A change to Instagram's API required us to request new permissions from influencers during their application process. While working on this technical requirement, we chose to also undertake a user experience audit and upgrade.
Our analytics data showed widely variable dropoff rates in the influencer signup flow, especially around the account connection step. We were especially interested in reducing those rates because this was when we would need to ask for additional permissions.
We began by running a set of moderated remote usability studies on our current application experience with two groups of influencers: current or former users and those who had no previous experience with the platform.
We also performed user interviews on a second set of influencers about their experiences with signing up for influencer marketing platforms more generally, as we were interested in learning about industry trends and pain points within this market beyond the scope of our own platform.
Our usability sessions demonstrated that the account connection step was indeed the "toughest" one for influencers. Influencers who had been around longer and used other platforms were often able to move quickly through that step, while newer influencers often became confused around the relationship between a Facebook account and an Instagram account (this also explained why the dropoff rate was so variable for different brands–it depended on the type of influencer they were inviting). Our interviews validated this separation of knowledge and provided us with insight into how we could better coach newer influencers about the specific types of accounts they could connect.
The most significant change we chose to make was adding the option to sign up with email rather than just Facebook. This helped avoid conflict between Instagram and Facebook tokens, especially when an influencer had both personal and professional Facebook accounts. We also clarified the help messaging when Facebook and Instagram authentication tokens were in conflict, and allowed multiple tokens to be active where possible.
We also considered the full UX flow for a user, including possible modes of failure and steps for remediation:
We received positive feedback internally on the addition of email login, especially from the strategic perspective of reducing our reliance on Facebook and Instagram as identity verification. Despite arguably adding complexity to the process with a secondary form of login and asking for additional permissions, the dropoff rate for this step of the signup process was reduced by 7% overall, and was further reduced for outlier user segments who had previously had the most difficulty with the step.
We continued to monitor dropoff rates throughout the application funnel to spot and fix any irregularities. We also began to monitor the use of email login and compare it to the use of Facebook login to see if there was value in the feature. We began to iterate on additional improvements that could be made to the micro-interactions and overall feel during the login process: