, every time, without depending on memory or attention.
In multi-tagger workflows, the problem becomes even more pronounced due to variation in judgment. In both cases, relying only on human checking introduces inconsistency and makes the process harder to scale. What is needed is a system that can apply the same logic, every time, without depending on memory or attention.
Rules in SPAN: a built-in safety net for every workflow
Rules in SPAN templates provide that consistency. By defining simple logical checks - such as validating durations, ensuring correct time sequences, or identifying unusual tagging patterns - SPAN can automatically flag events that need attention.
This works equally well for a single user or a large team. For individuals, it acts as a real-time safety net, helping catch mistakes that might otherwise go unnoticed. For teams, it ensures that every event is evaluated against the same standard, reducing variation across taggers.
Instead of relying entirely on manual effort, rules introduce a structured quality layer that runs continuously in the background. This allows workflows to remain fast, while ensuring that data remains reliable and consistent.
Note: These rules in SPAN are programmed based on the json-rules-engine syntax as can be seen in this page, so that you can extend it at any time in the future from the “Rules” tab in Template Management page. Even non-coders can modify these simplistic JSONs and add rules as per their requirement.
Important note (in SPAN)
In SPAN, Rules are specific to a template. To learn how to create a template in SPAN, refer to this article
Related articles
Related Articles
Access rules page
In this article, you will learn how to access the Rules page of a template. Before you begin To learn how to create and open a template in SPAN sports performance analysis platform, refer to this article. Access rules page After opening the template ...Use multiple rules with all (AND logic across rules)
In this article, you will find the rules document for using multiple rules with all (AND logic across rules). Before you begin To learn how to access the rules page, refer to this article. Use multiple rules with all (AND logic across rules) Note: ...Multiple rules for a template
In this article, you will learn how to write multiple rules for a template. Before you begin To learn how to access the Rules page and view some rule examples for a template in SPAN sports performance analysis platform, refer to this article. ...Save rules for a template
In this article, you will learn how to save rules for a template. Before you begin To learn how to access the Rules page and view some rule examples for a template in SPAN sports performance analysis platform, refer to this article. Save rules for a ...Flagged events - rules in action
Flagged events are those that do not adhere to the rules written for a template. In this article, you will learn how to check flagged events once multiple rules are set for a template. Before you begin To learn how to access the Rules page and view ...