The Nature of CRO Has Changed

The nature of conversion rate optimization has completely changed in the last six months. This is what seven and eight figure brands need to change in their existing CRO workflows.

For my own brands, and the people I know doing serious volume, this has completely changed the workflow and approach for getting a higher conversion rate out of your store.

Only successful brands could produce high quality content. Experienced content creators were gatekept. Statics were made in Photoshop, Canva, Figma by hand, every single detail. Not AI.

But now the bottleneck is no longer content generation. Now the bottleneck is using that content on your site in the most data-driven and optimal way.

Content is no longer a bottleneck for anyone. Even stores doing less than 100K a month are producing seriously good UGC content, static ads, and have decent structure on their landing pages and product pages.

So what is the new differentiator?

UGC is mainstream. AI generated content is getting better and better. Everyone is using it regardless of store size, and the creatives produced are better than ever.

So now the problem you are left with: you have tens of thousands of reviews. You have thousands, maybe hundreds of UGC videos. You have AI generated statics and content. Meanwhile, your product pages, video sections, and review sections have been the same, showing the same six videos and reviews for the last months, half a year, or even a year.

This is especially true if you have multiple products that you are focusing on simultaneously, because the operational cost of managing the content across multiple product pages is very hard to do manually.

If it makes sense to test hundreds of creatives weekly in Meta Ads, it also directly translates to the fact that it does not make sense to have a product page that remains static for months at a time. And not having your best content being put to active use using all the behavioral data that you track on your store.

The current CRO toolkit is fundamentally limited

The current toolkit that most CRO agencies promote and that most big brands have adopted: A/B testing, using heat maps and visitor recordings manually, and trying to see what works in your ads and using that insight to try to improve the product page — without even testing.

Before you listen to any conversion rate agency saying that A/B testing is going to solve your conversion rate problems, let me take you through the logic of how the current situation actually is.

You have A/B testing. You have your product page. You have all your assets — your videos, your images, your reviews. And you could try to A/B test. Swap out the selections, see what works. Look at visitor recordings, heat maps.

Why A/B testing breaks down at scale

If your goal is the highest possible conversion rate on your site, then A/B testing does not anymore make sense. It is sequential. It requires massive amounts of data to only make one decision. And it does not consider interaction effects.

This is a huge problem because your whole product page works through interaction effects. You cannot just show one element to someone and they will convert, or one section. It is about how all these pieces work together to take the visitor through their objections and give them clarity on the product's benefits and how it will work out for them.

The typical cycle of A/B testing that most CRO agencies offer as their flagship service takes an extremely long time just to test, for example, 10 different elements. After 14 weeks you're left with five elements that work better than the other five. The problem: if you want to test content, you have to do that individually for each product. And if you spend 14 weeks to find a video that works better, in that time span you have also produced probably a hundred more videos that you should be testing as well.

And this is not even mentioning the human labor that goes in and the operational drag that A/B testing leads to. The goal is not to do work. The goal is to get results. And if you isolate the result — which is the highest possible conversion rate — then A/B testing is a fundamentally limited way to achieve this.

A/B testing is extremely valuable in the beginning when you are making your structure, your product page, and getting your fundamental assumptions right. But for any high seven or eight figure brand, you already have a good foundation. The challenge now is: what do you fill that structure with to get the highest converting site without adding impractical operational drag?

The easy choices are already made

What usually happens: every brand adds a review section. They add product video sections. They improve the media gallery. They add trust badges and all the basics. These are very simple psychological decisions — you need social proof, you need to show the product in action, visual storytelling is more efficient than written text.

But it's after these easy choices that it becomes hard. Now you can't, with certainty, know that any change actually improves your conversion rate. And if you then make a lot of choices that are 50/50 whether they improve or decrease your conversion rate — since your choices are more or less random unless you test them — but if you A/B test them, you are moving so slow that the value of validating the decision is not even worth the workflow process needed.

You can't know which reviews convert best

Every well converting product page usually has at least two well converting review sections. One is the typical Loox, Judge.me, Yotpo review grid at the bottom. The other, most common, is a curated review carousel further up on the site.

The problem: if each product has hundreds, maybe thousands of reviews, and 80% of those are five star reviews — how can you and your team possibly know that these are your best performing reviews for this product? And know that possibly another selection of reviews wouldn't give you a 10 or 15% increase for this product page? And what if the reviews need to be different for returning and new visitors?

The same goes especially for videos. Just like in your ads — if you publish two videos, one might have a 4.2 ROAS while the other has a 1.7. This is the whole point why you even test creatives in the first place. But the website is just a giant creative. There is no difference. It's just that there haven't been any good tools available to solve this problem without extreme operational drag, which is why most people have just accepted this conversion rate ceiling and defaulted to manual human decisions.

You already trust an algorithm for your ads. Why not your website?

The funny thing is that every ecom owner and CRO agency understands this. They just don't know how to solve it. And I can prove that to you, because the workflow that currently exists that everyone relies on for their performance is Meta.

For your creatives testing: you give Meta your creatives. Nowadays people don't even use targeting anymore because Meta's algorithm is so efficient. And instead of relying on manual bids, trying to call Mark Zuckerberg to ask for specific ad distribution — you decide to trust the algorithm to deliver you the results that you want. And if you don't get those results, you focus back on creatives.

So the question I have, and the nature of CRO that I see changing: why are you relying on a human or a team of humans to try to use your content and behavioral data to get you the best possible results?

Fair enough if you haven't found a better option. But this is the exact problem and conversion rate limitation that me and my team have been trying to solve for the last year and a half.

The workflow changes exactly like Meta

The part of it that you can be really excited for is that it does not only get you better results — it completely removes the layer of operational human friction that stands between your data and what's on your website.

With Eevy, your workflow changes exactly like Meta. You focus on collecting and creating the best possible content you can. And Eevy uses an evolutionary algorithm to see the performance patterns and interaction effects in your conversion and behavioral data — to make daily data-driven decisions for what video content, review content, and image content is converting the best in your sections on your store.

It is absurdly synergistic because it means that not only will you get better utilization — you and your team can also focus your workflow entirely on getting better content. Improving your review collection flow to get image reviews, video reviews. Getting creators to record videos that you think will work really well on your site, that might also work well in your ads.

And possibly worse — maybe you are not even using a human to make these decisions. Maybe you have the same reviews, the same videos, the same media gallery for the last six months.

A direct, practical solution

I don't see the value in bringing up a problem like this without arming you with a direct, practical solution you can implement right away to remove this whole bottleneck.

We have just launched our public Shopify app after eight months in private testing with selected seven and eight figure Scandinavian brands. And have now made it so simple to use that we are releasing it publicly without it going through us as a CRO agency.

If you want to try how it feels to go from the old workflow to this workflow, go to eevy.ai or the Shopify App Store.

We work closely with our customers and offer live support with me and my team — where ironically, we are not using AI for that — because we want to identify every possible opportunity we have to provide value through the app.

In my opinion, the current apps available on the Shopify app store have become extremely complacent in their ways, to the point where a smaller team like mine can compete on both value and price.

We built it out of a necessity to solve a bottleneck we felt was limiting us no matter what we did. Our goal was to help with conversion rate optimization for other stores, as well as our own store, which is closing in on eight figures. We just found that no matter what we tried, we have the content — our content could get us way better performance than we have — but there was just no way for us to put that content to use.