Starting in 2015 and spanning over a year, Eventbrite rebuilt its consumer purchase experience from the ground up. This included a rearchitecting the back-end services that power the purchase, and reimagining the human-facing interactions. It also meant shifting the company's approach from thinking in terms of individual features and screens to a more holistic, user-centric mindset.

My Role

I was the lead User Experience Architect, and led the initial research. Following a groundswell of interest, I was also tapped to lead the early concept work, in close collaboration with Tom Censani, Eventbrite’s head of Product Design. As the project progressed, I would eventually find myself facilitating design work across nearly every UX Architect and Product Designer in the company in order to launch development forward.

Goals

Initially the goal was simply to identify a handful of usability issues in the existing flow. After a heuristic evaluation, the goals were rewritten to focus on building a purchase experience that aligns with how users think about buying tickets, as well as people's behaviors before, during, and after checkout.

Heuristic Evaluation

Before starting this specific review, I developed a heuristic evaluation framework for the team so that we could perform such evaluations rapidly and more consistently across the team and over time. The Eventbrite design team had crafted design principles some time earlier.

I aligned and translated these into heuristics based on an amalgam of those published by Nielsen, Norman, Tog, Schniederman, and other. The end result was a set of heuristics that, while somewhat broad, were approachable and easy to apply or understand — even for those who might be less confident with some of the science or conventions that support them.

Accessibility
Affordance
Brand & Visual Identity
Efficiency & Performance
Learnability & Discoverability
Simplicity
Trust & Transparency
User Control

Performing the Heuristic Evaluation

After validating these heuristics with my colleagues on the UX team, I set about applying them to the set of interfaces and artifacts a typical person would encounter when getting tickets to an event on Eventbrite. Eventbrite’s consumer purchase experience starts with deciding to buy, which takes place on an event listing. A person would then select which tickets to purchase and begin checkout. Checkout could be as simple as entering a name and email and hitting a “Complete Checkout” button, or as byzantine as filling out a hundred fields while an eight-minute timer counts down off-screen. Completed checkouts are punctuated with an order confirmation that scatters one’s attention with competing CTAs and re-engagement hooks, but doesn’t make it clear where you’ll find the item you just exchanged information and money for.

Findings

The issues were many, and the severity often high. Each of aspect of purchase had been designed and developed, historically, in isolation as separate “features.” On one screen, a button might take one form, and on the next it would look and behave entirely differently. Transaction and registration were tightly coupled together, meaning checkout would take longer if an event organizer had many registration questions, and incrementally longer for each additional ticket on the order. Many aspects of the experience had no small-screen equivalent, and those that did were stripped-down shells of the desktop experience — and neither of these felt anything like purchase in the native apps.

I presented the findings of my heuristic evaluation to the company, along with recommendations for the 12 most severe issues, should we decided to fix them one at a time. For many who saw and heard my findings, it was the first time anyone had framed consumer purchase as a single, continuous experience. For many others, it was simply a wake-up call. Too many things were broken in the current experience. Several key stakeholders felt it was clear that we needed to throw out our assumptions about consumer purchase on Eventbrite and rethink the experience and the services that powered that experience.

Rebuild

The company spun up a small team to start hammering toward a solution. I was tapped to join an Engineer, a Product Manager, and our head of Product Design in rebuilding the whole experience, end-to-end. I won't map out the entire experience here, as it's far too large and complex. Instead, I'll spend some time talking about the design process and some key decisions that had significant implications down the road.

First, we identified the need for a purchase experience that was integrated. On Eventbrite, the event listing is the main representation of the event itself. A purchase should not only start from it but should also feel like a continuation of that experience. The approach that we decided to try was putting the entire purchase (ticket selection, group registration, buyer and attendee details, payment, etc.) into a modal. This would allow the purchase to happen entirely within the context of the listing. On large screens, this would be apparent as the

Another key design decision was to create discrete stages of the checkout flow and use those to signal to buyers roughly how many steps remain before they are done. Initially, we employed the metaphor of subway stops and constrained to number of stages to three so checkout would feel simple and quick from the outset.

Independent of how this presented in the UI, this approach would mean building a system that could easily afford us the ability to experiment with re-ordering stages, or even removing stages entirely. This would be important if we ever decided to try separating the transaction stages (which are typically short and simple) and the registration stages (which are highly variable from event to event, and are often so long that they discourage buyers from completing their purchase) — something I strongly advised.

Finally, a critical aspect of any new infrastructure we built was that it should be portable. This could mean a UI that could be transplanted into other contexts to enable a consistent look-and-feel for purchases initiated in event browsing experiences or from external sites. However, "portable" could just as well mean that our infrastructure be service-based, with solid API endpoints that would allow 3rd-party partners or vendors to create their own ticket purchase experiences on their platforms, powered by Eventbrite. This would mean addressing a huge need for any event organizer: greater distribution.

Iteration

This particular project was driven by strong engineering leadership. The benefit to this was an team eager to build something great. The challenge, for me and my product design partner, was to ensure we were able to maintain a healthy design process. In retrospect, we agree that we didn't do as much validation or divergent exploration early on, but the progression in designs was rapid and dealt with highly-complex scenarios and edge cases.

Release & Post-Release

Implementation was a long, difficult process. My engineering counterparts were, in many ways, rewriting the entire codebase that supported purchase. As we progressed, we uncovered new scenarios, edge cases, and states that would require the me to respond quickly and comprehensively with new UX recommendations. At break-neck speeds, development was difficult to keep up with, but the gargantuan effort gradually started appearing in the live product.

The new experience was released via Eventbrite's home-grown A/B testing framework, which would enable us to control how many people would see the new flow and to track user behaviors within the new experience. We started with simple free events, gradually rolling out support for more complex registrations: paid events (i.e. where tickets aren't free), complex registration setups (i.e. when an organizer asks survey questions during registration), group registration, discounts, and more.

As of this writing in early 2016, I continue to support the roll-out of the rebuilt experience, including ongoing validation research and recommendations for iterative design progressions to further improve the purchase experience for ticket buyers.

Reflection

This was my first time leading UX and design for a complete redesign/rebuild of a consumer-oriented transaction flow, and I learned much about typical buyer behaviors. This project also challenged me to maintain a healthy design process in the face of overwhelming complexity and development that isn't going to slow down. I made a huge impact in setting up a consumer purchase experience that is much more pleasing for users, but which also put Eventbrite in a position to experiment rapidly with even more ways to improve the user experience of this critical transaction point on its platform.

I also got to see, first-hand, the power that even lightweight research methods can have in effecting real change.

Consumer Purchase