Data Tags vs Data Metrics: Clear Differences and How to Use Them

1. What Is a Data Tag?

Data tags describe objects and consist of a tag name and tag value attached to a specific target, turning raw data into clear, distinguishable information. Common business objects such as users, products, channels, and marketing activities can be tagged.

Example: User A is male, 28 years old, lives in Beijing, logged in 5 times in the last 30 days, and has never made a purchase. Each of these attributes is a data tag.

Key points to remember:

Fact tags: Describe objective, relatively immutable facts (e.g., gender, birthplace, registration time) directly extracted from raw data.

Rule tags: Generated by applying predefined rules to data (e.g., “hot‑selling product” based on top‑10 daily sales).

Model tags: Predicted by algorithms (e.g., “high phone‑upgrade intent” or “potential home‑buyer”), which are dynamic and probabilistic.

2. What Is a Data Metric?

Metrics measure the level of something overall or in a specific aspect. They are quantitative values that require a clear calculation method.

Continuing the user example, the tag “logged in 5 times in the last 30 days” is descriptive, whereas the metric “average login count of all users in the last 30 days” quantifies that behavior.

In e‑commerce, “sales of 1 M in the last 7 days” or “product X sales of 500 K” are metrics.

Three essential components of a complete metric:

Dimension: The perspective (e.g., “Beijing region”, “users born after 1990”).

Aggregation method: How to compute (sum, average, ratio, etc.).

Unit: The measurement unit (currency, count, percentage).

Metrics must be numeric and have an explicit formula; otherwise, they are not true metrics (e.g., sales = unit price × quantity, login rate = logged‑in users ÷ total users × 100%).

3. Aligning Definitions and Ensuring Consistency

Inconsistent metric definitions lead to misleading decisions. For example, daily active users (DAU) may be defined as “any login during the day” or “login time > 10 minutes”, producing different counts.

Therefore, the first step in data work is to standardize metric definitions, calculation formulas, and scope.

4. Converting Between Tags and Metrics

Metrics can be derived from tags: “high‑net‑worth client migration rate” uses the tag “high‑net‑worth client” plus a calculation.

Tags can be derived from metrics: the tag “private‑bank client” may be defined by the metric AUM ≥ 5 M.

5. Practical Application Workflow

When analyzing problems, first build a tag system, then calculate metrics to pinpoint issues.

Example: Conversion rate dropped from 15 % to 10 %. By segmenting users with fact tags (new vs. old) and rule tags (high‑spending vs. low‑spending), the analysis revealed that new users’ conversion fell from 20 % to 8 %, while old users remained stable. This insight directs optimization efforts toward new‑user strategies.

Another pattern: first define tags to identify target groups, then use metrics (conversion rate, paying users, average order value) to evaluate the effectiveness of campaigns.

Overall, mastering the distinction, proper construction, and combined use of data tags and metrics resolves most data‑related challenges.