Now we have reached the point at which machine learning (ML) is powerful and mature enough to make a critical difference to value creation. Increasingly, ML is part and parcel of how we grow and succeed in improving people’s lives around the world. It is a mission-critical enabler for us.
A positive cycle
From image-recognition to the ability to interpret text, ML technology has advanced rapidly in recent years. As a data-rich business, we have the fundamental asset – the essential ingredient – to really make the most of this technology’s strengths and potential.
Importantly, a natural positive cycle accelerates value creation – the more quality data you can flow into ML, the better your algorithms will be. Better algorithms make better tools which create better services that, in turn, attract and keep more customers, which creates yet more good data to flow back into your ML. It is a quality and quantity game – one where we have a distinct advantage in our markets: strong local businesses generating volumes of great data for ML. So, for example, when we train open-source image-recognition tools on our proprietary data sets for classifieds, we obtain much more accurate models than otherwise possible. These models, in turn, serve to deliver a more personalised buyer experience and a more streamlined seller experience.
Focusing for greater impact
The key for us is to focus on the greatest value creation. As such, we are firstly ensuring that we design and implement strategies for data – data acquisition, data management and data use as well as sharing inside the organisation.
Secondly, we are focused on acquiring and retaining the right amount of talent in the right places, organised in the right way for our ambitious, amalgamated, high-growth business.
We therefore have a group of data scientists wherever we work around the world, along with a relatively small team at the centre, coordinating the whole, sharing best practice and fostering innovation across the organisation. We implemented this structure during the year and it is working well.
The central team has three key tasks
- To help all organisations in the portfolio activate the tools and opportunities necessary to get the value of ML realised as fast as possible – to accelerate.
- To scale – ensure we use ML efficiently throughout the entire organisation, to serve customers better and improve our operational performance and efficiency.
- To embed ML as a super-utility across the organisation – a horizontal layer of competence and technology that everyone uses, much as we use electricity today. This naturally leads to a new and exciting era of ML-by-design.
Creating value in many different ways
We use ML to create value in many different ways across the group. For example, you can use it to increase the trust and safety of interactions between buyers and sellers, or to make a service simpler and more streamlined. All these individual improvements combine to create greater value.
In Classifieds, ML has a big impact on convenience – making the platforms as simple and as convenient as possible to the end user. For example, OLX uses ML to help buyers identify items of interest. By interpreting what the buyer is actually looking for, the algorithms are able to suggest the most relevant items across the OLX catalogue. ML is also improving trust and safety on platforms, or improving the sellers experience for instance by suggesting the sale price for items.
In Payments and Fintech, ML is supporting advances in fraud detection. We are also able to offer groundbreaking new credit services to underbanked people in India, for example, who have simply not had access to such services.
Our ML tools make it possible for us to offer micro-credit to these customers, which really makes a difference in their lives.
In Food Delivery, we are using ML to increase and enhance automation, improve demand and supply prediction, optimise and personalise search, and ensure faster, more reliable deliveries – all of which makes for happier restaurants and customers and, in turn, fuels the extensive growth of our food delivery businesses.
We are working on deepening and extending the understanding and use of ML across the group so we can move faster and incorporate more advanced tasks.
Training is critical – from education and coaching for senior leaders to enabling a large portion of the entire workforce, not just our engineers, to understand the technology: at any point in time, several hundred associates are participating in ML education programmes across Naspers. Our aim is to capitalise on ML across the group to accelerate the way in which we create value by improving people’s lives. ML is an exceptional tool in our business and we are determined to make the most of it.