TG activity data vs gender screening: Screen for activity first or screen for gender first? A guide to optimal sequencing for cost and efficiency

In Telegram’s accurate customer acquisition, tg active data and gender filtering are the two most commonly used screen size dimensions. But when planning tasks, many teams will struggle with a question: should we screen for activity first, or for gender? Depending on the order, the final consumed balance may vary by 20%-50%. This article will start from the itemized billing mechanism of KK-DATA, compare the cost difference, data quality and applicable scenarios of the two sequences, and help you find the optimal execution path.

Why does the screening order affect TG customer acquisition cost?

The key to understanding the impact of order on costs lies in KK-DATA’s billing method: Each screening task is deducted based on the number of test numbers, and the unit price is different for different test types. Both TG activity detection and TG gender detection include the basic step of “activating detection”. If you do A first and then B, the second task will repeatedly detect the “activation” of the same batch of numbers, resulting in additional costs. However, the more core cost difference comes from the detection base in the second step - dimensions with high screening rates are done first (for example, activity is usually only 30%-50%), and the second step only needs to detect the remaining valid numbers, and the total detection volume is greatly reduced.

Under billing by item, each inspection is a cost

KK-DATA does not have a subscription package and adopts the balance recharge + item deduction model. The estimated cost will be displayed before submitting the task and will be deducted from the balance after the task is completed. Suppose you do a TG activity test (including activation test), and then perform a TG gender test (including activation test) on the same number, the “activated” status of the same number will be tested twice, and a fee will be deducted each time. Although the unit price of the two tests may be different, there is indeed a waste of cost in the duplication part.

Differences between activity and gender detection

• TG activity detection: First detect whether the number is activated for TG, then determine the number’s activity within the specified time window (such as 7 days/30 days/60 days), and finally return to the active status.
• TG Gender Detection: First check whether the number is activated for TG, then match the gender, age, avatar and other fields based on the tgid, and finally return the gender determination.

Both include “opening detection”, but the subsequent dimensions are different. If you do activation first and find that only 40% of the numbers are active, then the second step of gender detection only needs to target these 40% of the numbers (about 400), and the total number of detections is 1000 (first step) + 400 (second step) = 1400 times. If you do the gender first, the gender determination may cover all numbers (assuming the gender ratio is not filtered), and then perform active detection on all 1,000 numbers. The total number of detections is 1,000 + 1,000 = 2,000 times (actual active detection cannot be skipped, so it is still 1,000 times). It can be seen that ** be active first, the base will become smaller in the second step, and the total cost will be lower **.

---

Scenario 1: Screen activity first, then gender (recommended order)

Applicable scenarios: Limited budget, pursuit of high conversion rate, and the gender ratio of the target population is not extreme (for example, 30%-60% are men).

Operating steps:

1. Use the KK-DATA “TG Active Detection” task, set the active window (recommended 7 days or 30 days), and detect the original number pool (for example, 1000).
2. Export active numbers (assuming the result is 400 active). These numbers have been activated and have been active recently, with high conversion potential.
3. Use these 400 active numbers as sources, submit the “TG Gender Detection” task, and obtain the gender and age fields.

Cost Analysis: The first step is to detect 1,000 items, the second step is to detect 400 items, and the total number of tests is about 1,400 times (the actual second step still includes activation detection, but the base is small, and the total cost is significantly lower than testing the gender and then testing the activity of all numbers). More importantly, you only invest in gender detection fees for active users who are actually valuable to reach, thus avoiding wasting your balance on inactive zombie accounts.

Data Quality: Exclude inactive accounts first, and the subsequent gender field is aimed at users who are actually likely to reply, with higher accuracy.

---

Scenario 2: Screen for gender first, then screen for activity

Applicable scenarios: The gender of the target group is extremely scarce (for example, you are only looking for male users on TG, and the proportion of males is less than 10%), and you do not have strict requirements on activity level (you only need to be contactable, and you do not need to be active recently).

Operating steps:

1. Submit the “TG Gender Detection” task and screen out males from the original number pool (assuming that 50 males are found out of 1,000).
2. Export these 50 male numbers, and then submit the “TG Activity Detection” task to check whether they are active.

Cost Analysis: The first step is to detect 1,000 items, the second step is to detect 50 items, and the total number of tests is about 1,050. Since the base of the second step is extremely small, even if repeated testing is enabled, the total cost may be lower than active first and then gender (if the proportion of males is extremely low). But if the proportion of men is high (such as 30%), then the second step is to detect 300 items, with a total of 1,300 times. It is not as good as first active and then gender (400+400=800 times? Note here: If active first and then gender, the first step is to detect 1,000 items, and the second step is only for active ones. Partially (assuming 40% = 400 items) detects gender, but gender detection will be enabled repeatedly, with a total number of 1000+400=1400 times; and gender is first followed by activity, the first step is 1000, and the second step is active detection for males (300 items), the total number of times is 1000+300= 1300 times? It seems lower at first glance? The reason is that the active detection itself also includes activation, so the base difference depends on the unit price of the two tests. Generally, the unit price of tg active detection may be higher than that of tg gender detection, but you need to check the real-time price of the console. . The more important point is: When you screen gender first, the first step of activation detection has been completed for all numbers, and the second step of active detection will repeat the activation detection. However, if you only screen out a small number of men, the second step has a small base and the total cost is controllable, so Scenario 2 is suitable for targeting with a very low proportion of males.

Risk: If the gender ratio is not extreme, screening for gender first may result in the active testing base in the second step being still very large (for example, the female population is also very large), resulting in obvious repeated deductions.

---

Efficiency and cost comparison table of the two sequences

Comparison dimensions	Active first, then gender (recommended)	First gender, then active
Number of steps	2 tasks	2 tasks
Second step base number	Active number (usually 30%-50% of the original amount)	Target gender number (possibly 10%-50% of the original amount)
The status of repeated testing activation	The second step of gender detection will be repeated detection and activation, but the base is small	The second step of active detection will be repeated testing and activation, and the base depends on the gender screening rate
Estimated total number of tests (1000 original, active rate 40%, male proportion 50%)	1000 (active) + 400 (gender) = 1400 times	1000 (gender) + 1000 (active) = 2000 times (assuming gender screening is not reduced)
Estimated total number of tests (1000 original, active rate 40%, male proportion 5%)	1000 + 400 = 1400 times	1000 + 50 = 1050 times
Data quality	Exclude inactive accounts first, and subsequent gender data will be more effective	May screen out more inactive accounts, wasting the cost of the second step
Adaptation scenarios	Most general customer acquisition, limited budget	Extremely rare gender orientation, small-scale testing

Conclusion: Unless the target gender ratio is extremely low (less than 10%), it is always better to be active first and gender later.

---

Best Practices: How to Design Your TG Screening Task Sequence

Step 1: Determine the active window of the target group

When selecting TG active detection in the KK-DATA console, you can set the active time window (7 days/30 days/60 days). For promotions or instant messages, it is recommended to use a 7-day or 30-day window; for brand building or long-term reach, a 60-day window is recommended. The shorter the active window, the more accurate the filtered numbers are, but the smaller the base number.

Step 2: Filter for gender in active results

Use the active number exported in the first step (it is recommended to include the tgid field) as the source data of the new task and submit it for TG gender detection. Note: Even if both numbers are active, gender detection will still include activation detection, but because you have confirmed that the number is activated, this cost cannot be avoided. However, the base is small and overall controllable. If you want to completely avoid activating repeated testing, consider merging activity detection and gender detection into one task? Currently, the KK-DATA platform can only select one detection type (such as tg activity or tg gender) for each task, and it cannot be detected simultaneously at one time. The two-step approach is therefore current best practice.

Step 3: Use “Data Deduplication Warehouse” to prevent repeated submissions

When submitting a new task, if your original number pool contains duplicate numbers from previous tasks, it is recommended to enable KK-DATA’s “data deduplication warehouse” function. It automatically filters out numbers that have been detected in previous tasks to avoid charging the same numbers again. Note: Deduplication only applies to the number itself and does not differentiate between detection types. If your two missions are for different detection types (active vs gender), even if the numbers are the same, the second mission will still be charged (because the detection types are different). But a deduplication warehouse can prevent you from mistakenly submitting the same detection type with the same number multiple times.

---

Which order is more suitable for your business scenario?

Scenario	Suggested order	Reasons
Promotion of B2B business tools (target male users)	Be active first, then gender	The proportion of men may not be low, be active first to filter out zombie accounts, and then promote active men
Cross-border e-commerce beauty promotion (target female users)	Active first, then gender	Same logic as above
Looking for extremely scarce genders (for example, a certain type of male only accounts for 5% of the total account sources)	Gender first, then active	Quickly find all men (a small number) first, and then do activity detection on them to avoid wasting active detection on a large number of non-target accounts
Very low budget and only want to send a message once	Gender first, then activity (or only do gender detection)	If you only want gender data and don’t care about activity, you can do gender detection directly
Community recruitment (no gender distinction)	Just do activity detection directly	Save costs, no gender required

---

How to use KK-DATA to efficiently execute the best sequence

Specific operations in the order of “active first, then gender” are performed on the KK-DATA platform:

1. Generate number (optional): If there is no ready-made number, you can use KK-DATA’s “Global Number Generation” function to generate random numbers or designated numbers in 240+ countries for free. It is recommended to generate numbers for countries where TG users are active (such as Russia and Southeast Asia) to increase the initial activity rate.
2. Submit TG active detection: Create a new task in the console, select the detection type “tg active”, set the active window, and upload the number.
3. Export active results: After the task is completed, download the CSV or TXT file containing fields such as tgid, active status, etc.
4. Filter Active Numbers: Keep numbers with an active status of “Active” locally or through a script (or use the exported active list directly).
5. Submit TG gender detection: Use the above active number as the source, create a new task, select “tg gender” detection, and submit.
6. Export final data: Obtain the final results including gender, age and other fields.

---

FAQ

**Q: Does KK-DATA support simultaneous detection of activity and gender in one task? **

Answer: Currently, only one detection type can be selected for each TG screening task, and activity and gender data cannot be obtained at the same time in one task. It is recommended to perform it in two steps: first do TG activity detection, export the active number and then do TG gender detection.

**Q: Why is it cheaper to screen for activeness first and then screen for gender? **

Answer: Because TG activity detection can usually filter out 50%-70% of inactive numbers, the second step of gender detection only needs to be performed on the remaining valid numbers, and the total number of detections is greatly reduced. On the contrary, first screening for gender may retain a large number of inactive numbers, and subsequent active detection still requires detection of all valid numbers, resulting in a higher total cost. Unless the target gender ratio is extremely low (less than 10%), “active first, gender later” is a more economical strategy.

**Q: Is the age field in gender detection accurate? **

Answer: The age field is inferred based on public data or models and is not accurate at the ID card level. It can usually be used to divide approximate age groups (such as 20-30 years old, 30-40 years old). In the KK-DATA console export field, the age field can be used as a reference and is not recommended for scenarios where precise age is required.

**Q: If I have a large number of duplicate numbers, how can I avoid repeated charges? **

Answer: Before submitting a new task, you can use KK-DATA’s “Data Deduplication Warehouse” function. After uploading the number, it will automatically compare it with historical tasks and filter out the numbers that have been detected. Note: Deduplication only applies to the same detection type. Fees will still be deducted for different detection types.

**Q: Where can I check the unit prices of different detection types? **

Answer: After logging in to the KK-DATA console, you can select the detection type on the task creation page, and the system will display the real-time unit price of that type. The specific price is subject to the console display, please see the official website billing page for details.

---

👉Log in to the console to start screening numbers Two-way contact customer service https://t.me/kkdata_robot For more usage guidelines, please refer to Official Document