Author’s Note: All the good ideas in here? They’re from our very excellent Director of Analytics Aleksander Krupski. He’s how we pull off insane conversion rate optimization stunts with Google Optimize for our clients.
We’ve had a blast with Google Optimize ever since Google released their testing and personalization tool in March 2017.
Super quick and easy experiments. Sharp insights and reports.
And the price — 100% free — is just right. 👌
First, we just used the Google Optimize tool to boost our own conversion rates. Fast-forward to today: we use Optimize to drive growth for clients, quickly making (and test) changes on landing pages, websites, even entire funnels.
The kicker: we’re doing it without relying on our client’s developers.
What does that mean?
You don’t gotta be super technical to crush it with Google Optimize.
Like most Google Optimize tutorials, we’ll show you how to set up Google Optimize, how to run your first experiment, and how some of the finer-grain features work (like Google Optimize’s new personalization abilities).
Along the way, though, we’ll teach you a little A/B testing kung fu. After all, a tool is just a tool. Finding good conversion rate optimization tactics are ultimately what wins the day. To do that, you gotta experiment.
Your Google Optimize Tutorial Begins…
Setting Up Google Optimize
Installing Google Optimize is easier than punching through a wet paper bag.
Especially if you’ve already hooked with Google Analytics and/or Google Tag Manager. In fact, one of the best parts of Google Optimize is how seamlessly it integrates with other Google products — especially for reporting, as we’ll see.
Google documents the process exceedingly well, so I’ll be borrowing liberally from them and providing screenshots.
In order to use Optimize, you’ll need the following:
- A Google Analytics account.
- The Google Analytics code snippet installed on your website.
- The Chrome web browser.
- The Optimize Chrome extension to create experiences in the Optimize visual editor. If you only need to view experiment results, you don’t need the extension.
—Set up Optimize
So to break it down, you need Google Analytics, Chrome (I’m a Firefox man myself so this was a minor inconvenience), and Google Optimize’s Chrome extension — turns your browser into an pretty sturdy page builder.
Next, put Google Analytics on your website. Yep, just do it.
(Google probably used Optimize to test that button copy).
You’ll create a new account…
…then connect Google Optimize to your analytics…
…and find this nifty little box in the right-hand column.
Yes, it’s time to Deploy Optimize. Jazz hands. 👋
If you just slapped Google Analytics on your site, this next bit of code will look very familiar. If not, you may have to do some digging.
This is as tough as it gets, to be honest.
Find the Google Analytics code in your website, then add the line for the Optimize plugin (or Snippet, as I’ve seen it called).
The next step gives you a little more code to implement.
“Minimize Page Flicker.” The heck is that? It’s when your original content shows up for an instant before your test content does. We don’t want that (ruins the illusion and freaks out your visitors).
Rad. Now you’ll navigate to the site you’re setting Google Optimize up for and — if you have the Google Optimize Chrome extension installed — confirm that you have, in fact, deployed Google Optimize.
If you’re like us, and you set up Optimize through Google Tag Manager, you’ll want to double up and install it through Analytics as well. Why? Google says so.
Which means there may or may not be Google self-interest involved. 😉
When you deploy Analytics tags through Tag Manager, it is still recommended that you install the Analytics tracking code with the Optimize plugin directly on the page (as opposed to deploying Optimize through a tag in GTM)
—Optimize and Google Tag Manager
Still not set up? Try these:
If all went well, you’ll be ready to create your first experiment.
Running an Experiment with Google Optimize
“Hey, do we have a GIF showing the entire process from start to finish, like one of those animated recipes on Instagram?”
Yeah, we got one of those.
Let’s go through this step-by-step.
Step 1: Create a New
Copy Rant: In the latest update, Google changed their terminology — what used to be called “experiments” are now “experiences” … though in some places (like the page title), we still see “experiment.” Just know if you’re doing any reading on your own and you see “experiment,” the author or article likely means “experience.” OK, rant off.
Go to your Google Optimize dashboard.
See the big blue button that says “Create Experience?”
The right column pops out or changes and asks you to describe this magical new experience you’re preparing to create.
From here, there are four kinds of experience you can create:
A randomized experiment with 2+ variants of the same web page.
A round robin-like test that tries multiple page elements in different combinations to find the winning arrangement.
Test two pages that are radically different from one another (including by URL)
Google’s latest experience is a bit out of place amidst the other three types.
The Personalization experience isn’t a comparative test. Instead, you create a variant (or choose a variant from another experience you’re running) and show the variant to select visitors.
The targeting rules — and there are a TON of them — work the same way they do when you set up segments for other test types (e.g. A/B, Multivariate, Redirect). Think “Who” and “When.”
See for yourself:
Google Optimize has a robust knowledge base around targeting rules.
Here’s the short of it:
URLs: Decide which pages the test runs on — larger images on product pages, for example.
Google Analytics audiences: Show the variant to a specific audience. Only available if you upgrade to Google Optimize 360. This may be worth paying for (we’ll let you know after we try it).
Google Ads: Target segments you’ve already identified in your Google Ad campaigns.
Behavior: Great for first-time visitors and/or visitors coming from sites you specify (a competitor, perhaps??)
Geo targeting: Good ol’ location-based segmentation. Great for pushing content to spur IRL action.
Technology: Segment by Browser, Operating System, and Device category. Get really specific, e.g. Google Nexus 5
First-party cookie: Just what it seems. If you give a guest a cookie, you can offer them milk next time they drop by.
Query parameter: Key off of your UTMs to target people from certain campaigns or other tactics you’re running.
Data layer variable: Creates a virtual scrap of paper and jots user behavior on it to refer back to later. #oversimplification
How you use all those rules and experience types depends entirely on what you what to find out.
For that, we’ll need a hypothesis!
Step 2: Set A Hypothesis
The entire success of Ladder (and our clients) is built on one simple thing: a growth test method that succeeds 300% more often than the industry average.
It’s worth quoting our CEO Jon Brody on this at length:
First, we ask a question: Which funnel stage do we target with a test, and why?
Next, we do our research: We analyze the entire marketing funnel of a business we’re working with in order to identify the biggest growth opportunities. At the start of any relationship, we actually perform a full-funnel growth marketing audit to illuminate the growth levers.
After that, we pick tactics – out of the largest possible array of options from our database – that target those growth opportunities.
For each tactic, we write a hypothesis: What is it that we’re aiming to achieve with the test in a way that will increase ROI and boost a business’s KPIs? Every hypothesis is there to push the needle on a metric to drive growth, not just maintain status quo baselines.
—Why Our Growth Test Success Rate Is 300% Higher Than Market Average
All conversion rate optimization hinges on testing.
So, here’s how we wrote a hypothesis for an in-house growth test we ran:
Changing the CTA copy from “Talk to a Strategist” to “Get a Free Growth Audit” may help us attract more leads. If this tactic nets us a 10% higher lead conversion rate over four weeks, we’ll know it’s a worthwhile copy change.
Notice that the hypothesis includes:
a) what’s being tested
b) a goal
c) a test duration
You can’t achieve success if you don’t define it first.
Look below and you’ll see we entered this hypothesis in the Ladder Planner, one of our proprietary marketing tools (you can now get access to automated insights and a growth tactic database too) that tracks all the experiments we run for ourselves and our clients.
Naturally, this hypothesis goes into Google Optimize for reference and consistency.
Step 3: Set Goals
This is our favorite part about Google Optimize.
We set up tons of Google Analytics goals for both ourselves and all our clients…
…and Google Optimize ties directly into Google Analytics.
The implications? You can immediately connect your experiments to the goals you’ve already set up. All you have to do is Add an Experiment Objective.
Select “Choose from list” and the goals you create in your associated Google Analytics view will show up in the drawer that pops out.
Now we’ll set up which goals we want to track with this experience. Since we’re tracking whether or not we get more submitted forms in the hero section, Hero Form Submission will be our primary goal.
We can select up to 3 goals, so what else should we choose…
How about bounce rate? Does the new CTA make people stay on the site longer and actually click around?
A tertiary goal might be our Contact Form Submission goal, which tracks the submission of our more detailed contact form further down the page.
So here’s our final setup:
Step 4: Set Targeting
Google Optimize is all about you carving up the pie — e.g. your site traffic — in as many different shapes and slices as you want.
We’ll want 100% of our traffic to get both the original and variant versions of the CTA copy, and we’ll want each to get 50% of the traffic. We’ll have the test show on page load, since we want them to see it immediately. You can also have a test load on a custom event, which lets you do some other fun things like load different forms for different visitors when they click a CTA.
As we covered above in Personalization, you have access to a ton of Rules that determine who the tests get shown to. Get creative with these.
Step 5: Create & Edit Your Variant
What goes into a good test variant?
Here are some tests we’ve run for inspiration:
- Test of the Month: How Ladder Launched a Facebook Video Ad With an 11% CTR
- Test of the Month: June 2018
- Test of the Month: July 2018
- 12 marketing experiments we ran in February 2017
In Google Optimize, let’s create our variant. Name it whatever you like — we called ours “CTA Edit.”
Once you’ve done that, click on the variant and you’ll be taken to the page you chose to test.
Next, select the element you want to edit. Ours is the Hero Section Call to Action Button.
And that’s it! Click done at the bottom-right, hit save then done at the top-right, and you’re ready to…
Step 6: Run Your Test
> Start Experiment!
Our experiment is now live, and we’ll soon gather data from our visitors.
Tracking Your Experiment
Now, it’s worth saying that Google Optimize has a really nice performance dashboard.
But to be perfectly honest, nobody needs another dashboard to log into!
(That’s why we pipe them all into Ladder Spotlight, the proprietary marketing tool we use to analyze every digital touchpoint of our clients’ businesses and find growth opportunities.)
Luckily, all of this should be connected to Google Analytics.
Yep, all our tests are right there!
Under Behavior > Experiments, you can see every single experiment you’re running and have run.
Wow, looks like our new hero copy is crushing it!
In just 8 days of testing, there’s an 88.5% probability that our hero copy variant is better than the original.
Now, I’m going to export this data to Google Docs, and integrate yet another tool from the Google suite — Google Data Studio — to make these results look even more astounding (as if that were possible).
…but that is for another blog post.
So, happy experimenting with Google Optimize. If at first you don’t succeed, that’s normal. After all, even though we’re 300% above average in our test success, our tests still fail 57% of the time.
That’s why you need persistence.
Persistence, and a solid A/B testing tool.
— Google Analytics (@googleanalytics) November 1, 2018