How Data Can Ruin Your Conversion Optimization Strategy

by Neil Patel

Last updated on July 27th, 2017

We conversion optimizers are in love with data. So much in love that it can skew our strategy or mislead us into thinking we don’t need one.

Data serves a purpose. All those statistics, analytics, reports, and numbers are intended to drive at something deeper and more enduring, which, in turn, enables the holistic optimization of a site for conversions.

When I consider the scope of the conversion optimization industry, I see a relentless focus on data, and I’m thrilled! It’s rigorous, informed, and scientific in thinking.

On the same token, I sense that data has a mind-bending impact on some optimizers. Often, instead of following strategy, they are following a false trails of numbers.

Where does this lead them? Sometimes it leads to dead ends. Other times it leads to marginal improvements. Usually, it leads to false expectations and disappointment.

I encourage a broader view of conversion optimization that uses data and strategy to efficiently increase results.

Data without an underlying strategy is useless.

The best one-liner on the role of data in conversion optimization comes from Avinash Kaushik.

Avinash is one of the most data-driven guys I know. He lives and breathes analytics in mind-blowing ways. Few people on the planet can touch his expertise in analytics.

avinash kaushik


But what does he say about the role of analytics in conversion optimization?

A data-first strategy, defined as above, is nuts. It will only slow down your progress and allow your competitors to crush you like a bug (even if you are a top player in your market today!).

You should reject data-first.

You should accept data-with strategies.

Bravo, Avinash!

As magnificent as data is, it is useless without strategy.

So what exactly is “strategy”?

A conversion optimization strategy is a process-oriented, systematic, heuristic-driven approach to gradually improving conversion rates and revenue on a website.

Sheesh, that was a mouthful; but it works!

To make this approach more tangible, let me provide you with a flyover view of a working strategy.

This is a hierarchy of optimization, and it was inspired by Kathy Sierra.


The above image is an excerpt from my infographic. You can see the whole thing at

Did you notice how much data talk there is in that strategy? NONE!

Okay, there is some data lurking beneath the surface. But you’ve got to realize that data doesn’t drive strategy, data supports it.

Many optimizers are feverishly testing, organizing, and crunching data, all in the name of optimization. But they are doing so without an eye to the bigger, broader framework of conversion optimization.

Pick a strategy. Any strategy is better than no strategy. And any strategy is better than a data-first approach.

There are a variety of strategies from which you can choose.

WiderFunnel prefers a cyclical model of improvement for conversions:


Image Source

Their LIFT analysis is a way to analyze the specific features that need to be optimized.


Image Source

LoginRadius shows how conversion optimization is a logical next-step in the process of business and marketing goals. This linear approach to marketing helps you figure out how to structure your entire conversion approach.


Image Source

The Convert blog has this great diagram for a CRO strategy.


Source (incl. sic.)

However you like to represent your strategy, I honestly don’t care. What I do care about is that you have a strategy, and that you use data to generate flesh out of the bones of said strategy.

Strategy first, then data. Remember Avinash’s wise words: “A data-first strategy…is nuts!”

Data is good for split testing.

So what is data good for?

In a word, everything except for strategy.

Split testing is where data really shines. My advice to incipient CROs is to pick a test and run with it. You’ll soon discover the siren song of data. It is powerful.

Running your first split test isn’t hard. You simply need the right tools and a spoonful of knowledge. Then, you’re ready to roll.

Okay. Hold on to your swivel chair. Some of the data-driven discoveries of conversion optimizers are absolutely shocking:

Take this one for example.

Security badges are normally a way to improve conversions. But not for icouponblog. When they dropped the badge from a checkout, they boosted conversion rates by 400%!


They removed it, replacing the spot with a search bar and links.


Study Source

Almost instantly, conversions skyrocketed.

You wouldn’t know that unless you relied on the data of split testing to give you an answer.

Now, take another split test that will leave you scratching your head. Express watches decided to remove a price guarantee and replace it with a trust symbol.

Check out the price guarantee. Looks nice, right?


They replaced it with a little Seiko authorized seller box.


Study Source

Results? A conversion increase of 107%

These two contrasting examples go to show that you can’t always trust your hunches. You need data to make the right choice.

Moving forward it is important to remember that data has value, but it is not a panacea. Data is one part of a larger whole in the wild and wooly world of conversion optimization.

Data doesn’t tell the whole story about your customers.

Any marketer in his or her right mind craves customer data. Comprehensive and reliable customer data is the key to a successful conversion effort.

Let me share with you some of the customer data that optimizers look for.

RetentionScience uses this set of five ecommerce data points to shape their big data approach.


Image Source

Every one of these features is data driven.

  • Credit score is a number.
  • Length of time on site is a duration.
  • Items purchased is a quantity.

Data are not limited to numbers, however. Here’s one outlay of the types of data that companies collect on their customers.


Image Source

And data goes deeper still! Today’s companies are using in-store sensors and other navigational data to track customer behavior.


Image Source

IBM’s 360-degree view of the customer shows you all the possible data points that they collect. They say, correctly, that “the journey that a customer takes with your company is an individual one.”

Thus, data overload!


Big data with all its risks, trust issues, and breaches is like a god: it’s omniscient and omnipotent in the minds of some, and controls everything they do and say.

But even big data can’t illustrate the full picture of the customer.

Each of the data-driven images above missed out on a crucial component: psychographic information.

Psychographic data are less quantitative and more qualitative, like a story. You can’t successfully quantify stories, sensations, attitudes, and the organic interactions that humans have with brands and life. As such, they must be interpreted and implemented differently.

More importantly, they tell us things that quantitative data cannot.

JeremySaid points out the data-driven contrast between demographic customer data and psychographic customer data. Here’s what he writes:

Conversion optimization, as you’re well aware, is the practice of turning visits into conversions. In order to rake in more conversions, you have to understand a person’s motivating factors, and that’s exactly what psychographic research is designed to explain.

He has a good point.

Demographic customer data provide us with exceptional information that contributes to conversion optimization. But no matter how accurate, secure, and reliable this data, it can’t give us the psychological insights into customers.

For example, demographic data shows us the customer’s age. That’s really useful.

But age merely informs us of a generational segment to target, not explicitly how to target them. A more actionable bit of information comes from psychographics. Now we know the customer’s age and their personality — they are highly aggressive.

Now we’re getting somewhere.

A 25-year old male who is aggressive is far more informative than a 25 year-old male.

Jeremy Smith’s explanation simplifies it in a helpful way:

  • Demographics = who is your customer?
  • Psychographics = why does your customer do what they do?

With that simple contrast in mind, you can see the enormous value of psychographics for conversion optimization.

Unfortunately, that psychographic good stuff can’t be pulled from Google Analytics or stacked in an Excel pivot table.

You have to get this information through careful surveys, long conversations, and intuitive awareness of your customers’ experience.

Data does not apply revenue-generating change.

Now, let’s get down to the ruthlessly practical. This is the last point in this article.

It’s also my most actionable point.

Data, no matter how awesome, do not optimize your website. You optimize your website.

This may seem like a facile distinction, but it needs to be emphasized. Some optimizers collect terabytes of data on their customers, collating, crunching, dissecting, and massaging it to death. Then, they iteratively split test the website until it’s worn out by the changes.

And then what do they do?

They don’t change a thing! I’ve seen it happen. The glistening pile of data and neatly arranged report is the only thing that they have to show for their labors.

Step back for a moment and consider the ultimate goal of conversion optimization. What is it?

The goal of conversion optimization is to increase revenue!

You can only increase revenue if you implement strategic, holistic changes that are supported by data.

When I wrote my Definitive Guide to Conversion Optimization, I made it a point to drive straight to home plate with this. Here was what I wrote:

“You need to optimize for revenue not just conversions.”

Calling it conversion optimization can be slightly misleading. Your task as a CRO is much bigger than mere conversions. It’s about revenue.

I gladly spent a quarter million dollars on CRO one year. Why? Because of revenue. My lavish expenditure on conversion optimization netted me millions of dollars in revenue.

In my early days of optimization, I would spin out split tests like nobody’s business. I saw massive upticks in my conversion rates — 30-40% increases!

But what I didn’t see was revenue increase.

The testing data was making me blind to the much bigger issue — unchanging revenue! Eventually, I refined my testing and focused more on the end product, revenue.

I got smaller split testing increases and ran fewer tests, but I achieved way higher revenue. Why? Because I got my gaze off the data, and onto a goal-focused strategy.


Don’t get me wrong. Data is awesome. I love it, use it, collect it, study it, cite it, quote it, read it, source it, and even built two businesses that collect it.

Clearly, I dig data.

But data can ruin your conversion optimization strategy. Wrap your head around these three points, and you’ll avoid ruin:

  1. Start with a strategy, not data.
  2. Use data where it matters most — split testing.
  3. Go beyond the data to get high-quality information about your customers.
  4. Instead of simply collecting data, make revenue-generating changes on your site.

Making this change will take your conversion optimization strategy to a whole new level, and give you lots of revenue to show for it.

How has data enhanced or limited your conversion optimization efforts?

No Comments


Get updates on new articles, webinars and other opportunities:

Neil Patel

Neil Patel is the co-founder of Crazy Egg and Hello Bar. He helps companies like Amazon, NBC, GM, HP and Viacom grow their revenue.


Comment Policy

Please join the conversation! We like long and thoughtful communication.
Abrupt comments and gibberish will not be approved. Please, only use your real name, not your business name or keywords. We rarely allow links in your comment.
Finally, please use your favorite personal social media profile for the website field.


Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Show Me My Heatmap

@CrazyEgg Thanks for no longer truncating naming conventions of exported snapshots. Saves so much time!

Nicole Mintiens


What makes people leave your website?