The most important metric that any tech company centers around is growth. In B2B SaaS, it’s typically ARR growth. With app-based B2C companies, it’s commonly growth in monthly active users (MAUs), or some other measure of the total number of users that start and continue using the service.
Some app based companies have to balance the pressure of adding users, with the rigor of curating the app’s ecosystem. For some, it’s about platform integrity (bots in a dating app degrade the user experience) and for others it’s about promotion abuse (new signups receive a significant discount). The marketplace/peer-to-peer industry has been wrestling with the fundamental tradeoff between converting every new user that signs up, with ensuring they capture the highest quality, low risk users.
There’s effectively no friction placed on a user that signs up for an email newsletter on any given website. There’s also no real risk of somebody putting in false information. Conversely, getting approved for a mortgage requires two forms of government issued ID (verified by a human, in-person), extensive evidence that you’re financially qualified to re-pay the mortgage, and 4-8 weeks of bureaucracy before you’re approved and you can close on your house.
In marketplaces, the optimal level of up-front scrutiny obviously falls somewhere between these two extreme examples, and will depend on how critical marketplace integrity is to the success of the platform. The more a marketplace has to lose by letting in unwanted users with ill intentions, the more friction they’ll place on new user signups. However, this friction comes at a high cost to conversion rates, and consequently to growth.
The best way to show unwanted users this additional friction, and desirable users a seamless signup process, is by creating a progressive signup flow. In other words, only place additional scrutiny on users that warrant it.
For example, a popular ride-share app had a problem with abuse of their $25 new user promotion. Each phone number (the first question on their signup page) can only receive this discount once, so abusers of this promotion would quickly set up SMS enabled voice-over-IP (non-fixed VOIP) lines to create a multitude of accounts, and ride for free on these promotions. Rather than ask all their users for multiple corroborating pieces of identification, they use Ekata’s Phone Intelligence API to cull out sign ups using non-fixed VOIPs, and only ask those users for additional information (such as email and address) to prevent promotion abuse.
Similarly, a top grossing dating app had too many fake account signups spamming their user base with unsolicited advertisements. These bots were using similar types of SMS enabled non-fixed VOIP phones to pass two-factor authentication, making them look like real and legitimate users. With a simple check of the phone type, these bots are asked to provide information they cannot do in an automated way, which slowed their authentication into the app to a standstill.
Beyond simple rules-based checks on phone line type, many apps use a machine learning model to determine which onboarding flow a user should receive. These models can leverage more complicated signals, such as carrier (there are over 5,000 carriers in the US, making it impossible to write business logic around them), or level the of ownership data available by companies like Ekata.
If email matters more than phone number for your marketplace, try verifying the ownership and authenticity of an email and the originating IP address before asking for additional information. For example, if the IP address appears to be low-risk, the email address is established and connected with the user signing up, do a quick 2 factor authentication of the email and let the user get up and running on the platform. If the email address is not well established, go to a second signup page where you request a phone number and based on the risk associated with that phone number, you can either let the user in, or keep going to another, higher level of scrutiny.
Variations in signup flows will yield varying impacts on conversion rates, false positives, and false negatives. By layering on 3rd party validation of the contact information provided during the signup flow, you can ask for less information from the user, but have more insight into the most appropriate signup flow for their risk segment. Contact us to learn more about our phone, email and identity validation products, and how to build risk segments based on our data insights.