Opticon London 2016 — top take-aways from the AB-testing conference

Optimizely (the online AB-testing service) organized the Opticon conference this year again. Here are my top take-aways. If you are into split testing, personalization and data driven online methods, this is for you!

This year’s Opticon was one of the few events, where I have found all the presentations useful and informative. The greatest take-aways for me were though the inspirational thoughts from Optimizely’s clients and experts. But let’s see, what we had there:

Announcements & new features

The conference started with big announcements. It was centered around 4 new things, that Optimizely has just implemented into their new product set:

1. Improved Web Experiments: The AB-testing got a new, nice, more easy-to-use UI.
2. Personalizations: Dan Siroker (the CEO anyway) announced their new personalization solution. It can pull any data (usage data, events, sources, etc.) to target your audience and you can release it by AB-testing to measure the success of it.
3. Recommendations: An automatic recommendation engine, that can be set up in seconds and also can be released by AB-testing.
4. Optimizely Full Stack: This part is more interesting for developers. The “Full Stack” is a collection of open-source SDKs covering all the major platforms (Android, Objective-C, Java, Ruby, Python, etc.). With that testing — and the above mentioned new features — became easier to implement and it also makes things more scalable for bigger players.

These together are called “Optimizely X”. It comes with a 1-month free trial, which is quite nice.
+1: The show had many live-demos! So it was fun!

Useful stuff

I have found out about two freely downloadable practical “tool” on the conference. If you are an expert in CRO, these will be very valuable for you too. Both were coming from an agency called conversion.com:

1. Chrome Extension for Optimizely. This tool is very handy, when you set up your AB-tests and you want to check out all the different versions on Desktop or Mobile and learn if everything’s working properly. LINK.
2. Hypothesis Form: Tired of explaining to colleagues over and over that you need hypotheses, researches and goals before you set up an actual split test? When someone comes to you with a “good idea” next time, ask him/her politely to fill this form first. Raise consciousness on testing! LINK: https://conversion.com/hypothesis/

Inspiration

And finally I will highlight here 3 inspirational thoughts from rest of the presentations.

#1

You should look at personalization projects not from your company’s, but from the users’ point of view. Maybe your buckets won’t be based simply on ages/genders, but something that is driven more by users’ needs.

#2

Two important things from Secret Escapes’ Rohit Gupta:

Means — you don’t need a PhD degree to run proper AB-tests, but you need to follow the rules.

And even if you are very data backed, you need to think creatively and out of the box all the time. Go further, than simply testing button colors and stuff!

#3

Lastly, Dan Siroker highlighted the most important thing: “Our job is NOT to have all the answers. Our job is to enable our organization to be brave and experiment.”

Thanks for reading!
Enjoyed the article? Please let me know by clicking the 💚 below. It also helps other people see the story!

Tomi Mester
my blog: data36.com
my Twitter: @data36_com