⇓ More from ICTworks

The Simple Way to Solve Your Technology Gender Gap Problem in International Development Projects

By Guest Writer on January 25, 2018

chatbots Gneder Gap At FarmInk, we recently launched a new Messenger chatbot called the Africa Farmers Club. It provides farmers with local trending news about their crops and a way to connect with other farmers in their local area. While we’ve been excited by the speed at which we’ve gained traction over the last few weeks, one thing has been bugging me: our gender gap.

In my previous life I advised mobile money providers on how to better target women in their markets and reduce the, often persistent, gap between male and female users. In some really bad cases the proportion of women using mobile money in a market was as low as 30%.

That’s why I was pretty horrified to see that as our user base was growing so was our gender gap. In the middle of last week we hit the 70% men mark and I knew we had to act.

Why Did We Have a Gender Gap?

The first thing was to understand the route of the problem. I analysed usage data at each stage of the funnel and, to my surprise, found almost no difference whatsoever between male and female behaviour.

Men and women who try out our chatbot are equally likely to finish the registration process (which involves telling us where they are located and the types of crops they are farming). Men and women also have very similar retention levels over time: 82% of men and 81% of women who register are still using the bot on day 7. This gender parity still holds at day 14.

It became apparent that female farmers wanted to use our bot just as much as men. The issue was that we weren’t marketing it properly to them.

Our Experiments to Find Better Marketing Messages

Gender Gap Chatbots

Much of our advertising was through Facebook Ads so I decided to analyse the performance of the ads for different demographics to understand where we were going wrong.

You can see a selection of the best performing ads above with the % of total clicks on the Sign Up button that came from female users.

Analysis of these ads highlighted some interesting behavioural trends. Unsurprisingly the ads with pictures of women in them seemed to appeal more to women. But it’s not as simple as this.

Ads 2 and 4 seem pretty similar, they even show the same woman, but ad 4 worked much better with women. Clearly there was something in the style of the photo and the messaging that made ad 4 more appealing to women.

The other clear message to come through this analysis was that not a single advert saw more than a 38% share of clicks come from female users. Getting the photo and the messaging right was not going to be enough to close our gender gap. In fact the more we continued to use these ads as a marketing channel the more our gender gap was going to widen. We had to change tack.

Putting Women Front and Centre in Our Marketing

Gender Gap Chatbots

I decided to experiment with running a Facebook campaign for a week that targeted only female users. On the Facebook ad platform you can select a demographic to see your ads and so I ran the campaign selecting only adult women in Kenya, Tanzania and Uganda.

I also created a series of ads with a variety of different photos of female farmers and a variety of messages to see what would fly with our target segment.

Above are a selection of the ads we ran…

In terms of cost per click the ad on the furthest right returned the most clicks per $ spent. However, there are other ways to cut ad performance. In terms of overall reach ad number 3 was by far the strongest receiving over 1,000 likes within a couple of days. In terms of comments and shares ad 2 was the highest performer as hundreds of women tagged their friends (or themselves) in the comments.

The least well performing ad was on the far left which, interestingly, was the only ad that didn’t specifically reference ‘lady farmers’ or ‘farmers like you’ in the messaging.

It’s not possible to say conclusively what type of imagery and messaging works best for attracting female farmers and this is most likely because women are not a homogenous segment. What works with one person may not work with another.

However, what is clear is that if you want to attract female users to your service you need to put women front and centre in your marketing.

A few days later I ran the numbers on our split of male and female active users…

Gender Gap Closed

In less than a week we’d (pretty much!) closed the gender gap in our active user base. While we’ll now open up our marketing efforts to both a male and female target audience we’ll continue to monitor these numbers and take concerted action if and when we see a gender gap start to emerge.

Why We Care About Gender Gaps

For those wondering why we’d go to such an effort to target female users, this is not just about fairness.

Africa Farmers Club is a community: farmers share their success stories, their failures, their advice and their questions. It’s important to the quality of our service that women’s voices, as well as men’s, are heard.

As we see from our ad campaigns, women are more attracted to stories and messages that come from farmers who look like them. Early analysis also suggests that our female users may actually be more likely to share helpful information than our male users (more on this to come!).

For these reasons it’s vital to the success of our chatbot that we grow and appeal to a gender diverse user base.

By Georgia Barrie, Co-founder of Farm.ink and originally published as How we closed our gender gap in 5 days


Filed Under: Marketing, Women in Tech
More About: , , , , , , , , ,

Written by
This Guest Post is an ICTworks community knowledge-sharing effort. We actively solicit original content and search for and re-publish quality ICT-related posts we find online. Please suggest a post (even your own) to add to our collective insight.
Stay Current with ICTworksGet Regular Updates via Email

2 Comments to “The Simple Way to Solve Your Technology Gender Gap Problem in International Development Projects”

  1. Eko says:

    Could you clarify, what the actual number of men and women (not in percentage) completing the registration by the end of experimentation?

    • Georgia says:

      Hi Eko, thanks for your question. We typically prefer to measure our user base not in terms of completed registrations but in terms of monthly active users. If a user is active it gives us a better sense that they are engaged with the product as opposed to registering on a whim and forgetting about it. We are not disclosing exact figures at this time but I can tell you that during the experiment we registered tens of thousands of new female users.