Marketing for SMEs: Jevon Le Roux on the E-Commerce Mistake Costing Millions

Online retailers put a lot of effort into getting a shopper to the buy button. But what happens after the sale can undo that work surprisingly quickly.

I spoke with Jevon Le Roux, co-founder and CEO of Keeyu, about why many e-commerce businesses only discover an order problem after the customer complains. We looked at what it means to move customer service from reacting to tickets to preventing the failures that create them.

• Watch on YouTube
• Listen on your preferred podcast app

Introduction

Jeremy: You have a rich work history, and it all connects to what you have created now. Can you give us a rundown?

Jevon: I dropped out of high school to be a professional surfer. I did that for five years, and it taught me marketing, determination and how to get comfortable with failure. You need that in a startup because you need to be comfortable with rejection and failure.

Jevon: After surfing, I tried to launch my first business by building the Ezekiel surf brand in South Africa. It was an absolute failure because I had no idea what I was doing.

Jevon: I then worked for Billabong as a sales agent for seven years before building Nike’s Hurley surfwear brand in South Africa with my own working capital. People said we could not compete with Billabong and Quiksilver, but I had identified a gap in the market. Within three years, we went from last to a leading brand.

Jevon: I later moved to Australia and worked with brands including SurfStitch, P.E Nation and Papinelle. Much of my career has involved building businesses from nothing, helping promising businesses scale by improving their operations, or turning around businesses that were struggling.

The imbalance after the buy button

Jeremy: When I looked at Keeyu and then read your history, it made sense. You had been on the other side of the table, dealing with these problems yourself. How did that lead to the business?

Jevon: You should not build a solution and then look for a problem. You live the problem and then build the solution. That is precisely what happened to us.

Jevon: At SurfStitch, which was a large direct-to-consumer e-commerce business, I noticed how many people worked in marketing and product to attract customers, move them through the funnel and convert them. There were far fewer people making sure customers had a good experience after pressing the buy button.

Jeremy: That imbalance matters because acquiring the sale is only part of the commercial picture. The return on an advertising campaign also depends on margins, repeat purchases and what happens after conversion, which is why I often encourage retailers to look beyond a headline return on ad spend. Our return on ad spend calculator is designed to help businesses work through those numbers before deciding what performance really means.

The order failure that exposed the problem

Jevon: The issue became impossible to ignore when I was CEO of P.E Nation at the start of COVID. We had planned a physical warehouse sale, then had to move it online.

Jevon: A large burst of traffic caused the website to stop syncing properly with the warehouse. The website kept accepting orders because it believed stock was available, and around 1,000 customers ordered products that were not actually there.

Jevon: We had to call those customers, issue credit notes, apologise and hire extra people to handle the work. My co-founder Tracy said that if she could have seen the problems happening in real time, she could have prevented them from becoming a nightmare.

Jevon: That was when we realised customer service was built backwards. Brands often measure how well they fix a problem after a shopper complains. We thought they should also measure how well they prevent the problem from existing.

Why order problems stay hidden

Jeremy: What causes so many online shoppers to reach the point where they have to ask, “Where is my order?”

Jevon: After the buy button, many systems can touch an order. There may be a storefront, payment gateway, warehouse system, order and inventory management, carrier integrations, delivery notifications, returns software and a help desk.

Jevon: Each system performs a specific function, but they are not necessarily connected to the customer promise. That promise might be: order before 2 pm, pay for express delivery, ship the same day and receive it tomorrow.

Jevon: When something fails, it can fail silently. The brand may only find out when the customer says the promise has been broken. The service team then has to search across systems to understand what happened and fix it manually.

Jeremy: This is a useful distinction for any online retailer. Bringing more shoppers into a store through paid advertising does not resolve weak operations behind the sale. More demand can make those weaknesses more visible.

From reactive support to proactive detection

Jevon: Keeyu connects those systems through an orchestration layer. It does not replace the existing tools. It sits across them and links their activity to the promise made to the customer.

Jevon: If a failed payment, stock problem, system sync issue or lost parcel threatens that promise, the platform identifies it in real time. It then selects and launches an appropriate automation so the issue can be addressed before the customer needs to complain.

Jeremy: I thought of the orchestration layer like the conductor of an orchestra. The conductor does not play every instrument, but coordinates the whole performance from beginning to end.

Jevon: That is the difference between reactive and proactive artificial intelligence. A help desk or reactive AI can answer a question after a ticket exists. The proactive model is intended to prevent the ticket.

The business case for preventing tickets

Jeremy: This is already operating with paying customers rather than being only a concept. What results can you discuss?

Jevon: We have worked with brands including Puma, Helly Hansen, Quiksilver, Billabong and EHP Labs. The product addresses the cost attached to customer service tickets by reducing the need for some of those tickets to occur.

Jevon: EHP Labs reduced its tickets by 55%, which translated to an almost half-a-million-dollar saving over 12 months. Based on what it paid us, I estimated that as roughly a tenfold return on investment.

Jeremy: Jevon’s example is a reminder that customer acquisition cannot be assessed in isolation. Our five-year retailer growth case study also shows why sustained e-commerce growth needs to be viewed over time rather than reduced to one campaign result.

Complexity matters as much as order volume

Jeremy: I assumed order volume would determine where this type of system makes the largest difference. Is that right?

Jevon: Volume is important, but complexity is often more important. A small Shopify store operating from a garage is not our customer. A help desk or returns solution would usually be a more appropriate first tool.

Jevon: The need becomes stronger when a retailer has a customer service team, multiple storefronts or countries, marketplaces, social selling channels, click and collect, ship from store and more complex fulfilment arrangements. Each added element makes the customer journey harder to monitor across separate systems.

Building a category people do not yet understand

Jeremy: You are presenting retailers with a different way of thinking about customer service. Has that made the product difficult to explain?

Jevon: It has been a journey. We are showing people a world they have not experienced before, so the first hurdle is understanding the concept. The second is believing the return on investment, and the third is deciding whether the business has the capacity to implement it.

Jevon: We have worked to simplify our language and reduce the implementation burden. But building awareness of a new category remains hard work.

Jeremy: You also mentioned that your team is now on the fourth version of the product. That learning, the integrations you have built and the way the system learns from each business all contribute to its defensibility.

Jevon: We have made mistakes, particularly with the AI automations, and this is our third path in that area. We had to build them specifically because there was no suitable solution available out of the box.

The founder’s job is partly storytelling

Jevon: When we started around two and a half years ago, many people did not understand what we were doing or believe there was a problem to solve. But the three founders had worked in e-commerce for decades and could see it.

Jevon: Customer service exists because things go wrong. We believed there should be an equally significant category focused on preventing them from going wrong.

Jeremy: The idea feels clear once it has been explained, but arriving at that explanation took time.

Jevon: One of a startup founder’s key roles is storytelling. My number one role is getting better at telling the story.

Jevon: We now compare a help desk with a hospital emergency department. An emergency department cannot prevent a heart attack; it can only triage the patient after something has happened. In the same way, a help desk manages a ticket after the problem exists, while our aim is to prevent the ticket.

Practical takeaways for online retailers

• Do not stop assessing the customer journey at the buy button.
• Map the systems that touch an order and identify where failures can remain hidden.
• Measure avoidable complaints and operational failures, not only response time after a ticket is raised.
• Consider complexity across storefronts, markets, fulfilment methods and service tools, not just order volume.
• If you are building a new product, start with a problem you have experienced and understand deeply.

You can learn more about Jevon and the platform at Keeyu, or connect with Jevon Le Roux on LinkedIn.

---

This conversation is part of Marketing for SMEs with Jeremy Yang. Watch the full episode on YouTube or listen on your preferred podcast app.