What Global eCommerce Taught Me About Building Software

One of the biggest changes in my thinking as an engineer came from working on eCommerce products that real customers depend on.

Before that kind of work, it is easy to think good software means clean architecture, nice interfaces and interesting technical decisions. Those things matter. But production teaches you very quickly that reliability is a feature.

A storefront is more than a website

An eCommerce storefront sits directly between a business and its customers. Every page, button and interaction is part of the buying experience.

If the page is slow, someone leaves. If the product information is confusing, someone leaves. If the cart behaves strangely, someone leaves. If checkout doesn't inspire confidence, someone leaves.

There isn't always a dramatic bug report at the end of those failures. Sometimes the customer simply disappears.

Performance is not decoration

One of the biggest lessons is that performance should be considered part of the product from the beginning.

Images, scripts, third-party services, animations and unnecessary dependencies all have a cost. Developers often experience the website on fast machines, good screens and reliable connections. Customers don't necessarily have any of those.

This matters even more in the markets WebAfro wants to work in. Building something that feels fast on a developer's laptop isn't enough.

The checkout deserves respect

There is a strange tendency to spend enormous amounts of attention on the homepage and treat checkout as an implementation detail.

I think it should be the opposite.

Once someone has decided to buy, the software should get out of the way. Totals should be clear. Errors should make sense. Forms should be predictable. The customer should understand what happens next.

Nobody remembers the checkout that worked perfectly. They remember the one that failed.

Boring software is underrated

The best production systems are often boring.

They don't require everyone to understand an elaborate architecture before making a change. They don't depend on one person remembering a strange workaround. They don't break every time a dependency updates.

They are predictable.

That's something I want to carry into WebAfro. There is enough complexity in running a business already. The software shouldn't create more for no reason.

What this means for smaller businesses

You don't need to be a global brand to benefit from these principles.

In fact, smaller businesses can have even more to gain. If your website is responsible for a meaningful percentage of your enquiries or sales, every unnecessary step matters.

• Make important pages fast.
• Make the next action obvious.
• Don't hide important information.
• Make errors understandable.
• Test the actual customer journey.
• Measure what people do instead of guessing.

These aren't glamorous engineering principles. They are useful ones.

The standard I want to bring home

Working on global commerce products showed me what happens when software has to perform at a serious level. WebAfro exists partly because I want to bring that mindset into more of the work happening around me.

Not by copying global companies. By taking the engineering discipline and applying it to the actual business in front of us.