In Ghana MTN has chosen a mixed model of tiers and percentage. To send less than 35 USD costs about 0.70 USD for a complete transaction of deposit/send/withdraw. For amounts over 35 USD a percentage of 2% is charged (1% for depositing and another 1% for withdrawing).

The positive thing with a percentage is that the simplicity of just having one number that need to be communicated instead of 4-7 with a tiered structure. For instance M-PESA’s tariff sheet could look a lot better if they used fixed percentages for deposits/transfers/withdrawals.

The simplicity has some downsides as well. You as the service provider loses a chance to influence the usage behavior of the service. Are you for instance going to market yourself as a service that is cheap on lower amounts for the poorest or be competitive on the larger amounts? With a tier structure the price structure can be fine tuned to attract a specific customer segment.

Another benefit of the tier structure is that’s actually easier to calculate with. It was not mine or Gunnar’s proudest moment when we realized that it had taken us 20min to calculate the total charge of a transfer done with percentage fees. In the end it was simple math but it is still far from the simple adding and subtracting of a tier structure. An example: Carlos want to send 20USD to his mother, or in other words he wants her to receive 20 USD. The question then is how much he need to deposit if the fees are: free to deposit, 1% for transfer and 1% to withdraw?

As it is done in two steps it becomes 20/(0.99*0.99) = 20.406081012141618

As you see this is far from a simple figure. The easiest way for Carlos is probably to just add 1% (0.2USD) two times and then a little extra and deposit 21USD. But I still think it could source of confusion for both customers and agents. If this is the normal use case or if it is more likely that Carlos doesn’t need to know the exact amount the receiver gets is still unknown to us. If that is the case a percentage fee is probably better.

With a tiered structure the customer has some extra cheap amounts built into them. For example if you were to withdraw 15,000KSh from M-PESA in Kenya the normal amount for withdrawing this is 350Ksh. If the customer istead chooses to withdraw 5,000 and then 10,000 the fee is only 275Ksh. That is a difference of 75Ksh which for some can be worth the hussle of an extra withdrawal. If and how common this is also needs to be researched.

Even though a tier structure is most popular at the moment it is far from perfect and I hope we will see new creative tariff structures in the future.

On behalf of GSMA Mobile Money for the Unbanked and together with Caroline Pulver, from FSD Kenya, we have written a report titled “What makes a Successful Mobile Money Implementation?“, looking at the implementations of M-PESA in Kenya and Tanzania and comparing them.

“Can the success of M-PESA be repeated?” is another paper we have written on the same topic but with more focus on Tanzania and which includes findings from our field study.

Recently a dispatch of our project findings was published on the world bank’s CGAP blog. The blog has been an inspiration and a great source of information from the beginning of our work so it is a great pleasure to contribute to it.

The three services are marketed differently depending on the regulatory setup and all of them have different agreements with banks. ZPesa and Zap are both promoted as something more than a way of sending money. ZPesa is described as banking on the mobile phone - a place where you safely can store your money. At the moment there is a variety of payment services that are being introduced, first is the ability to buy pre-paid electricity using Zap, which both M-Pesa and Zpesa are following to implement.

As we already know from earlier research,branchless mobile banking services are still at a stage where most of the users, about 90%, already have a bank account. A figure which also is accurate in Tanzania at the moment. The fact that the early adopters have banking experience is not surprising but hopefully the number of unbanked users will increase. One of the things holding back the adoption in the unbanked segment right now is the lack of awareness of what these services can be used for. The concepts of banking and saving when promoting these services has not yet been communicated. Though many are aware of the services, as brands, they are not viewed as financial tools but as an alternative way of sending money, among both users and non-users, banked and unbanked.

Some users have adopted a behavior that involves keeping savings on the phone. One of these people is Hadija who we met in Kariakoo, Dar es Salaam’s busiest market place. There she is making a living by selling textiles at a stand in the market. The cash from each day’s commerce has to be kept safe and saved in order to re-stock the shop at the end of each week. Her bank is located in the city centre and a visit there means leaving her shop to spend two to three hours in queues. Three months ago she stopped going there to instead visit the M-Pesa agent to deposit her earnings. The agent is located in a shop next to the market and she knew the proprietor from before. However, few of her friends at the market use the mobile phone as a financial tool so she does not represent the average user, but for Hadija m-banking has made a big difference.

Contrary to our initial assumption, we did not see the usage behavior evolve to become more advanced over time, going from only sending and receiving towards keeping a balance. Users like Hadija, who were comfortable with keeping a balance, often got introduced to that functionality when they signed up for the service. Many times that was the main reason they chose to use the service. This shows how important it is to communicate what needs these services are able to fulfill, especially towards the unbanked population.

Right now Tanzania is entering a second phase of m-banking, where the three companies are changing their strategies to reach out more efficiently with their services. What strategy will work? What will prove to be key to enroll a critical mass that will initiate the network effect? It is this being on the verge of really taking off that makes Tanzania an interesting place to study the mechanisms of m-banking today.

Transferring money is an important service that has been improved a lot with the help of the mobile phone. But when it comes to services that help people out of poverty, loans and savings are believed to be more effective. Despite this, and the massive interest from development organizations we have not seen a service that is used or promoted for savings and very few that pay any interest on the money.

CGAP believes savings will come more or less naturally when more payment services are hooked into the m-banking systems. And if we take M-PESA, Zap and ZPesa as examples, they all are shifting focus from money transfers to payments. From what we reckon it is infringements in the payments space and not the savings space that is the reason to why banks react so negatively to m-banking and MMT (mobile money transfers). This probably because it is in payments that the money will be made and not in savings. Surely M-PESA can and is used for saving money by a few but growing portion of the users. This is saving in the notion of a balance or for the use of keeping money safe. For other forms of savings such as savings with interest, loans or insurances one has to turn elsewhere.

So who will offer and focus on savings if there is no interest from either the MNOs nor the banks? As before it will be Micro Finance Institutions. This is what they have done for a long time and what they are good at, so instead of seeing m-banking as competitors to MFIs it should be seen as a way of helping them.

The m-banking services are platforms on which other financial services can build on top of, as Triple Jump are doing with MFIs in Kenya and Tanzania. Isn’t it time to shift the focus from giving grants to build additional platforms to instead  invest in testing if these platforms actually can make the MFI business more profitable and attractive?

Direct deposits works like this; when sending money, the sender finds an M-Pesa agent but instead of depositing the cash into the own account the sender deposits the cash directly into the recipient’s m-banking account. The step of transferring the money between the sender’s and recipient’s account via the mobile phone is thus skipped.

In many cases the sender doesn’t even have an M-Pesa account, he or she always deposits directly into the accounts of others. This behavior is made possible because you do not have to provide an ID when you are making a deposit, and the agents do not check if the account in which you are making the deposit belongs to you or not.

Both Pauline Vaughan, head of M-Pesa in Kenya, and Jacques Voogt, head of M-Pesa in Tanzania, can confirm direct deposits is common and that it is seen as a problem. The transfer fee the users are skipping is the only fee that the MNO do not share with any agent. So, it is the fee that the MNO makes the most money out of that suddenly disappears. To deposit is free with M-PESA, so the only fee the users pay in direct deposits is to withdraw the money. This makes it quite cheap.

Another, and greater, downside of using the service this way is that this method takes away some of the advantages of m-banking. The sender does not get an m-banking account that has extra functionalities and benefits associated with it, such as keeping money safe and paying bills. Also, it becomes cheaper to receive money if you have an account.

From our experience the main reason for direct deposits is not to circumvent the transfer fee, it is only that many people believe that this is the appropriate way of using the service. Once again the source of the problem is a lack of awareness of how the service can be utilized and the benefits of those different uses.Direct

Tanzania has the unique situation of having two m-banking services with the same basic functionality but that has been marketed in totally different ways because of different regulatory deals. M-Pesa that has been marketed only as a way of sending money and ZPesa that has been promoted as mobile banking and a bank in your pocket. This has of course impacted the way agents explain the service and in turn the user’s understanding of it. Both M-Pesa and ZPesa have been operational in just over a year, so the users have had the same time getting used to the service. Still we see a difference in the way they are used. In the case of M-Pesa it was very few that held a balance on their account, most people sent or withdrew the whole amount at once. Z-Pesa users more often held a balance and either sent or withdrew the amount needed. To be fair, one thing that could have effected this result is that ZPesa has not reached the same popularity as M-Pesa. Their user base can be argued to be more dominated by early adopters who might be more open for additional ways of using the service than the average user.

Unfortunately we didn’t have the opportunity to study the usage for a longer period of time so we can not say that the usage does not evolve with time either, we think it does, but slowly. Our point is that it is important how the introduction of the service is designed. There is a lot of time to be saved if the service provider gives examples of uses with longer informative elements during the marketing compared to just wait for people to become comfortable with different types of more advanced uses.

The service provider probably have the biggest influence over the usage. How they educate their agents, how they market the service and what additional services they introduce to the m-banking platform, utility bill payments is an example of the latter. In CGAP Focus Note 46 Gautam Ivatury and Ignacio Mas have identified that marketing is one of the reasons customers use transfers and payments rather than banking services. The reason for this being that transfers and payments are more closely related to the MNO’s previous business model. They are also afraid of regulatory issues if they would step into the field of traditional banking.

There are cases where users quickly have adopted a savings behavior. An example is the Oportunidades, a project for paying social grants in Mexico. After only a handful of payments people began to trust the system, they stopped with withdrawing everything at once to instead start keeping a balance and save some of the money. Surely trust in the system plays a vital role in the initial phase of using a financial service, but it isn’t the only thing hindering more advanced usage such as savings - many times the case is that people are unaware of other functions.