By Pedro Rente Lourenço, data scientist at Vodafone Group Data and AI Centre of Excellence and speaker at Imperial Business in the City: Innovating Through Big Data and AI
Vodafone is focusing on artificial intelligence (AI) and Big Data technologies to grow our business, improve internal efficiencies, and contribute to social good, while having as a first priority the privacy and security of our customers. Over the past three years, our Big Data and AI department has grown from a small start-up venture within Vodafone to have global presence in 24 countries.
As food for thought before the Imperial Business and the City event, ‘Innovating through Big Data and AI’, I considered three things I have learnt from our growth journey about how to create value from investments in analytics and associated technologies like AI and machine learning:
1. Create a compelling offer to attract and develop talent
A key barrier to innovation is the availability of knowledgeable talent with a combination of AI and machine learning skills, business insight, and strategic communication skills. With more and more companies realising the hidden value in their data, it is not surprising that data scientists with these skill sets are in high demand.
As a data scientist, what I value is the opportunity to make an impact through my work by solving meaningful problems with high quality data.
A country with a market size of 12 million people might generate over a billion data points per day on the Vodafone network. With this vast data lake, we address business issues like optimising our customers’ experience of Vodafone products. But we also use our data for positive social impact, through aggregations of large-scale pseudonymized datasets.
I recently returned from Mozambique where I was presenting my work on malaria epidemiological studies. We used Vodafone’s pseudonymised network data to model the movement of people across the country, generating insights about how the disease spreads, which the relevant stakeholders can then use to develop elimination strategies.
Companies should be careful not to neglect the people element of their strategy. What makes your company attractive to a data scientist?
2. Combine short-term wins with long-term vision
Our Big Data applications are delivering personalised offers to our customers across the globe, increasing sales and improving customer experience.
As of January 2019, we are using AI learning to underpin over two thirds of our proactive communications to customers, to drive a predictive, proactive, and personalised experience.
But what is this personalised experience? For example, by looking at the past experiences and interactions of customers who have opted in to share their data with Vodafone, we can recommend the best smartphone for their needs, or the correct data bundle based on their previous usage, which in turn improves customer experience.
As an example of the impact of our personalised offers, in 2018 Big Data applications contributed to the sale of 2.3 billion communication bundles to our South African customers, an increase of 51% from 2017.
Companies can focus too much on long term dreams and neglect the latent potential of existing data which, when exposed to the latest analytics techniques, can shed new light on current business priorities.
At the same time, having a solid long-term vision for data science is crucial to fully unpack the value lying in a company’s data. I truly believe in this ‘double-geared’ strategy to power quick wins coupled with longer term, high-reward projects. Striking the right balance between short and long-term goals is hard, but it is something that, in my opinion, needs to be addressed from the start, to have iterative approaches aligned with a long-term strategy.
3. Integrate data scientists into the wider organisation
Data scientists should not be the smart nerds waiting for a specific request to develop a model. Embedding this function into the business is quintessential to properly drive Big Data transformations. At Vodafone, we believe strongly in integrating data science into the day-to-day business of the company.
To have maximum impact, data scientists need to collaborate closely with business units and have an in depth understanding of company goals to deliver work which adds value.
It is not enough to recruit a team of coders and ask them to play with a vast quantity of data, hoping that interesting things emerge.