The GSMA Mobile Money for the Unbanked programme (MMU) has been following the growth of the industry for the past few years using its Deployment Tracker which monitors the number of live and planned mobile money services for the unbanked. While this tool gives us a good sense of the growth of the industry in terms of number of new services being rolled-out, it tells us little about the number of people using these services, and how they use them.
Two years ago, if you googled “number of mobile money customers”, estimates from various consulting agencies would come up including:
- Approximately 100 million active users of mobile money services worldwide in February 2011, then projected to rise to more than 200 million in 2013.
- 133 million mobile money users in emerging markets in 2010 and forecast that that number would increase to 709 million by 2015.
- 45 million unbanked people were using mobile money in 2009, a number they expected to reach 360 million by 2012.
We now know that these estimates were far from the reality: MMU 2012 Global Mobile Money Adoption Survey counted around 30 million active mobile money users and 82 million registered mobile money users. Other highlights from the MMU 2012 Global Mobile Money Adoption Survey are available here.
At the same time, there was also a lack of benchmarking so we had little means to understand relative performance and growth.
The MMU Global Mobile Money Adoption Survey was designed to address both challenges:
- The first objective of the survey is to allow us to assess, in quantitative terms, the state of the mobile money industry;
- The second objective was to develop a tool to enable mobile money providers to benchmark their performance and inform their strategic decisions around mobile money.
In this blog post, I’ll address how we designed the survey and discuss the four challenges we faced during the survey process: participation, transparency, comparability and comprehensiveness.
Overview of MMU Adoption Survey methodology
Every year, we send a questionnaire to providers of mobile money services for the unbanked, asking them to share with us, on a confidential basis, their numbers of mobile money customer accounts (both registered and active), agent outlets (both registered and active), as well as transaction volumes and values for a range of products. For the 2012 edition of the survey, we asked participants to share monthly figures from December 2011 to June 2012. Survey responses were checked for internal consistency, but all data were self-reported and have not been verified independently by the GSMA.
Challenge 1: Driving participation
When we launched the survey in 2011, we thought that driving participation would be our main challenge. Indeed, most of the metrics we were asking for are considered highly sensitive by mobile money providers, in particular their number of active customer accounts. In order to boost the participation to our survey, we offered each participant a customized benchmarking report comparing their performance to their peers both at the regional and global levels.
In 2011, 52 mobile money providers participated in our survey. This was considerably more than we were expecting. In 2012, we were delighted to see a further increase with 78 mobile money providers participating. In both years the sample of participants has been very representative of the industry, with over 60% of the total number of mobile money providers including the most well-known services. Participants have been a mix of MNOs, banks, and third party players from all over the world; some long-established while others were less than a year old.
Moreover, we learned that the benchmark reports have become an important reference and guide for operators. Several participants reported that they had been using our benchmarking reports to negotiate resource allocation for mobile money with their company’s executives and to update their business planning and strategies around mobile money. I believe we had underestimated the need for benchmarks in this industry, and in 2012, we redesigned these reports, adding new dimensions to the benchmarks and increasing their degree of customization.
Challenge 2: Transparency
Finding the right balance between driving participation and transparency was also important. Given the strategic significance of the data we collected, we decided that all data shared for this survey would be kept confidential by the MMU team. Without this guarantee, we doubt any mobile money provider would have participated. We only share aggregated numbers or anonymized data in our public report.
However, we are starting to see an interesting phenomenon whereby mobile money providers want to publicly share how they are performing according to our benchmarks. Last week at Mobile World Congress, four mobile money providers announced themselves as GSMA Mobile Money Sprinters and publicly shared some of their numbers: Telesom, easypaisa, UBL Omni, and Orange Money in Madagascar. We believe that as the number of mobile money success stories will continue to grow, an increasing number of mobile money providers will be willing to share their numbers publicly.
Challenge 3: Comparability
Comparability is a challenge at two levels. First, at the data collection level, we needed to ensure that the numbers we collected were comparable. The best way to ensure data comparability was to impose definitions to survey participants. However, the reality is that not all mobile money providers use the same definitions for metrics such as “active customer account”, or “active agent”. Again, there needed to be some balance between data comparability, and participation levels. If definitions are too strict, there is a risk of potentially excluding a significant number of providers from the survey.
For instance, there is a debate around how to define an “active mobile money account." It is usually defined as an account that has been used to perform at least one transaction within a given time period. However, some people look at a 30-day period, some others at 60 days and 90 days. To accommodate for these different strategies, we needed to ensure we could allow for some degree of flexibility.
The next challenge was about how to interpret these numbers. Indeed, you may have managed to gather comparable numbers from participants, but how do you interpret them? The fact that mobile money providers are very diverse means that they will have different ways of looking at numbers and measuring success: what does success in mobile money look like for an MNO? Is it the same for a bank? With the results from the Global Mobile Money Adoption Survey, we wanted to find a way of comparing the success of the participants. We decided to create a methodology that allowed us to do so, while taking into account the differences among participants. The result of this methodology is the graph below which shows that mobile money is a two-tier landscape.
On this chart, each line represents an individual service and how its penetration has evolved since its launch. On the x-axis, we looked at time in terms of number of months since launch, 0 representing the launch of the service. This allowed us to compare services at different stages of development.
The y-axis represents the level of traction of mobile money. To measure success, we used a ratio of number of transactions over the size of the addressable market. We used number of transactions rather than customers so that we could compare wallet-based services and services delivered over-the-counter.
More specifically, we looked at all transactions excluding cash-ins and cash-outs, which, most of the time, customers perform as a requisite first step in order to perform other transactions. We also excluded airtime top ups as this number is often strongly biased by promotions and bonuses encouraging customers to buy airtime using mobile money rather than other channels, and as such does not accurately reflect customer adoption.
We used addressable market as a denominator. For MNOs, the addressable market is their number of GSM customers. For banks and third-party players, we used the number of unique mobile subscribers in their market.
It was important that numbers were comparable and we built only so much flexibility in our definitions that we did not risk losing credibility of the comparisons. The flexibility was equally important as it allowed us to capture the complexity and variation in this industry and to encourage greater participation.
Challenge 4: Comprehensiveness
The last challenge was to balance all three points described above with comprehensiveness. There was so much we wanted to learn from mobile money providers and it was tempting to design a very long questionnaire with detailed questions ensuring that the survey was as comprehensive as possible. However, one of the key success factors for this Survey was the simplicity of the questionnaire. In addition, the questionnaire we circulated in 2012 was very similar to the one we sent in 2011. The core questions were about the numbers of customer accounts (both registered and active), agent outlets (both registered and active), and transactions (in terms of number of transactions and values for a range of product). The idea was to make the survey as accessible as possible in order to encourage participation.
This year, we wanted to learn more about a particular group of participants – the sprinters who are the fastest-growing mobile money services. To do so, we conducted further research with them on the drivers of adoption. In particular, we asked them questions about their organizational structure, levels of investment and profitability for mobile money, marketing and distribution strategies. It worked well and we managed to gather the additional information we needed to finalize our research on the drivers of adoption.
So this was the solution we’ve found to address the challenge of the comprehensiveness of the data: keep the questionnaire short and simple, and conduct further ad hoc research with a smaller group of participants when necessary.
Finding the right balance between driving participation, transparency, comprehensiveness, and comparability is the key to success for any organization willing to conduct similar data collection exercises at an industry level.
---- The author is a market intelligence analyst at GSMA.
Referring to, "For instance, there is a debate around how to define an “active mobile money account." It is usually defined as an account that has been used to perform at least one transaction within a given time period. However, some people look at a 30-day period, some others at 60 days and 90 days. To accommodate for these different strategies, we needed to ensure we could allow for some degree of flexibility."
What is the CGAP accepted definition of Active Base for the researches that you conduct?
Would be great if you could let me know.