Case Study: Chat Systems (Virtual Assistants)

Chat systems (commonly called virtual assistants) are implemented in your website to help convert potential clients viewing your website. As someone who is critical on site speed, I wanted to measure the performance impact of the most popular chat systems our clients use. How are chat systems impacting your website performance? Find out below.

The Testing Environment

  • I am using a fresh install of WordPress on WP Engine. The domain is https://mockingbird4.wpengine.com/ and is running the default 2020 WP theme.
  • The codes used are current clients codes.
  • The code is being implemented directly not the theme footer for the sake of brevity.
  • The async attribute is added to each script
  • Using GT Metrix to do crawls before and after.
  • Each script was ran three times on GT Metrix to try and get an an average. Screen shots were the last test ran with they were all close.

Clean Slate Stats

Blank WordPress Speed Test

Conclusion

The results I found were all pretty similar. Calculating some margin of error, all of them are relatively the same impact on site performance. Some things I found interesting were Apex chat has only three requests but each one is pretty heavy to load in. I was impressed how lightweight Juvo came to weight in compared to the rest.

In depth screen shots of each test is below the table

Chat System Extra Requests Extra Time Page Weight
Client Chat Live +29 +0.9s +329KB
Apex +3 +0.8s +285KB
NGage +22 +1s +131KB
Juvo +9 +0.8s +67KB
Smith.ai +9 +0.6s +380KB

Client Chat Live

Live Client Chat Stats

Filtered Requests

Live Client Chat Stats Filtered

Apex Chat

Apex Chat Stats

Filtered Requests

Apex Chat Filtered

Ngage

Ngage Chat Stats

Filtered Requests

Ngage Chat Requests filtered

Juvo Leads

Juvo Chat Stats

Filtered Requests

Juvo Chat Stats Filtered Requests

Smith.ai

Smith AI Gt Metrix Score

Filtered Requests

Smith Ai Filtered Requests