Giving day 2017 was a transition year for Cornell as we moved to a new fundraising platform: GiveGab. It was a successful year with 8,640 donors making 12,209 gifts for a total of $6,321,962 but a major pain point emerged: post-day gift processing.
Each gift being made to Cornell during a Giving Day needs to be uploaded into Oracle’s PeopleSoft. Before that, the data needs to be pristine, though, and we need to match donors with their current profile.
In 2017, this process was manual and it took about 30 people working full time 5 weeks to check 12,000 gifts through a shared Excel Online spreadsheet.
Processors couldn’t filter or sort the data, etc. We saw typing errors and the process led to redundancies (donors being researched multiple times).
I led some discovery discovery with processors to really understand both the bottlenecks in the process but also all the moving parts: our giving system is a complex ecosystem with a lot of legacy applications and different protocols and file formats.
A handful of in-person interviews were enough to surface process and usability challenges.
Based on this information, I outlined an initial concept for a new gift processing workflow through user journeys.
I socialized this new flow with a group of stakeholders and welcomed their feedback and questions during a standing meeting. This was a great way to agree on a flow without discussing an interface.
The group also started to get excited about the possibilities of the app.
Me and my team went through a couple of rounds of wireframing based on the finalized user journeys.
These were collective sessions where we’d review a version of a specific view for the app and comment/improve.
One of the improvements that we identified was a stronger algorithm to pre-process the Giving data and to start the donor matching process (do we know this person?).
We collaborated with the Business Intelligence team and established the basis for a tiered matching confidence score.
We then scoped the stack and started prototyping in ColdFusion (the language the division has been using for 15 years), feature after feature.
We also worked with our central IT colleagues to build the outputting to Peoplesoft. After a few rounds of testing and debugging, we got our systems to communicate.
The big day was approaching and we organized two training sessions.
We then invited people to review the app in our testing environment while we were finalizing it in production. We caught a couple of things that we were able to improve before going live.
The day went well with $7,827,284.41 raised, 11,748 donors, and 15,805 gifts—a 29% increase in the number of gifts!
Even with this increase in the number of gifts, the process was completed 5x faster.
Hours after the Giving form closed, we fed the data to the algorithm and loaded it in our database. Soon, the processors were reconciling the gifts.
With a 29% increase in the number of gifts, the processing team was able to go through all of them in 6 business days instead of 30.
It meant that the data was available faster for our Colleges/Units colleagues who could then steward their donors much closer to Giving Day.
We were able to improve the user experience for our donors, our university partners, and our colleagues of the advancement team (processors).
Because we understood well the nature of the process and the challenges that our colleagues had faced last year, we used feedforward to deliver a strong, positive experience.
Our colleagues enjoyed working with the app and the process felt simple and easy.
We made sure that the tone was encouraging every step of the way. This process was tedious, last year, and we had to work against this.
The processing team is used to working on legacy applications was UX is thrown in last—if ever. They appreciated these small details.
© 2007, Thomas Deneuville.