Gust Core
Problem
Gust.com, a web app helping over 500,000 founders raise early-stage capital, has a core application that acquires thousands of new users per month, but only a small percentage stick around or convert to paid services.
Solution
Run experiments, test using rapid prototyping, and execute using the Agile Scrum methodology. The end product was a guided web app that could tell any entrepreneur how much money they could raise, by what group of investors, and what their next steps were. We called it “algorithmic venture assessment.”
Step 1: User Definition and Exploratory
To simplify, we focused initial experiments on the Quantified Personas driving the most usage on the platform. To get substantive feedback, various websites were designed describing different value propositions. Open questions were asked in one-on-one interviews based on the approved Research Agenda.
Step 2: "Pretotype," Preliminary Validation, and Early Usability
With early feedback, a “pretotype” was built in Typeform based on an algorithm developed by one of our SMEs. The pretotype was used to further test user value, as well as get feedback on usability.
Step 3: UX Design and Wireframes
With insights from the pretotype interviews in hand, a full set of annotated wireframes was designed.
Step 4: Algorithm Refinement and Detailed Specs
Since our initial algorithm was built for a different purpose, we worked closely with an SME to revise it based on the new experience. Additionally, detailed specifications were created to further facilitate the development process.
Step 5: Build
Together with a team of two developers we built a minimal version for early alpha tests. The following is a screenshot of the backlog in Pivotal Tracker.
Step 6: Usability Research and Refinement
Before release, the earliest functioning prototypes were shown to targeted users to perform proper usability testing. This process informed short-term iterations and longer-term roadmap planning.
Step 7: Release, Measure, Analyze, Iterate
Our first releases were a series of limited, private tests designed around a strict learning agenda. Quantitative and qualitative data was methodically gathered and analyzed to inform further iterations.
Outcome
These early tests were designed to prove that algorithmic venture assessment was even possible. The feedback came back positive, with completion rates around 60% (vs. the existing rate of less than 5%) and satisfaction rates exceeding 70%. The module represents the first step in Gust's fundraising user journey, and will serve as a foundation for further core development.