Researching Registration

View project

Image collage showcasing the Leica Biosystems logo and the main registration page.

With the launch of the upgraded Leica Biosystems e-commerce platform in 2019, driving increased registrations on the site was a priority for the team. The numbers were doing well at the beginning of 2020 with the updated registration process from the previous platform, but with the impact of COVID-19, registrations had plummeted as uncertainty had settled in. Once laboratories started buying again, it was imperative that we made it easier for newer customers to buy from us.

Screenshot of the first step of our old registration process.

Pictured above: the original state of our two step registration process.

Initial plan

With the current iteration of registration live, I drafted a short-term plan for evaluative research and looked into user’s current behavior. Getting access to users would be difficult; the pandemic was picking up steam and many of our customers had far more important work than to participate in usability testing…even incentives didn’t arouse interest. Existing customers had already been through this process, so getting a fresh set of eyes on it would be a challenge. However, given that the registration process is very straightforward, I utilized recording capabilities in HotJar to monitor uninhibited behavior. While imperfect and it would not allow any opportunities for think-aloud practice, this could possibly give me an idea of what areas a user would have difficulty with.

Mapping the findings

Recordings had begun to roll in, and key issues began to reveal themselves. I used a Miro board to build out an affinity map to identify the key themes where issues were emerging. One finding was that a “Complete this Later” button on the second step of the process gave users an excuse to back out of the process. Coupled with the button, users were moving from a step that had 4 input fields, to one with 16. That is a significant jump, and could be construed as an intimidating change; recordings had shown users scanned the page quickly and, once aware of coming to a lengthy form, were happy to hit that button and leave to never come back. This was a well-intentioned feature that was backfiring.

Another issue was some users were being drawn to excessive copy on both pages. This copy, added after the initial design went live per a stakeholder’s request to inform users of potential pricing sync issues that may not be instant, seemed to be a point of friction. Without direct user feedback, it was difficult to tell if this was the case, but a number of users lingered on this copy, some even highlighting it or slowly mousing across it.

Password creation was also a sticking point. Prompting users to have to create a password was a stop-gap for several, where 12% of respondents had dropped off in the process at that point.

Also, there was a Google Captcha that had been added at some point. Sometimes in large organizations, changes can happen without you knowing.

Recommendations

Themes in hand, I applied a priority matrix where I was able to identify a number of ‘quick fix’ changes that would resolve some of these issues in the near term. The Google Captcha was removed, as it wasn’t really mitigating any supposed spam issue in the first place. The removal of the “Complete this Later” button was also recommended, and this realized a significant drop in registration abandonment. The second step of the process, in Addresses, was also streamlined to show fewer fields by default and thus appear less overwhelming. The rate of users completing this process improved significantly.

Recommendations for improved instructions for password requirements, which would progressively identify to the user what requirements they needed to satisfy, were implemented slightly later. This mitigated friction at this step by lowering drop-offs from 12% to 5%. Other changes were also proposed, such as removing the header elements and having a more streamlined, minimalist screen. Other changes were also recommended from some of the stakeholders, which helped achieve buy-in for future targeted evaluative research by involving more people into the process.

Importantly, all of these recommendations required a minimum of backend work, and instead just some light front-end touching. The developers were more than happy to take this work on.

Screenshot of the first step of our updated registration process.

Launch & further improvement

After the quick fixes were pushed live, I used this opportunity to work with customer service to get potential customers who were not yet on our e-commerce platform to test the new platform out. While the recordings were helpful, as well as analytics to track the impact of the changes, many labs were getting into a groove around the ‘new normal’ of COVID, and some customers were more amenable to assisting with this research. I ran usability testing sessions with a few of the users that customer service was able to source, and we were able to iterate on this feedback by making some minor tweaks on top of the larger changes I had recommended.

Ultimately, this led to a significant increase in conversions (as of the time of my departure, abandonment had dropped to an average of about 15%/month from a previous 40%/month).