Guide to base material screening and pit avoidance: 8 common mistakes and correct practices

Anyone who has done overseas customer acquisition operations knows that base material selection is a key step in determining the reach cost and final conversion rate. However, when many teams screen numbers in batches, due to the lack of systematic methods, they waste their budget and import low-quality numbers into private messages or community operations, resulting in an increase in account suspension rates and a decline in ROI.

This article summarizes the 8 most common base material screening errors** and gives the correct approach combined with practical tools (such as the KK-DATA screening platform). If you are a cross-border e-commerce, independent website promotion or TG/WhatsApp community operator, this guide can help you spend every penny wisely.

---

Mistake 1: Only detect “activated”, not “active”

Consequences

“Activated” only means that the number is registered with the platform, and does not mean that the other party has opened the application recently. Many operators get a batch of “activated” numbers and send mass messages, only to find that a large number of users don’t read them at all, the open rate of private messages is extremely low, and even frequent sending to inactive numbers triggers platform risk control.

Correct approach

Add activity detection dimension. For example, Telegram supports specifying activity windows (such as 7 days, 30 days, 90 days), and WhatsApp also has an activity field. Highly active numbers are suitable for instant event notifications and limited-time discounts; medium and low active numbers can be placed in the resurrection action sequence.

How to choose the activity window?

• Short-term activities (3-7 days): Select numbers that have been active in the past 7 days for the fastest response, suitable for flash sales and live streaming to attract traffic.
• Long-term community maintenance (30 days): Select users who have been active in the past 30 days to provide wider coverage while ensuring a certain possibility of interaction.
• Low active alternative pool: For numbers that have been active in the past 90 days but inactive within 30 days, you can try a wake-up strategy (such as coupons, red envelopes) to avoid discarding them directly.

Linkage between activity and subsequent operations

• Highly active numbers → Mainly promote instant conversion activities.
• Active numbers → Push brand content and gradually cultivate it.
• Low activity numbers → Limited time special offers or subscription reminders, if there is no response for a long time, they will be eliminated.

How to use KK-DATA to operate activity filtering?

When creating a task in the console, select the platform (such as Telegram), check “Active” for the detection type, and set the “Activity Window” (such as 7 days). After the task is run, the “Activity” field in the export results can be used directly for grouping. See Usage Documentation for details.

---

Mistake 2: Not doing cross-platform filtering, one number library dominates the world

Consequences

Many teams only have one phone number database, which can be imported directly regardless of platform. However, the same number may have completely different activation status on different platforms - for example, a certain number is registered on Line but not registered on WhatsApp. If you only use it on one platform, you will waste a lot of possibilities: you are spending money on an invalid number when it is clear that the other party is active on another platform.

Correct approach

Select multiple platforms to screen numbers at the same time based on the target market. For example, Line and Zalo are commonly used in Southeast Asian markets; WhatsApp is the main choice in South American markets; Telegram and iMessage can be considered in European and American markets. In KK-DATA, you can upload the number database at once and check multiple platforms such as Telegram, WhatsApp, Line, and Zalo. In the output results, the activation, activity, gender and other information of each number on each platform are clear at a glance, which facilitates subsequent distribution by channel.

---

Mistake 3: Ignoring gender and age targeting, resulting in misplaced delivery

Consequences

If you promote beauty products to male users or promote games to people over 60 years old, you can imagine the conversion rate. Most screening platforms (including KK-DATA) support gender recognition on Telegram, WhatsApp, Line and other platforms, and some also have age fields. But if the operator does not actively use these fields, it is equivalent to giving up the big move.

Correct approach

When creating a screening task, check the “Include Gender” or “Age” field. Taking KK-DATA as an example, Telegram’s gender detection results include an age range (for example, it can be interpreted as people around 30 years old, rather than ID card-level accuracy). After exporting, secondary filtering is performed based on the operation scenario, such as the combination of “female + 25-40 years old + active in the past 30 days”, which can greatly improve the conversion of words.

---

Mistake 4: Not deduplicating numbers and wasting budget on duplicate detection

Consequences

Screening numbers are charged on a per-item basis. If the same number is tested twice, it will cost twice as much. Some teams have accumulated tens of thousands of numbers over a long period of time and directly upload all of them for each new task, resulting in a large number of repeated deductions.

Correct approach

Use the data deduplication warehouse that comes with the platform. KK-DATA provides a cross-task number deduplication function: when the uploaded number has already appeared in a previous task and has a result, the system will automatically skip it and no new fees will be incurred. It is recommended to use the deduplication warehouse to clean the numbers before each task to ensure that each number is only detected once.

---

Mistake 5: Blindly using global number generation without verifying validity

Consequences

In order to quickly expand, many operators directly use number generators to randomly generate a batch of numbers, thinking that the newly generated numbers are “clean”. However, many number segments are actually not assigned by the platform at all, and the detection activation rate is extremely low. If you screen numbers directly, it means you have tested a bunch of invalid numbers, but the balance has been deducted.

Correct approach

After using the random generation module, immediately perform activation detection. KK-DATA’s “Global Number Generation” function provides number segments for 240+ countries/regions. The generation process is free, but the generated numbers must first undergo activation testing before being used in the next step. It is recommended to first generate a number segment for a designated country/region (such as the United States +1), and then select the “Activate” test of Telegram or WhatsApp to confirm that it is valid before entering it into the database.

---

Mistake 6: Focus only on one platform and ignore other potential channels

Consequences

Some operators have only been engaged in Telegram or WhatsApp for a long time, thinking that a certain platform is the “main battlefield”. However, users are distributed across multiple platforms. For example, Vietnamese users use Zalo heavily, Taiwanese users prefer Line, and Japanese users also commonly use Line. Only doing a single platform will lose a lot of potential customers.

Correct approach

Plan multi-platform reach strategies based on target markets. In KK-DATA, you can cross-check the same batch of numbers on Telegram, WhatsApp, Line, Zalo, iMessage, RCS and other platforms. After the results are exported, channel resources are allocated according to the rule of “the highest priority for a certain platform’s activity”. For example: Telegram and WhatsApp are active in parallel in the Brazilian market and can be enabled at the same time.

---

Mistake 7: Not using export fields for subsequent analysis

Consequences

Many operators only export the two fields “number” and “activation or not”, while abandoning a lot of valuable information, such as tgid, wsid, uid, avatar, age, etc. These fields are very critical for subsequent labeling, building user portraits, and group marketing.

Correct approach

When exporting a task, check all available fields. KK-DATA supports exporting to CSV/TXT format, including tgid (Telegram user ID), wsid (WhatsApp user ID), gender, avatar link, age, etc. Importing this data into CRM or operations backend allows for more refined personalized contact. For example, use tgid to send messages directly through the Telegram Bot API to avoid relying on mobile phone numbers and improve delivery rates.

---

Mistake 8: Ignoring cost control and account security

Consequences

Once the sieve number gets out of control, the balance will be consumed much faster than expected. Some teams use public accounts to recharge, or arbitrarily disclose customer service links on the Internet, resulting in loss of funds. In addition, not paying attention to real-time price changes and still submitting a large number of tasks when the balance is insufficient will also interrupt the process.

Correct approach

• Pay-as-you-go billing, no subscription package: KK-DATA deducts fees on a per-item basis after recharging the balance, and you pay for what you use. The estimated cost will be displayed before submitting the task. It is recommended to get into the habit of converting the scores before submitting.
• USDT Anonymous Recharge: The minimum amount is about 50 USDT, no need to bind a bank card, and the balance will be automatically updated upon arrival.
• Verify official customer service: The platform provides a unique customer service Telegram account in the official channel and documents, beware of fake customer service fraud.
• Task Notification: Turn on Telegram notification in the console, and you will receive a reminder when the task is completed to avoid repeatedly refreshing the page.

---

FAQ

**Q: My number database is very large. How many records can I submit for one test at most? ** Answer: KK-DATA supports up to about 1 million numbers in a single task. If the number exceeds one million, it is recommended to submit it in batches.

**Q: Is the age field in Telegram gender detection accurate? Can it reach the ID card level? ** Answer: It is not accurate and cannot reach the accuracy of the ID card. This field is the age group inferred based on the public account information (such as about 30 years old). It is only used as a directional reference and cannot be used for precise age screening.

**Q: Are the unit prices for different platforms (Telegram, WhatsApp, Line) the same? ** Answer: Not the same. The unit prices of different platforms and different detection types are different. For details, please log in to the console to view the real-time price, or visit the [official website billing page] (https://kkdata.cc/billing/).

**Q: How do I avoid checking the same number twice? ** Answer: It is recommended to use KK-DATA’s data deduplication warehouse and automatically compare historical task results before uploading to avoid repeated deductions. For specific operations, see Usage Documentation.

**Q: The generated number is free, but how much does subsequent screening of numbers cost? ** Answer: The number generation module is completely free. The screen number is deducted based on the number of tests. For details, see the real-time price on the console. It is strongly recommended to conduct a small batch test after generation to confirm the efficiency of the number segment before screening numbers on a large scale.

---

The above 8 mistakes are pitfalls that many overseas teams have encountered. Base material screening is not a simple “batch inspection”, but a combination of data science and operational strategies. As long as you avoid these minefields, the quality of your numbers will be greatly improved and the cost of acquiring customers will naturally decrease.

👉 Log in to the console to start screening numbers Two-way contact with official customer service: https://t.me/kkdata_robot More documentation and tutorials: https://docs.kkdata.cc/