Jack Pratten needed ten Spanish speakers. He got a business lesson.

Jack Pratten went looking for ten Spanish speakers on the New York subway. A friend’s advertising agency needed them for a tequila-brand copy study. He found willing recruits, brought them across the city, and discovered that only three spoke Spanish. A promising head count had become a rather less promising research sample. New York had supplied the people. It had declined to supply the screening.

The episode belongs to Respondent’s beginnings in 2016. It also supplies a useful way into Pratten’s career: the gap between finding someone and finding someone suitable. Much of the work that followed would occupy that gap. A research interview can have a booking, a budget and a beautifully prepared discussion guide, yet still fail before the first question.

A farm, a camera, an inconvenient answer

Pratten grew up helping on his father’s farm in Outback Australia; his mother was a landscape designer. Film came later, including work with his sister on a South African project involving Nelson Mandela. His own venture acquired, edited and sold camera-phone footage. Research into expanding it exposed a shortage of problems the product actually solved. The team began winding the business down.

He describes the early building instinct simply: “We were incredibly resourceful.” That quality can get an enterprise started. Knowing whether to keep building requires a different discipline. His next company would put access to other people’s experience at the centre of the product.

There is an appealing lack of ceremony in this sequence. The useful answer arrives after money and effort have already been committed. It refuses to admire either. For a founder, a customer conversation can be an expensive interruption to a cherished plan. It can also spare everyone the expense of continuing it.

The trouble with a private little list

By August 2018, Pratten was setting out a more systematic argument. In his writing on research marketplaces, he described an industry dependent on privately assembled participant lists. Researchers had invested in those lists; access could become a competitive advantage. The information was also fragmented between spreadsheets, computers and specialist collections. Finding a particular kind of participant meant navigating a patchwork of other people’s arrangements.

His proposed marketplace would widen access and make verification part of finding people. The important ambition was practical: give a researcher a route to a specialist audience without requiring a fresh recruitment operation for every project. A designer or product manager could then spend more of the available time conducting the interview.

Read as a statement of Pratten’s business interests, the argument is straightforward. Read as an account of wasted effort, it has wider appeal. Organisations can be remarkably sophisticated about analysing a conversation and remarkably improvised about getting a suitable person into one. The spreadsheet has a talent for becoming permanent precisely when everyone considers it temporary.

“I started recruiting research participants on the New York City Subway in 2016.”

Jack Pratten, 2024 panel book

He kept making the case for a conversation

Pratten returned to qualitative research in December 2018. His concern was the difference between collecting measurements and understanding the reasons behind them. A business could accumulate data about activity while remaining poorly informed about why customers behaved as they did. Interviews offered a way to investigate that missing explanation.

He also connected the usefulness of qualitative work to its practical demands: careful participant selection, capable interviewers and tools that made remote sessions easier to arrange. The case was for lowering the effort required to conduct worthwhile conversations. The interview itself still needed care.

That is a less decorative proposition than celebrating a new dashboard. It asks people to reserve time, choose participants and listen to answers they cannot fully anticipate. A chart can make an argument look settled. A person may insist on adding a complication. For a product team trying to understand a decision, the complication may be the entire point.

In his later writing for B2B software companies, Pratten applied this thinking to specialised products. An employee using a tool, a manager selecting it and a buyer approving its cost bring different experiences to the purchase. A broad description of the customer can conceal the details that matter to each of them.

His 2021 essay argued for interviews as part of faster product development and better decisions. The concern was product debt: choices made today that leave problems for tomorrow’s team. Shipping more quickly increases the usefulness of learning quickly, provided the learning reaches the people making the next decision.

The guest list matters, too

Another strand of Pratten’s work concerns whose experience gets heard. In a published conversation with product leader Oji Udezue, he explored inclusive research and product-led growth. Udezue’s background included Calendly, Atlassian and Microsoft. The discussion connected customer understanding with how teams organise their work, measure behaviour and communicate ideas.

Pratten’s interest here fits the recruitment problem from another direction. A company can be attentive to every answer in a study while overlooking the people who were never invited. Access to a wider range of participants gives a team the opportunity to examine experiences outside its immediate circle. What it does with that opportunity remains a matter of judgment.

The conversation also brought writing into the picture as a way for remote colleagues to improve ideas together. Interviews and written briefs serve different purposes, but both demand the courtesy of making an idea understandable to another person. Software has made the invitation easy to send. Understanding the reply continues to take work.

Remote work, with an occasional campfire

Pratten’s accounts of Respondent’s own gatherings are unusually useful for understanding the operator behind the argument. In 2019, he described a distributed company whose colleagues worked across continental US time zones. The San Antonio retreat mixed presentations, workshops, customer-support time and ordinary work with a food tour and a boat ride to dinner.

The itinerary was ambitious. The feedback included a familiar complaint: insufficient time for regular work. Here was a research company being reminded that a gathering designed to improve work could also interrupt it. Pratten recorded the criticism alongside the team’s strong agreement that retreats were worthwhile. The operational detail makes the account more revealing than a declaration that people matter.

At the following February’s Arizona ranch gathering, the schedule changed. Work came at the beginning; the final day was reserved for activities. Building relationships and aligning around the 2020 strategy were the two stated goals. The team later regrouped for a desert cookout and a campfire with Johnny Cash songs. Corporate alignment has endured worse soundtracks.

The account describes choices rather than a compulsory single adventure: colleagues could select from outdoor activities before gathering again. It ends with feedback, too. The pattern suggests a leader willing to treat company routines as things that can be examined and adjusted. Even the pleasant parts of running a business benefit from asking whether they achieved their purpose.

Eight years behind an API

In May 2024, Pratten put numbers against the labour of building Respondent’s participant network. He said the panel had taken eight years to build, exceeded three million people and was growing by roughly 60,000 a month. He also described 2,000 hours of monthly team work devoted to growth and participant quality. These were his reported operating figures at that time, rather than a promise about every subsequent month.

The product he was discussing was a Participant Recruitment API. Other research platforms could connect their customers with Respondent’s participants inside their own tools. Great Question and Hubble were among the early partners he named. The marketplace argument had acquired another form: the network could reach researchers without requiring them to make Respondent their principal destination.

A recorded launch event with Rally makes that approach concrete. Pratten joined Rally co-founder Oren Friedman to walk through external panel recruitment, including a demonstration of recruiting without leaving Rally. For the researcher, an integration can look like one fewer interruption. Behind it sits the accumulated work of attracting, qualifying and maintaining a participant network.

That distinction gives the figures their interest. An API is a technical route into a service. The service still depends on people who agree to participate, and on the continuing effort required to make those introductions useful.

The next questions came with surfboards

By his December 2025 update, Pratten was reporting profitability, a doubling of Respondent’s team and the arrival of another million verified participants during the year. He also reported AI matching and suggestions. These were company milestones as he presented them, with payments to participants occupying a prominent place in his account of progress.

In March 2026, he described bringing colleagues from Sydney, Manila and the US to Bali. There were hackathons, exchanges about everyday AI use and surfing. The business question was how Respondent might work if it were designed around AI from the outset. He raised possibilities including automatic study applications and much shorter times to fill a study.

Those possibilities were framed as questions. They leave room for the next round of work, including the awkward answers that a recruitment company should be well equipped to seek. Pratten’s career has already supplied a reason to take such answers seriously: a useful plan can survive examination, and an unsuitable one can make way for something else.

The subway assignment remains a good companion to all this technology. Ten names on a list were never enough. Someone had to ask the next question.