The brief looked familiar.
A mid-sized fashion brand was struggling with customers not finding what they were looking for: their bounce rate for visitors coming from certain creatives was skyrocketing; others were dropping off on the PLP, or after running a search; the take rate on PDP product recommendations was abysmal.
The brief was clear: review the entire upper funnel, identify opportunities to improve the UX, and execute.
So we did. We mapped their user flows, audited the product discovery UX, looked at their heatmaps. We identified friction points and proposed solutions.
But as we got to work, we started an uphill battle against an awfully maintained product catalog:
- Shoes were organized by gender, while apparel and accessories were unisex. A customer clicking on "Women" landed on footwear and found no path to jackets. A customer clicking on "Jackets" couldn't filter by gender.
- Attributes were filled in unevenly: different names for the same color, some products with detailed specs and others with none, variant names that followed internal naming conventions rather than how the customer thinks.
- A lot of the attributes customers want to filter or search by—occasion, fit, fabric—were not being captured in a structured format at all. There was no data to filter on, so there was no filter to build.
No matter how beautiful we made their search, filters, or recommendations, the data powering these features was too scarce and inconsistent to produce good results. Every improvement we made felt like putting lipstick on a pig.
And while we could have just shipped the cosmetic improvements and called it a day—that's what the brief asked for, after all—we thought they deserved better. So we went back to the drawing board: we scheduled an all-hands-on-deck workshop, involving everyone from their team who was responsible for or impacted by the structure of the catalog, and figured out a richer, cleaner structure that would allow the new UX to really shine. It wasn't easy, and it went from a one-month project to a multi-month nightmare of coordination, attribute enrichment, and data migrations. But it was worth it.
In systems thinking, you'd call the catalog structure a leverage point—the bottleneck that prevents everything else from working correctly. If you only work downstream of the leverage point, you will have minimal impact. But if you re-focus your efforts on the leverage point instead, that can unlock results you couldn't have dreamt of before.
Back to the project: our initial UX work wasn't wasted, but it was work at the wrong leverage point. The real problem was in the taxonomy, the attribute schema, the naming conventions. Once we helped the client restructure those around how customers actually browse, the discovery experience became much easier to solve. The interface hadn't changed dramatically—it just finally had something coherent to represent.
This is, we've come to believe, what CRO work looks like when it's doing its real job. Not shipping UI improvements, but using the discipline of studying how customers interact with a store as a diagnostic lens—one that, if you follow it honestly, keeps leading you to the real source of the problem.
Finding Leverage Is Only Half the Work
Unfortunately, not every project has a happy ending. Sometimes, you find the leverage point and still can't get anyone to act on it.
Not that long ago, a subscription brand came to us with a retention problem. Their churn was insane, with most subscribers canceling by month six, and they wanted some help figuring out a strategy.
In our very first conversation with them, we threw around some ideas: redesigning the cancellation flow with better copy and deflection offers; launching a loyalty program to incentivize customers to stay; building new CRM flows to remind customers of all the value they were receiving.
But when we dug more into the data, we realized they had a deeper problem: as is often the case with subscription brands, most customers were just canceling because they had too much product.
You see, this particular brand had built a subscription model around regular shipments of a single, durable product category. By month three, most subscribers felt like they had more than they would ever use. By month six, what started out as a fun monthly treat had become an expensive annoyance.
Just like in our fashion project, working on the cancellation flow, loyalty offer, or product education wouldn't do anything to fix the fundamental misalignment between what the brand offered and what the customer needed. And just like in our fashion project, the work wasn't useless, but it was not aimed at the leverage point.
As we continued our data exploration, we realized there was a strong opportunity to rethink the subscription model itself, moving from a single-product cadence to a membership that gave customers access to a much broader catalog, giving them more ways to spend with the brand on their own terms. Less accumulation, more agency.
But this was a very different project from the one we had been hired to do. What had started as a conversation with the Director of E-commerce now involved the CEO, the CMO, merchandising, logistics, and a willingness to rethink the business model from the ground up. The project was also very risky: while they already had the product they needed to support our proposal, everything else was a bet placed by a business that was already struggling. Yes, it could have turned the business around, but it could also have dug them an even deeper hole.
We built the case and presented it. They listened, asked many good questions, and decided they were fine with the problems they had.
This is the struggle with leverage points: at the end of the day, finding them is not enough. Because they're so cross-cutting, you will still need to make your case to people whose work will be affected. Some of these people may have never even asked for your opinion in the first place, and some of them may prefer things to stay the way they are, for a million reasons that you'll find very hard to understand. The deeper the intervention, the wider the coalition you need, and the more political the conversation becomes.
Becoming a Better Player
While there's no formula for getting people to act on leverage points, there are a few things we've learned—some of them the hard way.
The place to start is almost always the org chart, not the system. In the subscription project, we spent our opening conversations with the Director of Ecom. That made sense given how the brief was scoped. But the decision we were ultimately asking for required many, many more people—people who hadn't been part of the conversation and weren't that eager to be involved. By the time we figured that out, we were already in the wrong room. Before you start making the case, figure out who the actual decision-maker is, who influences them, and what their agenda is.
Once you know that, don't limit yourself to data, but build a story instead. When you're asking for something this uncomfortable, people need to be able to picture the other side and see a real account of what life looks like once the work is done. How would a more flexible membership model fit within the lives of our customers? What cool things can we do with product recommendations once the data is actually clean? The clearer that picture, the easier it is for someone to imagine saying yes.
But while the story serves as your destination, you still need to figure out the smallest thing you can propose that actually proves your point. A full business model overhaul is a hard sell for a business that's already struggling (ask us how we know). A more contained bet, like offering a broader catalog to a small cohort of at-risk subscribers, would have been a much harder ask to refuse. And if it had worked, the bigger case would have made itself. The more cross-cutting the change, the more valuable a small proof point becomes.
To be clear, none of these will guarantee victory. Organizations have their own, opaque logic, and sometimes—alright, most of the time—that logic wins. This can be frustrating: we have been in more than one room where we felt like we were screaming into the void, while we could have just laid back and done what the brief was asking for.
Despite all this, it's still a battle worth fighting. The whole e-commerce industry is, by its very nature, extremely focused on the short term. We want results, and we want them yesterday. That's what makes UX and CRO improvements so appealing—and granted, they do have their place.
But at the same time, investing in cosmetic improvements in the presence of much deeper problems can lead to a lot of wheelspinning. When that's the case, you need someone with the ability to take a deep breath, find the one thing that matter, and focus all their efforts on it.
If this feels like the kind of work your brand needs to do, perhaps we should have a chat.
