My Process for Product Development [Supplemental]
Seven steps to make better products.
Identify new opportunities through market research, internal workshops, and user interviews.
Create business plans for testing new growth projects.
Build internal stakeholder support for plans.
Work with product, analytics, legal, compliance, marketing, and customer service teams to develop and roll out initiatives.
Vet, negotiate, and manage relationships with strategic partners including banks, card schemes, and payments companies.
Measure and report impact.
If successful, scale growth projects pass the MVP stage. If not, end it…quickly.
Identify new opportunities through market research, internal workshops, and user interviews.
Generating new product/feature ideas in an established organization is relatively easy — people love to talk about their problems. There were quality of life fixes, total system overhauls, and everything in between. The challenge was and has always been eliciting real feedback instead of simply confirming a preexisting bias. That’s where asking good questions and leaving them open ended helped.1
At Flywire, I was working on consumer products which meant collecting a large number of survey results. Prioritizing the list was kept to a small working group (3-5 people) who had enough context and experience to spot what was worth pursuing. These “half-baked” ideas — often no more than a visual and two sentences — were then shown to select teammates for stress-testing and validation before being developed into a full business plan.
Create business plans for testing new growth projects.
Writing a solid business plan was arguably the most challenging part of the job. There were the process and technical considerations along with high-level launch plans. What was most important was the potential impact on revenue or cost savings.
This number is what we knew executives would care most about, and it’s where I spent most of my time. I wrote dozens of proposals that never made it past the working group, let alone to stakeholders, simply because the projections didn’t justify the investment at the time.
Build internal stakeholder support for plans.
Those that did make it past the working group required fine-tuning and lots of shopping around. I would meet with as many people one on one to get buy-in before bringing it to a formal stakeholder meeting for approval. It only took one experience of a “gotcha” question to never be caught off guard again.
Work with product, analytics, legal, compliance, marketing, and customer service teams to develop and roll out initiatives.
Once we had approval, getting all the teams to work on the solution was easy: they had developed and signed-off on the plan already. The challenging part was getting it done in a timely manner. Like kitchen expediter I was running from team to team making sure things were done in the right sequence. The bottleneck seemed to always be wherever I wasn’t looking at the time, leading to lots of “hey, just checking in on…” slack messages.
Vet, negotiate, and manage relationships with strategic partners including banks, card schemes, and payments companies.
In parallel with the internal work, external partners had to be more closely watched and managed. Literal language barriers in addition to a wide range of time zones meant lots of "notes for when you wake up” along with the occasional late night/early morning. In my experience, these relationships are the make or break for a successful and timely launch.
Measure and report impact.
I make sure everyone had access to the data and insights that was most important to them. This meant tailored updates and follow-ups being sent to each of the relevant teams. It took some extra time, but made a world of difference and helped cover me and the team when things didn’t go as expected.
If successful, scale growth projects pass the MVP stage.
Zero to one — getting from proof of concept to product market fit — is hard but procedural. Scaling from there is tough because you’ve got to continue refining and innovating while living with the weight of previous decisions. When it does work, it feels amazing as the veteran team which had gotten the product out the door was well equipped to navigate challenges. Often, we already had a plan ready to go and had just been waiting on a green light from management to go.
If not, end it…quickly.
The less fun ones were the failures and the duds. One stinging failure was due to decisions I’d made. In our effort to get an MVP live, we’d chosen the fastest of the two potential partners. Each had similar capabilities and, at the time, strategic value for the organization. Only with the benefit of hindsight can I see that choice as the key limiting factor to future growth. There were second- and third-order consequences which came to bite us that could have been avoided had we communicated better.
These steps are by no means original: in fact, they’re an adaptation from the job description of my role at Flywire. What they lack in novelty they make up in structuring the challenging process of making new things.
I have a lot of respect for survey writers. They walk the careful line of neutral language and high information density. However, many surveys and interviews don’t require a specialist but they do require tuning. I found that a mix of research and AI to be a good 90% solution. Some of my favorite prompts include:
"Simplify this question: …”
"Rewrite this question to remove bias: …”
"Convert this closed-ended question into an open-ended one: …”