ICTworks

An extract from Strong Connections

As we wind our way slowly through the market in Accra, we stop to visit several clients of Opportunity International Savings & Loan (OISL), a full-service Tier 2 bank that received its banking license in 2006. The Bank, as it is called, has recently run an IVR—Interactive Voice Response—campaign.

In fact, about two out of three of OISL customers have some form of literacy constraint. Either they have not had the chance for significant formal education or their preferred communication is in something other than English (the main written language) or Twi (the main spoken language). As in India, women in Ghana are overrepresented with literacy challenges.

These circumstances make IVR seem attractive. It is a technology that allows a user to interact through their voice and/or with the keypad on their device. In Ghana, we are attempting an IVR intervention to engage the clients of OISL through a series of voice messages in local languages. With all of the literacy challenges, we think that voice-based messages will be more effective in engaging users with the banking system.

For the IVR work, three local languages were selected: the main local language, Twi and two additional regional languages. (The bank will later add two additional languages in future interventions, moving the total to five.) The nature of IVR allows for either simple broadcast messages or an interactive format where the client is instructed to press a number, say one, to hear the second part of the message. Because many women, like Afia, only own a basic phone, they need to press a number to interact, and these automated responses are designed to be used by a group where the majority still have basic phones.

The technology platform for IVR allows the staff at OISL to monitor the calls and to see which are the most successful. They can also identify the calls that result in people having to listen to additional messages. When the IVR results are connected to the financial activity database, the team can measure the increased savings that the clients are putting away.

The Voice of Adjoah

The team at OISL tested several different types of messages and found that there was a more positive response for messages with a woman’s voice rather than a man’s. The previous December, a female staff person recorded the first message and called the character Adjoah, a common Ghanaian woman’s name.

Thus, although it was clear the message was from their bank, it was Adjoah who wished the clients Merry Christmas. After that, a series of automated voice messages were crafted to take the clients on a journey to understand and encourage them to save more money.  Messages were differentiated depending on how the customer responded.

The responses in our focus group reflected the users happiness with the messages:

• “Adjoah is very friendly.”
• “I liked the encouragement to save from Adjoah.”
• “I am so happy Adjoah calls me every week, like my sister does,”
• We immediately notice that all of the women reference Adjoah; they seem to like receiving messages from her.
• “Would you prefer a woman’s or a man’s voice for future messages?” we ask.
• “Definitely a woman’s!” exclaim several women almost in unison.
• “If it were a man’s voice, my husband might get jealous,” one customer says and several women nod in agreement.
• “Where’s Adjoah? We wanted to meet her,” says one woman dressed in bright yellow.

How would you describe Adjoah?

This launches an interesting and, for me, completely unexpected discussion on the characteristics of a person whom the women in the room consider reliable and trustworthy, and who reminds them of a close friend or sister. I am again reminded that in testing new technologies and digital interventions, it is hard to predict how they will be received. In the US or UK, many mobile phone users would be cynical about receiving voice-based messages from what is essentially a chatbot, ignoring or deleting them, and certainly not associating the character who voiced the message as a friend or trusted counterparty.

We encourage the women to tell us how the messages have affected them, especially around the encouragement to save more money. One woman tells us how the staggered messages helped her set up a goal of purchasing another fryer for her prepared food business and then plan out how much she would need to save each week.

She said the reminders were important, not because she would forget to deposit the money to save but because it allowed her to inform her family that Adjoah and the bank were encouraging her to save—which helped her fend off social pressures from family members wanting her to buy them something or lend them money using her extra earnings. The social pressures to support other adults in extended families is similar to what is faced in East Africa, but often not well understood by staff who did not come from modest beginnings.

Afia, a long-time bank customer,  relays the story of her conversation with her friend Mawusi. She describes how the messages make her feel that the bank OISL cares about her. She also laughs as she tells us of how Mawusi seems a bit jealous. “Mawusi did not receive any calls from Adjoah,” Afia says.

The very congenial, competitive nature between Afia and her best friend seems to be key to the success of the IVR messages. They represent a way to differentiate service and to make OISL’s customer outreach have more premium value.

Adjoah is caring and compassionate

Awesi, an older woman dressed in bright blue, echoes Afia’s sense that the messages make her feel as if the bank cares about her. After missing two regular group meetings recently, Awesi received a call from Adjoah that noted her absences and asked her to contact the bank if there was something wrong. She went to the bank branch in the market and told the manager that she had been in a car accident, but she was now recovering well. With knowledge of the car accident, the credit officer at the bank was able to reschedule Awesi’s loans, which enabled her to remain in good standing.

For her, Adjoah’s call initiated a series of steps that made her feel that the Bank was a compassionate institution with an understanding of her circumstances. Awesi is unaware that Adjoah’s messages are all prerecorded and that an algorithm prioritized the messages she received based on her lack of attendance. The intervention is fully automated and requires no staff involvement once launched.

For this group of women, representing the clients of OISL, there was a profound impact from the set of messages, and it is clear there is a strong identification with Adjoah as if she were a real woman, a friend. Unlike many digital services where the descriptions are often words like “efficient” or “modern,” the women describe Adjoah’s automated, recorded voice messages as “caring” and “compassionate.”

Scalable IVR Solutions

I scribble in my mostly illegible notes “scalable,” which I underline twice. Unlike previous outreach efforts that depended on staff, contractors, and materials such as printed flyers and workbooks, the work of IVR is conducted digitally with no printed materials. Messages can be prerecorded and sent to thousands of clients at once with a very low marginal cost compared to more manual interventions. One IVR intervention in India that delivered health messages reached almost two million people.

I am also reminded that for women like Afia and Awesi, the voice-based messages have additional elements that are attractive compared to text messages. To them, voice messages seem familiar and are infused with emotion; to the women being called, it sounds as if the caller is someone they know and might regularly interact with. In addition, since the messages are tailored to the individual, navigating a decision tree that is influenced by their actions, the voice-based messages can leverage some of the best lessons of behavioral nudges. In other words, they are structured to encourage the best outcomes for each listener.

For this project overall, the aim was to increase the women’s savings. The trial has been successful, as on average, each person participating did increase their savings balance compared to a control group who did not receive the messages. And the messages also seemed to encourage listeners who had been relatively dormant to re-engage with the bank, something that could result in substantial value to the institution if it led to lower attrition rates. The surprise is the resulting positive feelings generated by the voice message, feelings that can be leveraged to encourage customer engagement and loyalty.

This is an extract from Strong Connections: Stories of Resilience from the Far Reaches of the Mobile Phone Revolution by Rosa Wang, former Global Director of DFS for Opportunity International