As 2020 winds down, and we all smile at the prospect of a new year on the horizon, Veriff's team was still busy as ever making improvements to our cornerstone product - our identity verification flow.
During November we were focused on preliminary technical work on new features. They’re due to be released as AB tests during December. In addition to this we implemented a set of technical improvements, which don’t have visual changes, but allowed us to move certain parts of code from our SDKs to our backend.
This allows us to rapidly update additional parts of our SDK without having to roll out a new version of the SDK. We plan to continue with this approach during the upcoming months as this will also help in reducing the size of our SDK.
However, these aren’t the only updates we made, and there are a few which your customers will already be able to notice and feel the effects of today!
The biggest visual update in November was in Assisted Image Capture (AIC).
In our Android SDK we updated AIC, moving it to the end of the verification flow instead of being shown straight after every single image.
Why is this a good thing?
The main reason for this choice is that there are a number of potential customers out there whose internet connection is not that fast. This means it can take a bit of time to upload their images and then our backend systems then need a fraction of a second to analyze it.
When AIC was placed mid-flow then users had to wait after every single photo step for the relevant feedback. Moving the feedback to the end allows the user to move forward through the flow, and while they take their photos, we can compile the feedback in the background, completely unnoticed and not delaying the verification. There might be a brief pause after the selfie, but this image is usually the smallest in size and also easiest for us to analyze.
How did it perform in testing?
As is standard, we ran an AB test to measure the success of the new feature. Results showed that the new functionality had a 0.7% higher conversion rate than the previous order. More remarkably, the change helped to improve the conversion rate by 2.4% for everyone whose first photo during verification wasn’t of good enough quality.
In addition to the AIC improvements on Android we also made a few improvements in our web flow. We updated the visuals that AIC presents to users to help them correct their photos with new ones which better demonstrate the mistakes users are making in real life.
The AB test of this change showed an improvement of 0.3% to conversion rate.
We updated the design of our iOS demo app to bring it to the same design language that we implemented on our web and Android recently.
Our new-look demo app
We’ve also been working on various technical improvements as already mentioned at the beginning of the blog.
Our main focus this month went to improvements around our background video recording feature in our SDKs, particularly on Android. Here, we fixed a set of bugs which were affecting specific devices.
We’ve been constantly working on improving our barcode scanning. This month we updated our barcode scanning feature within our web flow. We were able to increase the amount of camera frames per second we can analyze by five times! (This is on average and from a phone camera). Having a higher frame analyzing speed allows us to detect a full barcode much faster than previously, particularly on the best performing devices. And it helps us get a successful scan on lower-end phones more often.
This wraps up the verification flow updates for November. There’ll be more to come in December, obviously gift-wrapped to reflect the season, but in the meantime you can keep up with us via our social media – Facebook, LinkedIn, Twitter, and Instagram – and make sure to have a happy holidays and enjoy the festive season as much as you can! Happy Veriffing! View all product updates here.