How Poket Dev made one request at a time its operating model
James Verlander is the founder of Poket Dev, a software development subscription with a flat monthly fee and one active client request at a time. Verlander explains how that constraint shapes delivery, why referrals now bring most new business, and what the company learned from letting work run in parallel.
A clear scope and one active work stream can make delivery easier to estimate and inspect. Verlander says that discipline became central to the service after parallel requests made timelines drift.
How did the business start, and what makes it different from others in your space?
Poket Dev started as a one-person answer to a pricing problem. Software agencies sell coordination: account managers, standups, layers between the client and the code. I stripped that out. Clients submit unlimited requests, one request is active at a time on a Trello board they own, and I build them myself. The flat fee is printed on the site, $5,995 a month with nothing upfront. The difference is structural, not a slogan: there is no bench, no junior handoff, and no meeting to schedule. Code lands in the client’s own private GitHub repo, and they can cancel any month and keep all of it.
How did you win your first customers, and where does most new business come from today?
The first clients came from places where founders already talk to each other: Indie Hackers, Product Hunt, Hacker News, where the subscription model could be judged in public. One case study carried most of the early weight, a 24/7 AI voice-clone host built for Mario Nawfal’s X Spaces. Clients like SoMeeAI and OneName Global followed and have spoken about the work on the record. Today most new business arrives by referral from people already running their roadmap on the board, and that is the channel I keep feeding.
What is one decision that changed the direction of the business, and one mistake that taught you the most?
The decision was enforcing one active request at a time. It caps throughput on paper and raised quality in practice: estimates stayed honest because nothing hides behind another project, and clients learned the board’s rhythm instead of fighting it. The mistake was the stretch before that rule, when parallel requests let slow work hide behind fast work and every estimate drifted. The fix became the product. One card active, revisions until approved, 72-hour turnaround target, and the drift stopped.
What has changed in your industry or in what customers expect, and how have you adapted?
Two things moved. AI tooling changed what one developer can ship in a week, and buyers stopped trusting hourly billing and long retainers. Both favored the subscription. AI and machine learning work is now core scope, from automation pipelines to voice-clone hosts, and the flat price answers the billing question before it is asked. Clients expect a partner that moves at their roadmap’s speed, so the board, the repo, and the 72-hour target are built for pace instead of meetings.
If someone were starting a business like yours today, what should they do first, and what should they ignore?
Pick one narrow promise and print the price. Say no to every request outside it, because scope discipline is the whole product in a services business. Build the operating rhythm early: one work stream at a time, decisions in writing, work the client can inspect in their own repo. Ignore growth tactics, paid funnels, cold-email blasts, vanity metrics. Do visible work for clients who will vouch in public and let referrals compound. The clients who stay are the ones who trusted the model before they paid for it.
Thank you to James Verlander and the team at Poket Dev for sharing what actually worked with Leaders Perception readers.
Want to tell the story behind your business?
Leaders Perception publishes interviews with business owners and leaders every week. Five questions, about ten minutes, no calls. It is free and every completed interview gets published.