MarketLinc Blog | March 28, 2018

By Lise Reddick

Part 3: Keys to Effective Testing with Live Chat

In this series, we’re breaking down exactly how to optimize your live chat investment through four fundamentals: targeting, testing, engagement, and measurement. If you missed out on part one or part two of the series, consider taking some time to go back. Or, if you’re ready, let’s dive deep into the importance of testing to maximize your chat performance.

As we explored in part one of this series, the adoption of live chat is not enough to attain great results. To improve performance, you must continuously look for opportunities to learn and optimize. Effective testing is a critical piece of the optimization puzzle. It can mean the difference between making blind changes based on a hunch and taking confident actions that are likely to achieve significant results.

The fundamentals of effective testing.

Effectively testing for chat optimization functions is much like the scientific method, which can be broken down into the following steps: ask a question, do background research, form a hypothesis, test that hypothesis through an experiment, and analyze your data to draw a conclusion. The scientific method applied to the testing of websites is often referred to as A/B testing (or split testing) where a control group is used to isolate the variable of the single element you want to test.

What live chat variables should you test on a regular basis?

There are endless factors that can contribute to an optimized live chat function. Here are a few things you can test to maximize your chat performance:

  • Placement
  • Imagery
  • Size
  • Font
  • Color
  • Messaging
  • Offers
  • Display Time
  • Display Duration
  • Display Animation

What are the results of effective chat testing?

While the testing process can be complex to navigate on your own, the knowledge you gain can be applied to significantly improve live chat performance. Here are just a few KPIs and optimization goals that you can measure and act upon as a result of live chat testing:

  • Conversion Rate
  • Average Order Value
  • Accept Rates
  • Revenue
  • RPV Uplift

An example of live chat testing.

Recently, a global leader in cyber security came to us with a question: “How can we increase incremental revenue through our live chat function on X, Y, and Z landing pages?” Through research, we found that website visitors spend 80% of their time viewing the left half of the page and 20% viewing the right half. We used this, along with additional analysis and market research, to inform our hypothesis: If we adjust the placement of chat to center left, then we can increase the Accept Rates on each landing page due to an increase in visitor engagement.

Then, we went to work to conduct a test to determine which placement generated the most incremental revenue. We analyzed the data to find that different placement performed better for every page. As a result, our client gained visibility into how to optimize the placement of their chat feature for each individual page, not just the website as a whole.

What you’ll need to get started.

Testing is one of the most important aspects of live chat optimization. The challenge comes when trying to isolate site visitors and controlling external variables (visitor type, campaign, site crash, etc.). Those just getting started with chat optimization often forget that results can differ not only based on page placement, but also based on visitor and profile type. To guarantee the best results and build on site performance over time, you’ll need a tool that can accurately:

  • Measure chat against no-chat in real time
  • Measure chat variations against one another
  • A/B split based on “like” visitors

The best live chat results come from continuous learning, fueled by effective testing. MarketLinc can help. Reach out to our team today and stay tuned for part four of this five-part series where we will explore the keys to effective engagement.