Skip to main content
How Can We Help?

Understanding AI Recommendations

Posted
Updated
ByEmily Gomez
AI Recommendations are suggested optimization opportunities generated from Bluvia A/B Testing Insights.
A recommendation gives you an idea of what may be worth improving or testing. Your job is to determine whether that recommendation makes sense for your website and optimization goals.

Review the Recommendation

Before acting on a recommendation, understand what Bluvia A/B Testing is suggesting and what you hope to improve.
Consider:
  • What is Bluvia A/B Testing recommending?
  • What information led to the recommendation?
  • What part of the visitor experience could be improved?
  • Does the recommendation support an important website or business goal?
  • Is there a clear way to measure whether the change works?
These questions can help you decide whether a recommendation is worth pursuing.

Prioritize Your Recommendations

You don’t need to act on every recommendation.
If several opportunities are available, consider which ones have the greatest potential impact.
You may want to prioritize based on:
  • Importance of the page or element.
  • Amount of traffic.
  • Connection to an important Conversion Goal.
  • Potential impact on the visitor experience.
  • Ease of testing the recommendation.
For example, an opportunity on a high-traffic landing page may be more valuable to investigate than a small change on a rarely visited page.

Running Multiple Recommended Tests

You don’t have to wait for one experiment to finish before testing another recommendation.
Bluvia A/B Testing allows you to run multiple experiments simultaneously. However, you should avoid running experiments that overlap with one another.
Experiments overlap when the same visitor could be included in multiple tests that affect the same page, content, or experience in a way that could influence the results.
For example, experiments may overlap if you:
  • Run two experiments that make changes to the same page.
  • Test the same CTA or element in multiple active experiments.
  • Run a page-level experiment while another experiment changes an element on that same page.
  • Run multiple tests where one change could influence the Conversion Goal of another.
For example, imagine you’re testing a new hero section on your homepage while simultaneously testing the CTA inside that same hero. If conversions increase, it may be difficult to determine whether the hero change, CTA change, or combination of both caused the improvement.
Experiments on separate areas of your website can generally run at the same time when they don’t interfere with one another.
When reviewing several AI Recommendations, consider whether the experiments could affect each other’s results before launching them simultaneously.
If two recommendations overlap, run them separately so you can more clearly understand the impact of each change.

 

Use Your Own Context

Bluvia A/B Testing can identify patterns and opportunities, but you have additional context about your website and business.
Consider the recommendation alongside your:
  • Brand and messaging.
  • Business goals.
  • Customer knowledge.
  • Current campaigns.
  • Website strategy.
  • Previous experiments.
If a recommendation doesn’t make sense for your website, you don’t have to use it.

What Happens Next?

If you decide a recommendation is worth testing, you can create an experiment directly from the Insight.
Bluvia A/B Testing will prepare the suggested change for you so you can review and refine it before launch.
See Accepting Suggested Changes for the complete process.

Related Articles

Continue with: