¶ How to accurately determine the volume of traffic losses?
Traffic losses are often measured incorrectly.
¶ Incorrect Approach
The ratio of the Total number of visits to black pages (in this example, 121) and the Total number of visits (in this example, 505) is calculated, leading to the conclusion that 24% of visitors landed on black pages and traffic losses amounted to 76%.
The mistake lies in including the Total number of visits (505), which encompasses visits from all types of bots (scrapers, parsers, web crawlers), various companies, and services, of which there are countless on the internet. Bots may visit your white pages with varying frequencies, and sometimes their numbers can be very high.
¶ Correct Approach
It is necessary to calculate the ratio of the Total number of visits to black pages (in this example, 121) to the Sum of clicks from all advertising campaigns (from advertising accounts) on white pages where the hoax.tech script is installed (for example, 150 clicks).
As a result, 80.6% of clicks landed on black pages, while losses amounted to 19.4%.
It is impossible to accurately estimate the magnitude of traffic losses with small traffic volumes. The minimum traffic volume required for effective analysis is at least 500 clicks per flow.
¶ What Influences the Magnitude of Traffic Losses?
The magnitude of traffic losses can be influenced by the following factors:
-
Matching the targeting of the advertising campaign with the settings of the flow
For example, if the flow allows countries US, CA, and only Desktop devices, but the targeting in the advertising campaign is set to other countries and devices, visitors from countries and devices other than those allowed in the flow will remain on the white page, increasing the number of losses. -
JS fingerprinting enabled in the flow
When JavaScript fingerprinting is enabled in the flow, bot detection is more accurate, but redirection occurs not on the server, as with PHP filtering, but from the visitor's browser via JavaScript Redirect. This type of redirection is slower, and external factors, such as the visitor’s internet speed or device performance, can affect it. -
"Weight" of the white page and black page
When JS fingerprinting is enabled, redirection occurs only after the complete loading of the white page in the visitor's browser (which is hidden from them by a special preloader). Therefore, it's important to use lightweight and optimized white pages with minimal heavy media or scripts. The same applies to black pages. Especially in combination with slow internet—such as on mobile devices, some visitors may not wait for the redirect and close the browser tab. In this case, our system won't be able to obtain browser fingerprints, and the visit will be counted as a white-page visit, even though the visitor meets all the flow settings.
If the flow is configured only with PHP filtering, then the weight of the white page plays a smaller role, but optimizing the black page is still important. -
Server Location for White Page Relative to the Target Audience and the characteristics of the server
When JS fingerprinting is enabled, it's also recommended to use a server for the white page located geographically as close as possible to the target audience to reduce ping. Using weak shared hosting can also negatively affect the redirection speed. -
Click Fraud
Click fraud is a type of fraud involving fraudulent clicks on an advertising link by someone not interested in the ad (usually competitors). It can be done using automated scripts or programs simulating user clicks on advertisements.
With click fraud, the daily budget of the advertising campaign quickly runs out without generating any conversion. In the Dashboard, 50-100% of clicks may remain on the white page, although some clicks may still reach the black page. -
Characteristics of the Target Audience
Some Tier 3 countries have low average internet speeds, which, combined with other factors, can increase traffic losses.
¶ What is the Norm for Traffic Losses?
When the targeting of the advertising campaign aligns with the flow settings, and there is no click fraud, with JS fingerprinting enabled, it's considered normal to have traffic losses of up to 30%. However, these losses depend on all the factors mentioned above and may be slightly higher.
With PHP filtering, acceptable traffic losses typically range from 10-20%.
In native advertising networks, if ads are placed on bot-driven resources, traffic losses can be significant.
¶ How to Reduce Traffic Losses?
To reduce traffic losses, it's necessary to:
- Ensure alignment of targeting in advertising campaigns and flow settings;
- Optimize white and black pages for fast loading;
- Use a server for the white page located closer to the target audience and offering adequate performance;