Convert More Site Visitors Into Buyers with the Laws of CRO

As you know if you’ve tried to do any CRO yourself – there are a lot of ways you can get lost on the way to a good strategy. A LOT.  

But, after over a decade of working with industry-leading brands to improve their conversion rates, I have some solid tips for how to get started. 

In this article, I will first cover what conversion rate optimization actually is, because a lot of folks get this muddled.  

Then, I’ll tackle two laws of CRO. The first one will be a bit more conceptual and really geared toward re-focusing how you think about CRO. And the other will give you a bit more tactical information to work with. 

Now, all this will, of course, only take you on part of your journey. We can’t arrive at great CRO in just a short blog. 

So, I will also share a discount for my book, Opting In To Optimization, to help you continue on your journey. Hopefully, you walk away with three important pieces for the rest of your CRO journey: new thinking, new approaches, and my book as a guide.

Let’s dig into the new way of thinking, starting with what CRO really is. 

What is CRO (conversion rate optimization)? 

True CRO can deliver amazing results. Some of the results our team has generated include 132% year-over-year growth, 659% mobile revenue growth, and an 86.7% increase in return visitors. 

Here’s how we define that kind of CRO. 

Real CRO is a data-backed system for increasing the percentage of website visitors that convert into customers, or more generally take any desired action. 

Let’s focus on the first piece. Data-backed. 

CRO isn’t about trying random tactics or best practices. It’s about using data, both quantitative and qualitative, specific to your brand. 

Which is why our first law is this. 

Law #1 Best practices are for beginners. 

Best practices are a bit like training wheels. They can add a lot of value when you’re a beginner (And there’s nothing wrong with being a beginner!). 

But the problem is most brands keep best-practice training wheels on for far too long. And that can have some negative effects. 

Interestingly enough, wine actually shows us a pretty good picture of this. 

After spending over 15 years analyzing business and wine, a professor of strategy and management at ESSEC Business School in Paris came to this conclusion. Adopting best practices is a good way to achieve average results. Worse, if you apply a practice that isn’t suited for your business, you can actually destroy value. 

So, what does this look like in ecommerce?

Brand collaborations were a best practice that really took off in 2020. It took off, in part, because it worked for really well some brands. When Justin Bieber merely mentioned a Croc collab, he sent their stocks soaring 13%. 

Now, it’d be easy to attribute causation there if you’re a footwear store. And say, “okay, then, collabs with cool names drive sales, let’s do that.” 

But then you get something like the Slack and Cole Haan footwear collaboration.

In theory, this was a good application of best practice. The laid-back sneaker was going to appeal to casual at-home workers, which was most of us in 2020. It was also supposed to tap into the twelve million professionals using Slack. That’s a lot of people. 

But just because best practices show a correlation between a certain action and a certain result in some brands, like Bieber and Croc, doesn’t mean it’ll cause the same result in others. 

In fact, the Cole Haan and Slack collab didn’t at all do what it set out to do. Why? Partly because this was a big audience miss. Slack is a big, cool brand. But folks who use Slack aren’t the same as a Slack community. 

The Slack audience doesn’t function like a celebrity audience. 

Ultimately, reviews were brutal, and Haan only managed to sell out after slashing prices by 50%. 

A total audience miss like this is just one inherent danger of best practices. They’re generic, so they rarely fit YOUR brand, cultural moment, or even, in some cases, geography.

To zoom out a bit, part of what we’re talking about here are two different cycles you can live by. And this framework has been particularly helpful for me.  

Applying best practices and trends is like prescribing before you diagnose. It’s like going to the doctor and, instead of hearing out your symptoms and history, the doctor immediately prescribes the top three most popular medications for you to try. 

You hear that and cringe, right? It’s dangerous malpractice.

The same approach is just as foolish in CRO. 

The alternative approach, the one effective CRO takes, is diagnosing and then prescribing. In this scenario, the doctor hears you out. If necessary, she runs some tests. She gathers the context and information she needs, then she makes a diagnosis. She diagnoses THEN prescribes. 

This is an effective approach to CRO, too. 

So, to bring home the big idea of this first law…CRO isn’t a one-and-done guessing game. And it’s not based on trends or best practices. It’s a scientific process based on iteration. 

Which brings us back to our CRO definition. 

We know we need data-backed inputs, not best practices or trends.  

And then, according to our definition, we also need to put those inputs into a system. 

What system?  

That takes us to our next law. 

Law #2 - Scientific method, not silver bullets. 

The scientific method is the system you use for making sense of your data:

• Contextual baselines instead of generic benchmarks. 
• Exploratory testing instead of sweeping redesigns. 
• Continuous improvement over massive tests. 

Let’s quickly look at each of these in turn. 

Contextual baselines instead of generic benchmarks. 

Benchmarking, in theory, tells you where your website stands compared to industry-wide averages. 

And it’s something I discourage our clients from looking at.

First, these averages are faulty. Most brands don’t share their true conversion rate. Those that do rarely know how to accurately calculate it. Or use bad data to do it. Or use totally different methods to get their final numbers. (We see this all the time when we get “inside” a client’s data.) Plus, there’s no way to confirm the accuracy of whatever number they share.

For the sake of argument, though, let’s say the numbers are remotely accurate. Even in that scenario, lumping together “like” brands for a shared benchmark rarely makes sense. 

For example, we’ve worked with three glasses companies at various points in The Good’s history. 

One catered to heavy computer users. The second to older adults and stylish readers. And the third to sports enthusiasts. All three of these were “glasses companies” but they had totally different audiences, advertising strategies, channel mix, and seasonality. 

Yes, they all sold glasses, but comparing their conversion rates would’ve been like comparing apples to pineapples to papaya — it wouldn’t have made sense or helped any of them.

That’s why you don’t start with benchmarks. You start with your baselines and aim for a conversion rate that’s always improving. 

Exploratory testing instead of sweeping redesigns. 

Next, you’ve probably experienced or seen this cycle: a website frustrates users, traffic stops converting into revenue, and someone starts a redesign.

The problem is, this is a bit like burning down an apartment building when you stop selling units. A redesign torches the building to the ground and starts again on the ashes. You wind up with new units — but there are no guarantee families in the area will want the style or layout you’ve chosen (even if YOU like it). In fact, there’s no guarantee you’ve solved the original problem at all. 

The other option is a slower, more measured approach. A scientific approach. With this method, you remodel room by room. You revamp one unit and gather feedback about what appliances, color schemes, and types of windows your buyers want. Then you take what you learn and apply that to the next unit — and the next, and the next — until you’ve remodeled every unit in the building. 

In both cases, you wind up renovating the building. 

But in the first approach, you make a lot of risky guesses. In the second approach, you run small experiments, gather data, and learn as you go. You rebuild around what your buyers want and are guaranteed to move more units because of it. 

The same is true of your website. 

Don’t throw the whole thing in the trash and start from scratch. Work piece by piece. 

Continuous improvement over massive tests. 

And third, the misconception that the size of the test is proportional to the size of the results. 

We’ve changed one word and improved conversions by 200%. Ad we’ve run giant tests that improve conversions very little. 

In fact, we now have a 1-10 rating system for the tests we run. 1 is a small test you can run immediately, and 10 is a full-page redesign kind of test that will take days or weeks to build. We were curious which size test has the biggest impact. So, we looked at our data – over a decade of it – and we discovered level 10 tests weren’t more fruitful than small, compounding tests. 

So, to sum up the law of scientific method, not silver bullets.

CRO isn’t a “one and done” guessing game based on faulty data like best practices or benchmarks; it’s a scientific process based on iteration and data. 

For More On The Laws Of CRO: A Special Discount On Opting In To Optimization For Tydo Readers

As I mentioned at the start of the article I want to help you arrive at a place of great conversion rate optimization. And we can’t do that in a short blog post. 

While you now hopefully have some tools for thinking about CRO, I have a special offer for Tydo readers to continue your journey. 

My book, Opting In To Optimization, condenses over a decade of optimizing for industry-leading brands, into a handful of immutable laws. 

Get your copy at 50% off with the code TYDOXTG by visiting: Opting In To Optimization.

Happy optimizing! 

‍