Case study

Achieving the near impossible

1 November 2005

Building a mobile bank from scratch in only nine months is a project any company would be hard pressed to find volunteers for. Building a completely new concept in banking is even more challenging and you can imagine it`s something even less people would voluntarily sign up for.

Surprisingly, both Standard Bank and MTN got an awesome response from the handful of experts they approached last year to put together the MTN banking concept. Nine months later there was a completely new cellphone orientated bank capable of changing the market as we know it.

Herman Singh, director of technology engineering at Standard Bank, says his company first breached the topic of working together with MTN about three years ago.

“Those discussions didn`t really go anywhere because the concept of truly mobile banking solutions wasn`t particularly urgent at the time. There was a need, however, to test such a service`s viability,” he continues, “so we went ahead with the implementation of mobile services in the forms of a WAP solution, viewing the new service as nothing more than another one of our channels, just like telephone banking or going into a branch in person,” Singh says.

‘Wait and pay`

The fact that customers needed to download an application to their SIM card in order to use the service seriously complicated issues for Standard Bank however. “It meant that we needed a solution that would work on various brands of phone and brands of SIM card. And because GPRS hadn`t rolled out yet, WAP was a terribly slow experience. “It was re-named ‘Wait and Pay` by those involved in the industry,” Singh chuckles.

Because of the problems with WAP, Singh says Standard Bank rather decided to focus on the ‘lowest common denominator at the time`.

The first port of call was to use USSD, but this too was a less-than-perfect solution.

SMS – way to go

We had common goals. Herman Singh
“So we had a scenario where WAP was simply not ready and various issues existed with USSD. The only other solution was to standardise on a pure SMS-driven solution,” Singh explains.

“Our thinking was originally focused on getting our SMS-based banking application onto every new SIM card going out to the market, but at the same time realised that in order to get into the market rapidly, we would have to work with one network.

“Because of our previous discussions and the fact that the operator was moving into new areas such as expanding into Africa and providing corporate bandwidth, that operator turned out to be MTN.

“Coincidentally, MTN was looking for a banking license at the same time, something that showed us that we had common goals. When MTN got its second-tier banking license we entered into serious discussions again.

“This time, things were a little different. We were not simply discussing ‘another channel` – it was a brand new bank, a new opportunity and the development of a strategic partnership/joint venture in a project known as MTN banking.

The power of two

“It would operate under Standard Bank`s licenses, but use the infrastructure MTN had built,” Singh says.

A little before Standard entered into its second round of discussions with MTN, Singh says the company bought a majority stake in a start-up called Ezuza, who was in the process of building a solution to target the unbanked sector.

“Between our interest in what Ezuza was doing and our discussions with MTN, we realised that a great deal of common ground existed. Through a joint venture with MTN we could concentrate all efforts within Standard Bank, MTN and Ezuza into a single goal,” he says.

In terms of the existing development, Singh says Ezuza was a little different from what MTN banking was. “Between August and October last year, we delved deeply into a technology and product review period, realising that we needed to modify and add to the product offering build within Ezuza.

Pay as you go… banking

He says the solution was aimed at a customer base that could relate to having a ‘pay-as-you go` banking service. “There is no monthly subscription charges, no minimum balance and users will pay service fees on a per-transaction basis.

“The solution is a complete end-to-end service that issues cards, has a firm branch infrastructure, point of sales offerings and all many different channels of communication. This is something parts of our user-base do not yet understand,” he says. “It`s cellphone-centric banking, not cellphone banking.”

As such, it is envisaged that every MTN SIM card distributed after January of next year will contain the MTN ‘Banking` menu and a full set of instructions on how to set-up the service.

“All it will take is a four to five step process,” Singh says, “and the user will have opened a FICA compliant bank account.”

MTN will also include a ‘cash card` with the each starter pack. “After opening the account, this can be used in an ATM immediately,” Singh says.

Not scaled down

“It`s important, however, that the market understands that we have not provided a bank with scaled down functionality, aimed at the unbanked sector,” he says. “We`ve built a full service bank with a focus on both the banked and unbanked markets, as well as one of the most secure offerings on the market.”

Much of the functionality used by Standard Bank and MTN in building the bank, was provided by Fundamo, who the companies first engaged with through Fundamo`s existing relationship with Ezuza.

Best choice

Craig Saks, COO of Fundamo says Fundamo was the best choice for this solution because it`s in the unique position of being able to offer a software package that enables customers to build a mobile bank. “This is not very common,” he says, “in fact, we only have six customers in Africa using different variants of this solution.

“The solution comprised our base solution, integrated into the Standard Bank branch and MTN`s infrastructures. The development took between six and seven months to complete, with testing and analysis taking between two and three months to complete at the beginning and end of the project,” Saks says.

Challenge 1 – tight timelines, no roadmap

“The biggest challenge in this entire project was undoubtedly building the infrastructure and testing the code,” Singh says. “We used rapid application development techniques to get the solution completed on time. This meant development was done in parallel to the finalisation of product specification.”

Saks agrees, stating that the most material challenge in building the solution was keeping in line with the tight timelines. “We had to start working on the project before the full set of requirements were bedded down,” he says.

To reduce the potential complexity of the solution, Saks says the team did three things differently. “We broke the development process down into smaller increments, which allowed us to focus on what we knew best at any one point in time. In many ways, we ended up deferring the unknown, but it`s good practise to let your developers focus on one thing at a time,” he says.

“The second thing we did a little differently, was insisting that we were involved in workshops,” he adds. “Our analysts worked hard to get the development team answers to the questions they had and because they were involved in the workshops, were able to supply further insight to the development team.

“The third was to drive the testing phase into the project as early as possible. When each code increment was delivered, we ensured it was thoroughly tested, to drive out misunderstandings, ambiguities and errors quickly. It also enabled us to incorporate the lessons learnt from each testing phase into the next increment of code delivery,” he says.

Challenge 2 – secrecy and security

Singh says 200 people were dedicated to this project during that nine-month timeframe. “Herein came another challenge,” Singh continues. “To have 200 people working on such a massive project, many different technologies in the mix and over 18 vendors and partners involved, keeping the project secret was a massive challenge. Every member of the team signed a non-disclosure agreement and kept all information under wraps until the project was launched.”

Jenny Hoffmann, CEO of MTN Banking says this secrecy was not only a challenge when it came to building the bank, but was also a major challenge when it came to moving the project from being a project to a fully fledged, stand-alone company.

“MTN banking was a below-the-radar project and as such, we were unable to do any internal marketing or sales. This had to be done in the day before and weeks after the launch of the company,” she says.

Challenge 3 – different backgrounds

“You must also remember that those 200 people came from two different backgrounds, namely banking and telecommunications. Therein lay a further challenge,” Singh continues.

“The banking market is very focused on risk management, compliance and robustness and the telco market measures time to market in the space of days and weeks, not months and years like the banking market does.

“It was incredible for Standard Bank to realise the speed of execution in the telco space and for MTN to see the robust and risk management capabilities of the banking market. It was an awesome combination, but challenging one to negotiate,” he says. Another practise, which kept things on track, Singh says, was the fact that the project team partnered with its vendors. “It had to be this way, since we didn`t have a firm specification for the infrastructure. Because we played the roles of partners, much of the commercial bickering usually associated with projects of this nature, was removed.”

Stable hand on the wheel

He says many CIOs can learn a great deal from the way the development of MTN Banking was undertaken. “The way contracts are put in place and drawn up changes substantially when you are in rapid application development mode. There are massive risks, so the project needs significant executive sponsorship. That was the case from both MTN and Standard Bank.

“It`s a given that there will always be resource conflicts and friction points. It`s imperative to keep a calm head and a stable hand on the steering wheel,” he concludes.