Jason Puckett

June 23, 2016

How Landing Page Testing Screwed Up CRO for Paid Acquisition

We Started From the Bottom (of the funnel), Now You’re Here (reading this article)

Paid acquisition is a major driver for many companies. Spending money on ads fuels customer acquisition, revenue and company-wide growth. Today, conversion rate optimization starts at the bottom. Landing pages. We in the performance realm don’t really call them landing pages, we call them Final URLs and they are a small part of our ads. More often than landing pages, performance marketers focus on keywords, ad creative, targeting, bids and our ad positions. Landing pages are somewhat of an afterthought, and rightfully so.

Don’t get me wrong, landing page design is important, but it should not be your primary CRO focus. Your ads need to be seen by the right people, contain the right messages, and be clicked on for a profitable amount… If these steps don’t take place, what is the point of optimizing the landing page?

Why are we not performing Full Funnel Testing?

The Introduction of the WYSIWYG Landing Page Editor

Back in 2009, a small company out of Y-Combinator changed the game for non-technical marketers. The product allowed marketers who could not code or design to change to their landing pages and measure the impact of these changes on their bottom line. What we saw next was explosive. The adoption of A/B and multivariate testing within the context of landing pages absolutely skyrocketed.

Optimizely boasts thousands of customers and claims that over 30 billion web experiences have been optimized through their platform, which I definitely believe.

Seven years after the introduction of the company, after A/B and multivariate testing have been adopted as marketing gospel, and after the dust has settled; let’s take a look at why we test landing pages first.

Why Do We Optimize Landing Page First? (bottom up)

This phenomenon of a “Landing Page First” mentality is purely circumstantial. Landing page optimization was an easy sell. These platforms solve a major problem. There are revolutionary companies in the testing space that have changed how marketers think about A/B testing. A/B and multivariate testing are now fully accepted and executed theories of marketing because of these landing page companies.

Building products that isolated landing page tests was definitely the best business model. It solved a major problem marketers experienced, and proved that marketers could obtain some quick lifts to key KPI’s by using these products.

Companies like Optimizely, SiteSpect and DataXu paved the way for A/B testing to be applied to many other aspects of the marketing mix: email subject lines, blog titles, ad copy, ad design, customer retention strategies. The possibilities are endless.

Yes, these companies are pioneers in the space but is a “landing page first” mentality still the right way to do things?

“Revolutionary” Doesn’t Mean “Best”

Somewhere along the way, digital marketers forgot about the stuff they learned in b-school. Yes, your “Target markets” still apply in the CRO world. Before we begin to build tests, we need to think about (1) who we are optimizing for and (2) what brings them into our funnel.

It is our philosophy that optimization needs to start at the top of the funnel. We think this is a marketing methodology iteration that will really drive industry-wide change. I started writing about these concepts two years ago, (Don’t Isolate Your Landing Page and Ad Copy Tests & Understanding Conversion Layers) and have spent the past 8 months really seeing these philosophies in action. AdBasis released the first version of our “Final URL Testing” feature set back in September of 2015, and have seen it used widely by many customers in a number of creative ways. Self admittedly it’s a primitive form of what the product will/should look like.

Being able to correlate specific elements on a landing page to successful ad copy or ad design (within the context of certain targeting parameters) is quite powerful. It will actually give advertisers and agencies some additional information that is not obtainable through Google Analytics and/or landing page testing platforms.

The possibilities are cool, but they’re also more fundamentally sound.

Testing in Isolation Violates the Principles of Multivariate Testing

If you’re not familiar with the basic concepts of multivariate testing, learn more here (What is Multivariate Ad Testing). Fundamentally, multivariate analysis is designed to understand the relationships between variables, if you don’t test your ad creative in relation to landing pages and targeting (keywords), these variables never interact, and thus would provide misleading results. One other important note is that marketers need to examine the elements within each ad variation, not the entire ad (colors, text elements, images used, etc).

What You Should Change: Test in the Order of the Funnel

Advertisers need to rethink the order of operations for building CRO tests (think PEMDAS). We believe that thinking of your testing in the order of the funnel makes the most sense. I am not saying that any one of these three “layers” is more important than the other, because if one is broken the others will not work. What I am saying is that tests need to be ideated and created in this order. People are not uniform, and thus a test shouldn’t be implemented by lumping your audiences, landing pages and ad creative together as one (or isolating them).


1. Targeting & “Purchasing Scenario”: This is where experiment structure begins. Selecting the demographics, geographies and the stage of the buying cycle you are attempting to target (keywords). This essentially tells who the test is being run on.

Select the “Purchasing Scenarios” that make sense for your brand. This could be ad group sets, demographics, geos or even products.

2. The Ad Content: Create an individual ad copy test for each scenario that you define (with multiple variations) . Each situation may potentially have different “optimal ad creative”.

Think about it, if you are segmenting your remarketing ads into buckets by “Abandoned Cart” vs. “Previously Purchased” or “Visited Homepage”, these need individual experiments and unique ad creative.

3. The Landing Page: This is where you can use those skills you’ve been working on for the past 7 years. Create multiple variations of landing pages and see what works. Hopefully you will be able to isolate the combination of all three layers together for true “Full Funnel Optimization”.

The Result:
Each “Purchasing Scenario” Will Have an {Optimal Ad Copy + Landing Page} Combination

Tying the Gap Together:

It is currently WAY too hard to track and test correlations between what is in your ads and the elements on your landing pages. AdBasis does a lot of this but not everything (self admittedly) Currently, our platform allows you to track the performance of two individual URLs within the context of an ad group set, and measure which ad creative converts best for that “ad group set + landing page scenario”.

Nobody else does this, so for now, we are satisfying the “realized need” of our customers because there is currently no better option. Using AdBasis for this is easy, you can get the “bottom line” results you need and our platform will give you the short term answer you need.

But there is so much more we want to do.

The question we answer currently is:

“Which landing page (Final URL) and ad content combination has the highest KPI performance within this ad group (or set of ad groups), set for each device?”.

But what if the question changed to:

“What is the impact if we use X on our landing page and Y in our ads?” (these could be phrases, words, punctuation, or combination of anything, really).

Simple Example: “What is the impact if we use the same headline in our search ads as we do on our landing pages for ad groups related to Product X?”

Advanced Example: “What is the correlation if we vary our headlines, but use an identical action verb in both our search ad and landing page copy?”

Oh, the Possibilities

As you can see in the above examples, the possibilities are endless. Developing an experiment plan and executing that plan will require work but the results will be worth it. Not only will you have improved your KPI’s, made money for your company and gained all the street cred in the world, but you may have actually learned something.

Understanding user behavior is an unbelievably difficult thing to do but with testing the ENTIRE funnel together, you might begin to pick up small things other companies don’t see. Maybe you’ll develop “Best Practices” that will help you get promoted, maybe you’ll be able to launch your own company… at the very least you will have had a little fun.

If you’re a marketer and you don’t love this stuff, you might as well pack it up and go home.

Wanna Chat? Hello@AdBasis.com

Wanna know more about the AdBasis platform? www.AdBasis.com/get-started

Have a question for us? We'd love to hear from you!

comments powered by Disqus


Popular Articles

Understanding Conversion Layers and Executing Ad Experiments

August, 14th 2014 - Optimization

3 Steps for Building a Paid Acquisition A/B Testing Plan

October, 15th 2015 - Optimization

All Things Tested - 3Q Digital's Sean McEntee

August, 31st 2015 - "All Things Tested"