On Sunday 16 June, President Uhuru Kenyatta told a religious gathering at a stadium in Nairobi: “When they see me remain silent, they should not think they are threatening me. I will flush them out from where they are.”
The hunt begins for Africa’s big data
Shoppers walk past empty shelves in an aisle of a Nakumatt supermarket in an upscale neighbourhood of Nairobi. The retailer, which was once among the most promising on the continent, is now buckling under mounting debt – reported to be as much as $145m – while experiencing higher operating costs. Poor management decisions and a weak growth strategy are to blame.
It did not have to be this way for one of Kenya’s most recognisable brands. Experts say the company’s failure to utilise the data it has gathered from its more than 1 million customers is the main reason for its financial problems. “If they had been very data-driven, they would have seen the downward trend maybe two years ago – [and said] ‘We’re not doing well, what’s happening?’,” Francis Waithaka, chief executive officer of Digital4Africa, a Kenyan company that helps businesses use their customer data, tells The Africa Report.
Had Nakumatt managed its data effectively, following the trend in the global retail industry, it would have been able to mine the huge amounts of data being thrown off by every step in the process. From incremental price increases from suppliers, to customer preferences, to information about the management of warehouses, staff turnover and individual store performance, there are thousands of data points to track and analyse.
While the phrase ‘big data’ is often bandied about and is generally taken to mean the use of analytical tools and new datasets to improve productivity, it is not a new idea. Nakumatt, for example, could have learned lessons from US retail giant Walmart, which has tracked the sale of every single item through a system called Retail Link since 1991. In doing so, it cut down its supply-chain costs and revolutionised the US supermarket industry.
It is now easier to track consumer behaviour than ever before through Google, Facebook, Twitter and WhatsApp. So how does this apply to African businesses and banks? At the heart of Africa’s big-data phenomenon is the mobile revolution that has led to 960m mobile subscriptions on the continent. This has created mountains of new information about consumer spending habits.
Across the continent, both homegrown businesses and multinational companies are turning to big data to inform their growth strategies. Data-focused start-ups are also beginning to sprout up. Banks and financial firms have noted this trend and are scrambling to use the data generated by their customers. About 85% of African banks surveyed are using big data to improve security, while 77% are using it to improve customer service, according to a PwC African banking report from October 2016. The International Data Corporation, a market intelligence firm, predicts that revenue from big data and analytics operations will increase by 11% in Africa and the Middle East this year to reach $2bn. It forecasts that growth will remain at about that rate for the next few years.
Mohamed Dabbour, chief executive officer for Africa at Millicom – a telecoms provider that serves 21 million customers in five African countries – says: “It’s clear that the future – in terms of growth, in terms of serving your customers, in terms of understanding them – is going to be understanding our own data […]. We sit on a tonne of data – even when you don’t do anything on your phone there is some data generated.”
Africa’s mobile revolution has done wonders for financial inclusion. Between 2011 and 2014 the proportion of adults with a bank account in sub-Saharan Africa rose from 24% to 34% – a 41% rise – according to the World Bank. The idea that the continent’s commercial banks could further piggyback off this and dramatically increase the number of bank accounts is grabbing executive attention.
Pieter Vorster, chief data officer at Barclays Africa, says: “There are three, four or five times more cell phones than bank accounts, based on the population that exists in [African] countries, so there’s a really enabling technology and enabling relationship [between telcos and banks].”
Telecoms companies are beginning to see the value of this data and to anonymise it and then sell it on. Dabbour of Millicom says he would consider selling data as a new revenue stream. “On an anonymised basis, if we find the right partner and see the value added for the customers and for us, we would probably consider [selling our data],” he says.
Data from telcos offer valuable customer insights for banks. Already known as a breeding ground for financial technology, East Africa is home to many of the most advanced partnerships between banks and telcos. In Kenya, the massive success of Safaricom’s M-Pesa mobile-money service has led to a deal with KCB, the region’s largest bank by assets, to roll out mobile-banking services in 2015. That partnership now has more than seven million customers. Next door, Tanzania’s largest mobile phone company, Tigo Tanzania, has partnered with 17 banks including CRDB Bank to launch mobile-banking services.
“The lines have blurred, particularly in Africa, between banks and telecommunications providers,” says Abdigani Diriye, research manager at IBM Research. “But how do you give a loan with a high expectation of payback, even for only a small value, to someone who has zero credit history and no assets? The answer lies in big data,” he says.
An obvious area of collaboration is determining creditworthiness. For that purpose, Millicom has partnered with JUMO, a South Africa-based, low-cost financial services platform that creates financial identities for small and medium-sized businesses using data from their mobile usage. “We are sharing our data with [JUMO]. […] We have the customer data. But on our platforms, they basically do credit scoring, they analyse the customers, they analyse their behaviours and they give them the credit scoring. And for each of them, they are able to give them a specific loan,” says Dabbour.
Mobile usage data, including airtime top-ups and calling behaviour, can be used to determine credit scores. “Every transaction that your customer does on your network has a record somewhere,” says Dabbour. Using these data can provide insights into the prospective borrower’s financial stability, location, social networks and living circumstances.
Social media is another source of data that is increasingly used by financial institutions to determine creditworthiness. Over the past three years, partnerships between banks and social media companies have grown significantly. But the rules of engagement are complex. “Facebook won’t just sell you their data – it’s not their business model,” says Vorster of Barclays Africa. “But they will co-create with you a mechanism where people identify their information to you legally and you can then use that information appropriately.”
Clouds of data
As the cost of data storage goes down, even more data will be created. A key change in the coming years will be the adoption of cloud computing, an option that allows companies to store their data remotely and access it through the internet. The continent’s biggest companies are already preparing for this. “The big transformation is going to be Safaricom,” says Corine Mbiaketcha Nana, managing director for East, Central and West Africa at tech firm Oracle. “We’re going to announce a partnership with Safaricom. They’re starting to initiate their whole transformation into the cloud.”
And as cloud-computing capacity increases, so will data-rich financial transactions. Some mobile-money transfer outfits are relying on data provided by telecoms operators to function. Strict anti-money-laundering regulations mean companies that carry out cross-border transfers need lots of information about their customers. But mobile-money accounts, which are linked to customers’ identity documents, have been key for the industry, says Rachel Balsham, deputy chief executive officer at mobile payments company MFS Africa. “Our business works because we leverage the ubiquity or the reach of the mobile-money agencies,” she says. “With mobile money, you have traceability, […] you can monitor transaction behaviour.”
Financial advisory companies are also using customer data to help inform investment decisions. “We look more deeply in terms of the client’s behaviour […], more about their purchasing behaviour and that kind of thing,” says Kathryn O’Neill, a senior associate at South Africa-based Aspect Advisory. “For example, in the agriculture industry you can identify potential for loan defaults in terms of the area they are in, whether there is a drought, the types of loans that they have – these sorts of different indicators that you could have a look at. We’re doing a lot of research in that area.”
One problem is that in the rush to amass data companies can overlook the quality of the information. With SIM cards being changed frequently by mobile users, data does not always stay fresh for long. “Big data is going to be relevant only if we’re able to use it in a timely manner,” says Ahmed Rady, Coca-Cola’s general manager for East Africa. “What we’re finding today is data relevance expires much faster than it used to in the past. If it takes you too long to find what you need to find from the data, it stops being relevant.”
Use it or lose it
Quickly making data useful is a job for a new crowd: the financial technology (fintech) start-up brigade. At an old tobacco trading dock in east London, UK, hundreds of software developers, investors and executives have convened to attract investment for their business ideas. Dozens of the start-ups here feature the use of consumer data as a central component of their pitches.
Cashoff, a data-analysis company that helps banks to understand their customers’ data better, is trying to attract new business. The Russia-based start-up crunches spending data generated by banks’ customers. Cashoff sorts this data into different categories, making it easier for banks to identify offers and services that would be appealing to their clients. “Banks can sell their own banking products with a conversion of about 15% or 20%,” says Dmitry Gorkov, Cashoff’s chief executive officer. The company works with 15 banks in Russia and the UK and could soon enter the African market after beginning talks with a bank based in Côte d’Ivoire.
Another data-focused start-up with financial relevance present at the London confab is Cognitect, which builds databases. Cognitect works with traders on risk management and prospective modelling by tracking metadata about transaction flows. “We work with a few hedge funds that use our technology to run simulations of potential market futures and then understand, if any of these scenarios happen or these 10,000 traders do this, what would our response be and then how we model all that based on historical data for future predictive analysis,” says Justin Gehtland, Cognitect’s chief executive officer.
But critics say a data-driven approach can exclude those in rural and low-income areas from the attention of financial institutions. “What is often forgotten in this frenzy around big data is that the unseen in big data are often the same as the unseen in official data,” says Morten Jerven, author of Poor Numbers: How We Are Misled by African Development Statistics and What to do About it.
How to bank those populations remains a hard problem (see page 54), with the poorest 25% of Africans of little interest to lenders. “If you think about letting Barclays take the lead [in using big data] and that will somehow lead us towards financial inclusion […] you would not and should not expect Barclays to do that,” Jerven says.
The proponents of big data do not see it as a silver bullet for development problems. But technology is certainly part of the solution. For example, in 2013, IBM worked with call data released by French telecom provider Orange in order to propose streamlined bus routes for the city of Abidjan. Elsewhere, in Pakistan, according to a report by the Brookings Institution, ‘the government implemented a biometric ID system to ensure that certain government payments could only be collected by women beneficiaries.’ The women with the new ID cards said that their social status improved and they were able to negotiate their family roles better. This then provides a new data set to authorities trying to reach the poorest groups.
Meanwhile at the corporate banking level, the excitement around big data has yet to transform into an item on a balance sheet, says Vorster of Barclays Africa. He sees that there is still a long way to go before data’s true value is fully understood.
“How serious are banks regarding the data that they have? I have yet to find a chief financial officer of a bank that will put the bank’s data as an asset on their balance sheet,” he says. “But people are realising more and more that data is an asset you can use.”