Lessons I’ve learned about identity verification in Africa

When I set out to build Smile ID (formerly Smile Identity) alongside Bill Bares in 2017, we had a pretty good idea of the enormity of the problem we were trying to solve, but not much clarity on how the solution would really come together.

I knew from my work in India that as demand for financial services grew, African consumers would need identity verification to be able to access financial products. But the identity landscape on the continent was fragmented, not just by country, but even within them.

At the same time, legacy identity verification organizations also fell remarkably short of the promise of providing adequate solutions for users. Lots of companies were using incomplete or out-of-date data sources or forcing users to submit paperwork in person. We needed to find a way to provide identity verification solutions that were digital, reliable, and fair to all Africans regardless of their location, their form of ID, and even their skin color.

Six years later, as we celebrate the milestone of 75 million ID checks (approximately 5% of the African continent) and the launch of our newest State of KYC in Africa report, my team asked me to reflect on some of the lessons of this journey.

Here goes.

Fraud is rampant, but most people are honest

As we began verifying identities across Africa, one of the things that struck us was the sheer rate of rejected onboarding attempts we saw. On average, around 25% of ID checks we processed were rejected as fraud, suspected fraud, or incorrect ID numbers that were blocked. That’s 1 in 4 potentially fraudulent users that we helped keep out.

However, we started hearing from our enterprise customers that lots of good users were also among those rejections — and they were complaining. They struggled with poor instructions, poor image quality, and high downtimes of the various databases that were used to verify them.

Most users were honest people looking to sign up for products, open accounts, and transact legally. And as we dug into the details, we saw that an overwhelming percentage of the actual fraud we observed came from small groups of fraudsters who repeatedly tried to bypass onboarding security systems.

These fraudsters were typically motivated by financial gain; in most cases, they sought to take advantage of promotional bonuses associated with referrals or signups. For example, one common strategy they employed was to create multiple signups to maximize referral earnings or bonuses. This case study from Nigerian fintech CowryWise highlights one such promotional campaign where certain users attempted to open as many as 25 accounts.

The proliferation of this kind of fraud led to high marketing burn rates and high churn rates as these promotion-seeking users tend not to stick around. But again, the vast majority of users don’t fall into that category.

Biometrics help keep fraud out

As we evolved our offering, we began combining textual checks against ID databases with biometric facial checks to ensure that the presenter of a credential was, in fact, the owner. But we found that as comfort with biometrics grew, thanks largely to selfies, fraudsters also rapidly evolved their methods. Today, the strategies they employ often revolve around a few categories of fraud that we’ve been able to isolate.

Stolen or lost IDs are perhaps the biggest threat to identity verification systems today. When scammers get their hands on a valid ID that they’ve either stolen, found, or bought on the dark web, they try to pass themselves off as the owner of the document.

And when a check is performed against ‘authority’ ID databases or algorithms that analyze documents for security features, the details check out because the IDs are real. And since the fraudster holds the ID document, they can easily provide additional textual information that completes a KYC process. That’s part of why textual verification alone is insufficient for stopping fraud.

Biometric verification, on the other hand, provides a solution here. Using an individual’s unique physical features, we can confirm their identity and ensure that they are, in fact, the owner of a credential. Facial biometrics, in particular, has emerged as the tool of choice since selfie cameras have become universal on smartphones. According to Statista, there were 414 million smartphone subscriptions across sub-Saharan Africa in 2022. And most of those devices are equipped with both front and rear cameras. As a quick data point: 43% of the fraudulent IDs that we caught through biometric verification in the first half of this year were cases where the face of the user did not match the face on the ID they submitted.

In an attempt to beat biometric verification, fraudsters often resort to ‘spoofing’ — a tactic where the scammer uses a picture or video of the actual ID owner to try to game the system. This kind of scam hinges on whether a facial recognition algorithm can tell the difference between a real user and an image or video. But we’ve also built a system of liveliness checks into our KYC solution. These checks typically require users to perform simple activities like smiling into the camera or moving their head in specified ways. And here’s another data point: 41% of the fraudulent ID attempts we caught in H1 2023 contained spoofed pictures or videos.

Digital ID is transforming Africa

As mentioned earlier, when we set out to build Smile ID, the African identity landscape was highly fragmented. Even within countries, multiple IDs were often issued by different agencies, some with overlapping functions, with no one clearly able to define what was an acceptable ID document in many instances.

While that’s still the case in some countries, the last few years have seen marked improvements in identity infrastructure across Africa. In partnership with the World Bank’s Identification for Development (ID4D), more than two dozen African countries have either made commitments to improve or have made tangible improvements to their ID infrastructure. This continent-wide shift has been led by countries like Ghana with the Smart Card, Rwanda with the Indangamutu, and Nigeria with the National Identification Number (NIN), which alone has enrolled over 100 million people. Even Ethiopia, the continent’s second-largest country by population, has recently started enrollment for its digital Fayda ID.

In line with ID4D goals, these new ID systems are focusing on providing foundational ID documents to citizens and residents, i.e. documents that serve as a primary form of identification. These documents, which typically serve as a country’s National ID, also serve as the base for functional ID documents like driver’s licenses, passports, voter IDs, etc.

The most exciting part of these new developments is the overall shift to digital ID systems and the increasingly widespread recognition that African consumers must be given the tools to prove their identities while online. While physical cards will still be printed, most governments are defaulting to digital systems that rely on unique alphanumeric codes assigned to individuals.

Although it’s still early days, the implications of widespread enrollment are far-reaching for inclusion and access. With these identity documents finally becoming ubiquitous, more people who have never been part of the global financial ecosystem can finally open accounts, perform transactions, get credit, own insurance, and lots more.

Final thoughts

I feel a strong sense of pride when I think about what we’ve achieved so far at Smile ID. The numbers validate the vision that we set out to build six years ago. However, we’re also keenly aware that there’s still a long way to go — and we continue to innovate.

For example, while we applaud the World Bank and various African governments that have been investing in increasingly digitized national ID programs, ID authority databases are still a relatively new phenomenon across Africa. And many of them (e.g., Ethiopia, the DRC, Egypt) are still just being built or are not yet commercially accessible. These limitations led us to invest in a broader solution to reach all Africans and the world on behalf of our customers.

Using ‘Document Verification,’ our digital KYC solution that makes up for the shortcomings of ID databases, in conjunction with biometric verification and liveliness checks, we’re currently able to verify over 8,500 documents across 226 countries globally and have the ability to add support for more documents upon request. We also have true depth of coverage across Africa, with support for at least 3 document types for every African country.

Despite the progress we’ve made however, ID verification is still not easy or accessible for millions of Africans. And knowing this fills us with a desire to roll up our sleeves and continue building.

Yes, we’ve come a long way and have now processed 75 million IDs, but it’s still just day one for Smile ID.

To learn more about Smile ID and identity verification in Africa, download the new State of KYC in Africa report.