Making software modern - the middle kingdom between legacy and cutting edge

The world of new technologies has never been a calm place. Yet recent years have been even more challenging for those companies trying to stay ahead of the curve. The best proof of this is just how quickly consumers can adopt novelties.

As the Harvard Business Review indicates, it took 30 years for electricity and 25 years for the telephone to reach a 10% adoption level in US households. For tablet devices, reaching a 10% threshold took less than five years.

The formerly-quite conservative world of retail- has witnessed huge change, firstly e-commerce itself, and then the omnichannel revolution; a game-changer in last few years. The first online shopping site was launched in 1979 by Michael Aldrich, and the global online commerce market value is predicted to hit $4.5 trillion in sales by 2021.

Key trends to watch in the near future

Currently, there are a number of key technological trends that every e-commerce business needs to keep a close eye on, just to stay on the same page.

• Mobile dominance - considering the shear size of mobile internet usage, this is of course an underestimate to say the least. In fact, more than half of the world’s internet traffic is generated by mobile devices. Moreover, the mobile-first Generation Z are beginning to drive consumer trends. According to Google, 71% of teens watch mobile video and 51% use social media. The mobile channel itself is getting more fragmented, as companies access their customers with mobile websites, apps and the most recent development in this space - Progressive Web Apps (PWA). These take the best from both worlds, providing next-gen fuel for mobile e-commerce development.

• Artificial Intelligence (AI) - ranging from sophisticated data analytics platforms to chatbots, AI is disrupting the entire e-commerce arena. According to AI Trends, 49% of consumers are willing to shop more frequently while 34% spend more money when artificial intelligence is involved in the sales process.
• Sophisticated loyalty programs - with the big data revolution powering modern marketing, e-commerce companies are developing sophisticated tech to power up the data gathering process. Loyalty programmes, which enable companies to follow customers across channels, are some of the more interesting tools.
• 5G and IoT - there are expected 64B IoT devices worldwide by 2025, starting from wearable tech to smart refrigerators, all harvesting data for companies, including retail ones. All these devices could reach full productivity only when powered by a 5G network, allowing sufficient transmission speeds. This will be a key issue in the next several years.

Of course retail and e-commerce companies face many more technology -  related challenges than those mentioned above. Generally, solutions are aging. For example, Product Information Management systems need constant updating just to keep performance at an acceptable level.

It is a real challenge to keep a company’s online resources up to date, with a constant influx of new trends, and fresh opportunities for reaching customers. As a result, live streams and product videos now need to be standard practice for online retailers to stay competitive.

These commercial trends and the changes to the way we naturally communicate, mean that companies need to be constantly evaluating how to invest in new technology in order to stay fresh, interesting and attractive to the modern - ever discerning - consumer. But that’s not the easiest of tasks.

Challenges - why my brand new online shop is already outdated

The old joke in the ICT business says that no matter how new the computer, it will be obsolete before the new owner gets home, or back to the office. The challenge isn’t only about maintaining performance or delivering full functionality. The key is to build an integrated system that delivers holistic experiences. A silo - based approach, with separate processes for app, website and the offline store is antiquated, as cross - analysis and delivering an omnichannel experience increasingly becomes the new norm for retailers and wholesalers alike.

Also, cutting - edge technologies are rarely fully compatible. At the most basic level, their programming environments are quite separate. Artificial intelligence is most commonly developed in Python, mobile in Java and Kotlin and IoT is going deeper into C and C++ on the device side.

At the core of all this technological chaos there is an e-commerce system that simply delivers. It is far from being a legacy software or antiquated per se, and although a relatively new Magento system or a recent SAP implementation will not be leveraging all of the trends mentioned above, these are still very good pieces of software for supporting business processes.

That’s why there is a need for some modern patchworking in business software. But delivering a patchwork solution can be tricky.

An uneasy solution

Delivering a patchwork software is rarely done by the in-house team of the company, as building it requires a multidisciplinary team with knowledge covering a wide range of technologies. Not only do any add-ons need to be thoroughly understood, but additionally the infrastructure and integration experts need to be fully on board. In fact, augmenting the existing solution with some modern building blocks can be quite risky, and there are various elements that can simply go wrong.

Here are some of the most critical threats to avoid:

• Avoid breaking down an entire system - that’s uncommon but seen. Usually, the breaking takes a form of some critical failure, be that confusing the logistics routes or some serious issues in load balancing, resulting in constant system downtime.
• Ease of further development - as building monolithic systems becomes a thing of the past, delivering software that is future - proof and easy to further augment by new technologies in the future is not a routine challenge. The company needs to design the APIs and policies to maintain consistency throughout the system without losing sight of the business goals. However, delivering the solution in a modular way makes replacing a broken or outdated component much easier than in monolithic systems.
• The solution needs to be optimized - any problems with performance hurt the business, as up to 53% of people will leave the website if the loading time exceeds 3 seconds.
• Technologies need to support each other - launching any of the new features without gathering data, or storing it in a centralized data warehouse is a huge waste of potential. Mobile devices or IoT are supreme tools to gather data which can be analyzed further by machine learning models. Keeping those technologies separate is not an optimal use of resources, including power.
• Thoroughly tested - performing both automated and manual tests to check the systems integrity and functionality is crucial, as seeing users as first-line testers can work when it comes to computer gaming in early access, but rarely brings anything positive when it comes to e-commerce.
• Avoiding hardware investments - although computing power is getting cheaper, it is not free by any means. Thus before scaling up the hardware (cloud or on-prem, doesn’t matter in this case), it is advised to optimize the software to the edge. In the end, the system will perform better at lower costs. Additionally, optimized software provides savings on electricity, reducing the company’s carbon footprint and providing more sustainable growth.

By following the principles laid out above, a company can deliver a system that is as future-proof as it can be, and one which vastly reduces an accumulating technical debt, which would otherwise have to be paid in the future as a cost of modernization and further development.

Summary

Between cutting edge and the legacy software, there is a whole world of solutions that are relatively up to date. A software may well deliver value and support business goals, and yet could be greatly enhanced with the addition of one or two critical components - be that a Progressive Web App or machine learning support.

Leading a total revolution and redesigning the entire system can be counterproductive, as it brings more risk and uncertainty with it. As the story goes, evolution is better than revolution, even if it takes more time.