Hi I've done A/B testing last year with testimonials. I can give you some non so technical advices.
Results were this: it was uneffective to have too many testimonials, it was uneffective to have rolling testimonials, like in a slider.
What's effective is a small number (2 appears to be great) of well chosen testimonials, with a link to a different page to read a longer list of them. But then not a lot of users click it and all will go back.
What proved to be very effective is to use the testimonials as an incentive to read real content. In the end we chose to link the testimonials to the case studies, which is the logical long form of a testimonial.
So one or two well chosen testimonials to use as an incentive to visit long form content did provide leads.
You can check how testimonials are used on this homepage (webflow site) www.kds.com