In most countries around the world, fewer women use the internet than men. On average, 12% fewer women are online. In the least developed countries, there’s a 31% gap.

Most worryingly, this gulf is getting larger over time – and that’s partly why it’s attracting more attention than ever. In the past year, the OECD, Plan International and the European Commission have all published reports drawing attention to the digital gender divide.

The gap is particularly pronounced in Africa and in South Asia. In India, 72% fewer women than men use the internet. In Pakistan, men are seven times more likely to have a mobile money account than women. Women are also far less likely to own their device. Recent research shows that just 33% of women in India own a mobile phone, compared with 67% of men.

This matters, not least because internet access improves livelihoods. And mobile services can reduce poverty. Research in Kenya showed that access to M-PESA (the country’s mobile money service) was responsible for lifting 194,000 Kenyan households (2% of the country) out of extreme poverty. An estimated 185,000 women switched from agriculture to business.

The causes of the digital divide are complex and multifaceted, but in some countries patriarchal social norms play an important role. Better understanding the barriers to women owning phones and accessing digital services is an important first step, and fortunately, more research on this topic is underway in India and elsewhere.

But influencing norms is notoriously difficult, and sustained change takes years. So where does that leave digital delivery teams in the meantime, especially those operating in countries like India and Pakistan? If they adopt a business-as-usual approach, surely their websites and apps risk exacerbating the impacts of gender inequality, rather than reducing it?

There is no magic bullet. But here are four thoughts for teams responsible for digital services in places with a stark gender divide – and a few ideas for leaders, too.

1. Adopt a gender lens

User research has never been more important – but it’s also essential to adopt a gender lens while doing it. Taking a gender-neutral stance risks reinforcing the status quo. For instance, women who are online may use the internet differently. In India, it’s not unusual for women to use a son’s or husband’s phone, for example – so tying a customer profile to a phone number or SIM card risks excluding some users.

2. Iterate to optimise for female users

Other internet-era tools can be helpful, especially when employed with a gender lens. For example in Pakistan, mobile money provider Jazz Cash succeeded in growing their female base by using behaviorally-informed and gender-specific language, which they tried and iterated through A/B testing.

3. Be part of the change

Delivery teams can also be alert for small but important ways to help shift gender norms. Depicting service users as women, not just men – in user guides, demos, icons and advertising – can help, especially for tasks like managing finances, and running a business, which can be riddled with gender stereotypes.

4. Choose your channels carefully (hint: check the data)

While conducting user research in low-income neighbourhoods in Bihar, India earlier this year, I didn’t find a single woman who would admit to using a social media or digital payments app – even those who owned their own smartphone. “They’re for men, they’re not for women”, was a common response.

Women were also less able to text. In fact, just 26% of adults in India can send an SMS, whereas 80% have made a phone call.

The choices you make about how people can sign up for and use your digital service will influence adoption among women (and indeed among men too). What about voice? Or enabling sign-up and/or using a service via a family member’s phone, a women’s savings group leader, or a microfinance agent? That’s not to say you shouldn’t build a smartphone app; but you should consider offering multiple options. You’ll grow your potential user base in the process.

In this rush toward digitisation, it’s important that delivery teams think of their job as internet-era service design – not just putting things online, and then adjusting (or not) for different population segments. As my colleague Tom Loosemore wrote recently, that means “many paths through one service”.

This shouldn’t just be on teams. Leaders can – and must – help. They can reconceive teams in their organisations that are responsible for an end-to-end service. They can make sure it’s never the case that the “online bit” of a service is passed over the wall to some developers. They give each team the skills and the mandate to own, design, iterate and improve a service, and all the paths through it.

Leaders can strive for a gender balance in their delivery teams – including in the most senior positions. They can ask questions about adoption and completion rates by gender. They can set a goal to reduce the gender gap in service users — but let their empowered team determine the best way to get there.