Where are we taking Cauzal?

December 3, 2020

Our story

We started Cauzal with one idea in mind: use Machine Learning and AI to maximize the impact of content.

“Good content converts more” is pretty intuitive but we found that existing solutions didn't answer the questions: what defines “good”, how much is “more” and what is the true value of content.

After 6 months of Beta testing, the results are clear: our approach to content optimization is sound but we need to go further and move from page optimization to automated site-wide optimization of content and user flow.

Here are our 3 focus areas:

There is a need for a more intelligent and robust measurement of content: what are the pages, keywords, concepts that are really converting and engaging. Not simply in the path of conversion, but really participating in converting visitors.
This combined with our predictive detection of propensity to convert by dynamic segments of visitors allows the system to highlight what changes would optimize their experience
Double (or quadruple) down on automation and scale.

Our first phase until now: retire A/B (split) testing

A/B testing or even Multi-variants testing is a good way to validate assumptions but not a very viable approach to unbiased optimization:

  • Split testing is looking for one winner; we argue that your audiences have many intents and sensitivities and there’s room for more than 1 winning messaging
  • It is very time and resources consuming; we believe Machine learning is here to do the heavy lifting, fast
  • It is static. What your visitors want now is not what they’ll want tomorrow and we think an adaptive and continuous optimization solution is a better alternative to random testing

Cauzal AI offers that: provide all the content you have in mind and let the system optimize the display for you. Easy to use, good performance. But…we realized this is not enough. There is still a bias and some “guesswork” in the equation. We’re fixing that now.

What we're introducing now

True incremental value of content

Understanding the true value of content requires more than what most web analytics tools can provide: incremental attribution.

In other words: when we analyze all the journeys of your visitors, what is the contribution of content (pages, keywords, concepts...), as opposed to simply being in the path of conversion. We’re using game theory to achieve that.

What you get: a complete picture of your site content and what’s working.

Attribution of content efficiency

Sequence and user flow

In addition to the content displayed at every step of their journey, we found that another important question to answer is: What is the next best step for this user?

Again, machine learning and chain analysis can help and we’re adding that to our analysis of website

AI-based recommendations

When we combine all 3 elements...

  • Prediction of propensity to convert and intent of visitors
  • Understanding of true performance of content, and
  • Processing of all user flows

...our model can take the step of actually recommending content optimizations:

  • What content should be displayed to your visitors at any step of their journey?
  • What is the best next step for your visitors?