Ask the Pleasure Bot

12 min readNov 12, 2020

Using Pleasure to Reach Young Kenyans with Reproductive Health Information

By Christine Oduor, PSI Digital Program Manager, and Florah Muchiri, PSI Digital Communications Associate

What if you could go beyond a search engine and get honest, validated answers to your sexual and reproductive health (SRH) questions — especially questions about sexual pleasure — right on your mobile device? Imagine the difference this convenient, 24-hour information source could make in the life of youth and young adults, who might struggle otherwise to find reliable, stigma-free information on these topics from other sources.

PSI piloted a chatbot to serve this purpose. By programming it with content about SRH health and pleasure, marketing it, and tracking use, we demonstrated that integrating SRH information with facts about pleasure can increase the engagement of young women and men with this critical health information. We started with the following hypothesis: If Kenyan young adults (ages 18–29) can access sex-positive SRH information privately and on-demand with digital linkages to health services via a chatbot, then they will engage in less risky sexual behaviors and increased contraceptive method use, resulting in fewer unintended pregnancies and sexually transmitted infections.

With funding from UK Aid, we designed the Kenya Pleasure Bot, named “NENA” or “Speak” in Swahili, over the course of three design sprints:

Sprint 1: We conducted preliminary consumer research, a literature review, and market research to verify the feasibility and consumer perception of this project.

Sprint 2: Building on insights collected from young people during Sprint 1, we developed and tested the chatbot prototype. To improve its functionality, we gathered feedback on how young people wanted to engage with it.

Sprint 3: Finally, we launched the pilot test of the refined chatbot over an eight-week period, using a phased-in approach. During the first four-week phase, users interacted with only SRH-focused content focusing on contraceptive methods and menstruation to establish a baseline measurement of engagement and an overall satisfaction rating. During the second phase, the team layered pleasure content into the chatbot and social media marketing communications reflecting as much. To read about the first two sprints, check out “Young People, and The Intersection of Sex and Pleasure” for Sprint 1. Read “What We Have Learned, Building and Testing a Youth-Powered Digital Sexual Reproductive Health Bot,” to learn about Sprint 2. Below, we are excited to share results from Sprint 3.

Sprint 3: Phase 1

WEEK 1

GOAL: Soft launch to test chatbot functionality with a pilot audience and catch any major issues we might have missed during development and initial testing in Sprint 2.

PSI, employing the digital expertise of its Global Services Hub in Kenya, soft-launched NENA on 27 August 2020 to university students through two WhatsApp groups with a membership of approximately 100 students. Links were posted directing students and their peers to the chatbot access channels: WhatsApp, Facebook Messenger and a microsite. This soft launch was used to test the chatbot so that we could address any immediate system errors before the official launch.

Over the first week, 46 targeted users helped to reveal key functional challenges they experienced while interacting with the chatbot. Some users reported issues calculating their next menstrual cycle as the bot did not properly recognize the responses they were providing. Users wanted a calendar within the bot that could help them track their next period and safe days for sexual activity. Thanks to this feedback, the developers had time to address this functionality before going live to a larger audience.

LEARNING: The soft launch is a critical step to take before going live with a larger audience. It provided additional information on areas of improvement on functionality and an opportunity to confirm that the chatbot was capturing certain data points of interest to both the project and the user.

WEEK 2

GOAL: Expand awareness to a larger group within our target audience and drive traffic to the chatbot.

In Week 2, we commenced the Awareness Raising phase of the project. PSI launched content through a digital marketing campaign that built awareness of what NENA is, what benefit it could provide to our target audience and how they could use it.

This campaign was initiated in collaboration with PSI’s independent network member PS Kenya and a Kenya-based digital marketing firm. The local NGO employed its youth social media pages and through the firm, PSI contacted 11 social media influencers (SMIs), all of which created a buzz around NENA and what it is. For the first two days of the campaign, they all promoted a video and two posters. The objective was to introduce NENA as a trusted digital companion with credible information at consumers’ fingertips. The initial days were meant to increase curiosity and ultimately drive traffic to WhatsApp, Facebook, and a microsite on the third day when the links were included in all communications. Below are four pieces of content specifically developed to introduce and explain NENA.

Social media ads targeting youth promoted on Facebook, Instagram, and Twitter during Sept 2020*

PSI held a WhatsApp forum with SMIs and daily debriefs with community managers who reported on analytics, the challenges, and the preferences of social media followers. In working to share the advertisements on various digital platforms, PSI experienced challenges in getting approvals to maximize the application. For instance, users without Facebook Messenger revealed that they were directed to the Google Play store upon clicking through the Facebook Messenger link.

From the SMI’s insights, we also learned that users preferred to engage the influencers in private to either ask personal questions or get more clarification about NENA. The followers also had an interest in knowing exactly who was controlling the chatbot and social media pages. We iterated to maximize privacy for the user.

LEARNINGS: Integrating a new application into the existing digital social media infrastructure revealed unanticipated access barriers for users that must be addressed before any application can go to scale.

• Pushing awareness through other channels like WhatsApp status, which uses SMI’s accounts and offers the user more privacy, does generate engagement. But due to privacy regulations, analytics that could show traction can not be monitored.

Users without the Facebook Messenger app require a special link to prevent them from being redirected to the Google Play store when they click through to access the bot.

WEEK 3–4

GOAL: Tracking traffic into the chatbot — with SRH content engagement

To create a baseline of user engagement with SRH content, the chatbot only provided information content associated with contraceptive methods and menstrual health during Weeks 3 and 4. Information for accessing the chatbot was promoted to our target audience through advertising on Facebook and SMI’s social media accounts. We also developed a microsite landing page that provided three ways, and corresponding links, to access the chatbot. SMI’s promoted the microsite on Instagram instead of the direct links as the platform only allows an individual to include one link per their bio.

Social media posts shared by SMIs targeting youth promoted on Facebook, Instagram, and Twitter during Sept 2020*

Though 2,421 registered unique users clicked on the chatbot, PSI observed a 31% drop-off rate from entering to reaching any significant branch with health content within the algorithm beyond the initial introductions and menu options. Only 1,673 received significant health information from the chatbot.

LEARNING: In order to measure differences in overall content engagement for the next phase of the pilot with pleasure content, PSI needed to set a baseline for the level of engagement with health content only.

Sprint 3: Phase 2

WEEKS 5–8

GOAL: Measuring success by chatbot engagement — with the addition of pleasure content

Over Weeks 5 through 8, we carried out Sprint 3’s second phase of the pilot. During Week 5, we layered pleasure segments into the chatbot and social media advertising for full activation by Week 6. During this phase, SMIs continued to share links through posts intended to drive readers to the chatbot. The SMIs were also briefed and given materials on how to create content around pleasure in keeping with the tone of the new campaign.

Social media posts on pleasure specific content

To gauge whether users would continue to seek out and interact with the chatbot in the absence of targeted advertising, the team suspended digital marketing efforts during the last two weeks of the pilot. The chatbot was monitored to see how many users returned, and how many new users joined organically. In addition, the team continuously reviewed qualitative feedback for insights into user experience, content preferences, and gaps. While PSI saw a rise in new and returning users for the first weeks with advertising, after it ceased, we saw a decline in returning and new users to the chatbot, with the traffic eventually hitting a plateau by the end of the last week.

LEARNINGS: The role of digital marketing is key in driving increased awareness, traffic to the chatbot, and ultimately continued engagement.

KEY TAKEAWAYS AND NEXT STEPS

We kicked off Sprint 3 with straightforward SRH content in Phase 1 and moved on to include pleasure content in Phase 2. The aim of Sprint 3 was to highlight how users interact with the chatbot specifically with the inclusion of pleasure content. We measured these interactions in terms of reach, engagement, and satisfaction rates — with an eye to the sustainability of the chatbot after paid advertising ceased.

REACH: The majority of chatbot users (52%) were between the ages of 18 and 24. However, users outside of this age range also interacted with the chatbot, including those under 18 years old (5%). This is important as regulatory requirements in Kenya impose age restrictions on viewership of certain content for the protection of children and women against sexual exploitation or degradation in cinematograph films and on the internet. Organizations with an interest in integrating pleasure content into more conventional sexual and reproductive health information will need to be attentive to this prior to the development of an intervention, with practical implications for content development and targeted marketing.

We also noted that the chatbot was serving a greater proportion of females (57%) than male users (41%).

Over the course of Sprint 3, we reached 4,867 unique users, of whom 1,871 became returning users. Phase 2 had a much greater proportion of returning users than in Phase 1 (49% vs. 27% respectively).

During the entire Sprint, we reached:

When the chatbot only contained SRH-related content on contraception and menstruation, we reached:

After adding pleasure content into the chatbot, we reached:

ENGAGEMENT AND SATISFACTION RATINGS

Pleasure content had a shorter lifetime in the chatbot over our eight-week Sprint 3 period (three weeks) in comparison to straightforward SRH content (five weeks). However, the same three-week period over Phase 2 yielded a higher engagement rate of 83%. The satisfaction rating was also higher during Phase 2 at 3.9/5 in comparison to 3.2/5 during Phase 1.

PowerBI dashboard screenshots

Across the three different channels on which the chatbot could be accessed, we noted significant differences. The engagement rate — or people initiating a session with the chatbot and receiving health or pleasure information from the algorithm — on Facebook Messenger and WhatsApp was 94% in comparison to 75% on the microsite. However, those using the microsite indicated the highest satisfaction rating of 4.3/5, in comparison to Facebook Messenger users (4.1/5) and WhatsApp users

(3.4/5). In previous feedback sessions with users, privacy was cited as a key consideration for choosing a microsite channel of engagement, which may have been a factor for increased satisfaction with this channel.

PowerBI dashboard screenshots

NEXT STEPS

The inclusion of pleasure content into the chatbot dubbed the Kenya Pleasure Bot has shown preliminary potential in increasing engagement and satisfaction of this information source for target audiences.

To analyze the linkage between engagement with pleasure content and engagement with SRH content further, the project would benefit from increasing the size of the audience using the chatbot through additional marketing. These further efforts could better answer whether clients were more likely to engage with and be satisfied by the health content in a chatbot with the inclusion of topics about pleasure, than if the target audience accessed the health content alone.

PSI would also seek to strengthen the consumer journey from pleasure to health content, with further integration and more complementarity across the content. To better understand the impact on health behaviors, we may also seek to add a follow-up self-reporting mechanism into the chatbot to track self-reported uptake of health services after accessing different areas of the pleasure and health content. Call-backs of a sample of self-reporting clients would allow for verification as well. Building off our implementation lessons so far, we aim to learn how the inclusion and further integration of pleasure and health in chatbot-based tools can provide safe spaces for young adults to access high-quality, curated information privately and on-demand, which ultimately could have a positive influence on health behaviors.

Ask the Pleasure Bot

Written by Sakibdit