April 1, 2020
Create a new experiment in Optimize

Create a new experiment in Optimize


[MUSIC PLAYING] JASON O’GRADY:
Hello, and welcome to Optimize, Google’s free
testing and personalization product. I’m Jason, technical
writer here at Google. And joining me today is
Optimized Product Manager Alex “Q.” Hey, Q. ALEX DUONG: Hey, Jason. JASON O’GRADY: So what
is an A/B test, anyway? An A/B test, also known as a
split test, or an A/B/N test, is a randomized experiment
with two or more variants of the same web page. So variant A is the original,
and variants B through n contain at least
one element that’s modified from the original. So how do you create
an experiment? Now, you may hear a
couple of new terms here, but don’t worry. We’ll describe all of
them later in this video. So you create an experiment in
Optimize in three quick steps. First, create a variant. Then, target an audience. And finally, choose
an objective. Now, Q, want to show
us how it’s done? ALEX DUONG: Happy to, Jason. Yeah. So let’s jump right into
the Optimized home page. Here we are. And what you’ll see
is, we think of it as creating an experiment that
contains all of those variants. So let’s start there. Create experience. I just give it a name. It’s conveniently
pulled in the URL, based on configurations
previously, and by default, selects the A/B test. Now, I have a hypothesis
about the call to action on the home page. So here, I’d click
Alternate, and I would go ahead and click Create. But we’ve already
got one, actually, down here that I’ve started. And so let’s jump
right to that one. And here we are
on the Setup page. You can see up at
the top it’s called Alternate Call to Action. First step, as Jason
mentioned, is add a variant. Think of this as essentially
creating an instance, or a copy of the page, where
you want to make those edits. So I’m going to click Add
Variant, conveniently name it variant 1, click Done,
and there we are. You have the original
and variant 1. I’m going to go ahead now
and make the changes that I wanted to by clicking Edit. This opens up the page. There it is. But remember, this is a
copier instance of it. And none of this is live,
so we can go ahead and make all the changes we want. And if need to, right here is
that little Undo button in case you want to undo anything. So in this case,
my hypothesis is, I want to change
the call to action, and believing that if I
change it to something else, it’s going to actually cause
visitors to stay and visit other parts of the site. So click Edit
Element, Edit text, and I’m going to change this
to, say, View our catalog. Now, I’m going to click Done
down here in the bottom right. Take a quick look. Everything looks
right, and so I’m going to click Save
and Done at the top. Great. That brings us back
to the Setup page. Now, step two, as
Jason mentioned, is to go ahead and select
who gets to see it– what we call audience targeting. I’m going to click Customize. There’s a whole range of
options you can see here. But for this demo,
our flower shop is based in Fort
Collins, Colorado. And I just want
locals in that area to see this alternate version. So I’m going to choose
and click Geography. City equals has
already been set, and I’m going to search
for Fort Collins. There it is. Select it. And then go ahead and
click Add at the top right. There we go. Now, step three is objective. This is essentially, what is it
that you want to measure or try to improve with this test? So let’s click Add
Experiment Objective. Click Choose From List. Here, there’s some
preconfigured options. In this case, I had
mentioned that our hypothesis is that by changing
that call to action, we’re going to get visitors
to stick around and visit more parts of the sites. So I’m actually
interested in page views. Select that. Great. Now scroll up. What we’ve got here, is we’ve
got the original version of the website. We’ve got variant 1, which is
a copy with some changes we’ve made, noted here by
the number of changes. And we are only targeting
those down here at the bottom. Those in Fort Collins are
going to be the only ones that see this, with a
50% distribution between the original version
and this alternate version. Everything looks good. All we’ve got to
do is hit Preview if we want to check it out. And then, finally, just
click Start at the top, and you’re ready to go. It’s that easy. JASON O’GRADY: Thanks,
Q. That was great. So what you just
saw is how to create an experiment in Optimize
in three easy steps. Learn more about Optimize and
the Optimize resource hub.

2 thoughts on “Create a new experiment in Optimize

Leave a Reply

Your email address will not be published. Required fields are marked *