Many articles have been written and much research was done among low education, low technology users in Africa. Most of these users are rural dwellers and organizations working with them are mostly focused on agriculture, financial literacy, credit facilities, insurance, or off-grid pay-go solar solutions.

This article gives an insight into how SMS information delivery works among this user group and what factors would influence action and behavior change. It dives into my experiences that have been drawn from many African countries, Southern (Malawi and Zambia), West (Nigeria), and East (Kenya and Ethiopia)

I have spent some time in leading research, user experience testing, SMS content testing among these users, mostly smallholder farmers who are the majority in rural Africa.

You see! Like all users, rural Africa users will only use products that they need. This would look like a general statement but, it is not. Many times, we sent our SMS messages to our users but there were no responses or any acknowledgment. Any action from the user would give us an indication that they are reading the content, understanding it and this would infer impact.

You wonder why?

My understanding of this user group is that the SMS delivery rate is a challenge, as I have stated in my previous article, due to network coverage issues, phone charging challenges, most of these users will go offline immediately they leave an urban environment.

Most organizations would work in urban areas to enroll these users, then afterward it is almost impossible to reach them. The SMS delivery rates have mostly been ~50% or even less across many African countries I have worked in. Again, even if it was delivered it is hard to know if the user read the message unless there is an action required in response to the message such as dialing a USSD code or replying to the message. I will go into more detail on the one-way vs two-way SMS technologies in my future articles. 

Now, through SMS, you can reach about 50% of your users, this is a decent % if you think about it. However, the question will now be, how many users would respond or take the required action in the message content. It will surprise you to find out that this may be even range between 0-2%. The response rate will vary depending on the information being delivered to the user. You may send 1 SMS and receive many responses and sent the next one to the same user group only to receive few to zero responses. What is the difference here? What makes the same user groups respond differently to different messages?

I will try to answer these questions giving insights on what I think would differentiate these messages and influence user behavior.

Need/Timing of SMS

I have spent the last 5 years working with organizations that based most of their end-user communication on SMS. Difference messages have generated different responses and changed the perception of users. At Totohealth Ltd an organization that works in the maternal space by sending informative messages to expectant mothers and those with children under the age of 5 years. These messages are tailored to suit each user and much work goes into customization.

In this space, we received more responses compared to other industries. You may ask Why? Look! This is a very sensitive time for every mother and they would be very keen especially to learn on what is happening to their child if a message focuses on for example development stages of a child, they will be keen to establish if their child has hit that specified milestone in the message. They would use the messages as a guide to be raising their children and be very responses in case they feel like something is going wrong. We also experienced a response surge both via SMS and calls when a message was sent with advice on an outbreak in an area. The users would call back to try and understand how to keep safe. This happened in Kibera slums back in 2015 when there was an outbreak of cholera and we were working with the local clinics to send out messages to help our users to stay safe, more so because they were either pregnant or had children under the age of 5 and they would be more vulnerable. The behavior changes, in this case, were directly impacted by the need for the content for these users. The content touched on a fundamental part of their lives and they would take it seriously, read, understand, and respond where applicable.  In the future, I will write about the customization concepts we used to personalize the experience for each mother.

This finding was also consistent with what I found out working with smallholder farmers. We would send great messages about the best agronomy practices and users would not respond to a level we wished, even when asked to call our call center for more information, this rarely happened. This behavior changed whenever there was a need! 

I remember this particularly happened when armyworms attacked maize 2-3 months after planting in Blantyre district, Malawi. As part of our agronomic advice, we sent out messages giving preventive advice to the farmers and giving them information on where they can get pesticides. Immediately the messages went out, we received responses mostly through calls to our call center asking for more information on pesticides, where to buy, which ones are effective, etc. Remember this was the same user group we were constantly trying to engage with limited success but now that they needed something, their response and engagement improved immensely. Thus, to deliver impactful messages it is important to look at the needs of your users and target that.

SMS Content and Wording 

The way an organization words the message is very key to both inform the users and guide action and behavior change. In our aim to get the best wording, we tested several messages among users, changing words, punctuation, and the language of delivery. It will surprise you to find out that one word can change the action that a user takes.

Another complication of wording comes in when trying to adapt to different languages across the African region. There are many local languages and message customization is a big deal. For example, in Ethiopia, there are companies that have been registered by the government to do the translation to Amharic, it is not possible to ask a colleague to translate for you. These companies are required to sign that they will be responsible for any miss-interpretation. I will tell you about my experiences in Ethiopia soon! Different country. 

In trying to understand which words we would use for our messages. We tested our message content farmers in Kaduna and Kano states in Northern Nigeria. We showed different messages to our users trying to understand which ones they understood. We gathered information by observation and asking questions.

In one of the messages, we had referred to a character Jacky, to us this was the name of someone. We observed that users would read the message in their language (Hausa) and smile, others laughed. We could not immediately establish why the users were behaving this way only to find out later that the word we used for a person meant “donkey” in their local language! Imagine this message, “Jacky bought product X and got a great harvest! You can do the same!” to them, this message read “Donkey bought product X and got great harvest! You can do the same!”

I submit that there are no universal messages, a message can work across countries. Therefore, there is a need for organizations working within these communities to really understand the local context, words that may be okay for one user group in a different country would be understood or interpreted differently in the next. Religion is also very key in the words that an organization can use, lets the organization is labeled disrespectful and users demand to be removed from the specific projects or programs

Segmented User Group Testing 

As stated in point two above, testing the SMS content among users is a crucial process in refining the messages to fit each user group. A message may resonate very well with users in the eastern part of Kenya and be offensive to users in the North. Thus, targeted testing of the message content among segmented user groups is key. In my experience, showing and testing messages with a smaller group of users before launching to the entire user base is important.

Testing will also give an organization a better understanding of the users and would make them tick, which words if used would influence them to engage more. Our testing in the past utilized the AB test methodology. Here, the same user groups would be exposed to different messages as the 1st step. We would then move to the 2nd step where we would take the best message in round 1, iterate and split it, then target other users in the same user group. The customization will continue until we get a message that delivers the best response based on our performance matrix. This is the message that we would use for the targeted larger group. The test and re-test may look like so much work, which is, but that would what is able to differentiate between 0% response and 50% response. Depending on what the organization needs to achieve, either triggering a response or informing the user, testing can take a different dimension. 1st the wording is tested on a face to face interaction, focused group discussion or interviews and 2nd the users are sent the messages in a controlled environment as described above.

Incentives and Reward Mechanisms

An incentive for users is a big discussion, some people say that if you must incentivize a user to use your product, then that product is not working. I partly agree with this statement, but it really depends on how one defines an incentive, most people think about it as a reward. I would wish to define it as a form of self-worth or appreciation attained for doing something.

If you are keen, you will establish that an SMS content may by itself have in it an incentive even without attaching monitory value to it.

We once promised our users 0.02USD vs 0.05USD/action. The action was to refer a friend to us. The message was the same, but the promise of airtime was different, this was in Zambia. There was not much difference in the response, but the same test generated a completely different result in Kenya and Malawi. The higher the airtime incentive, the higher the response rate, positive linear correlation! Thus, an incentive for one user group would yield completely different results compared to the other. Again, testing will be able to clearly show what kind of incentive matters to different user groups.

Delivery of content through SMS technology would work. However, it would be of great interest to the organization to keenly look at the above four factors. Is the content of the messages needed or required by the community or targeted user groups? Is the wording, right? Are the nudges in the messages able to deliver the message as needed and change the expected behavior? Is the translation right for the message? Has the message been tested in a segmented way among user groups, what do the users care about? Is the incentive, right? Unless all these questions are answered satisfactory, I submit it would be very difficult to try and create any impact using SMS technology. Unless the SMS is only to inform and does not require any action on the side of the user.

These questions are important in trying to come up with content that can be delivered through SMS. SMS is limited in many ways including the number of characters allowed per SMS, not being able to express emotion clearly, not being able to share pictures in the messages. Considering that most users in rural Africa use phones with limited screen space “mulika mwizi” or feature phones. In trying to solve these challenges, I will look into new technologies including WhatsApp for business and other platforms that are currently being used to deliver content thus is able to generate the required action, influence behavior, and impact rural African communities in my future articles.