In a sixth-grade classroom in Marysville, Ohio, Todd Kaufman met BASIC and discovered the particular thrill of making an indifferent machine obey. He later called programming “almost like a superpower,” a way to bend a computer to your will while accepting the constraints of its framework. The phrase carries the bright certainty of a child at a keyboard. Kaufman’s adult career has been an extended study of the less glamorous half of the bargain: the framework always matters.
Sometimes the framework is technical. More often, it is the set of promises, incentives and habits surrounding the code. Who decides the deadline? Who benefits when a consultant stays forever? Who absorbs the cost of a hasty feature? Who owns the value after a hundred people have helped create it? Kaufman became a developer, practice lead, agile coach, salesperson and entrepreneur, but the same concern follows the job titles. Good software cannot be separated for long from the conditions under which people make it.
A consultancy conceived in traffic
Kaufman met Justin Searls around a recruiting event in early 2009. Kaufman was then a practice lead at a Columbus technology company; Searls was a Java developer who wanted to use Rails. Twitter continued the introduction. By 2010, they were working for the same company and the same client, which required drives of more than five hours between Ohio and Michigan. The mileage gave them time to compare notes.
Their stories were about sales processes that committed delivery teams to dates and scopes before anyone close to the work had understood the problem. The founders later compared these conversations to the scar-sharing scene in Jaws: two practitioners recounting the bites. On one return journey they stopped at a roadside Chili’s. Outback Steakhouse had been rejected by Kaufman as too expensive, an excellent small detail in a story about a company that would prefer patient independence to theatrical spending.
“I felt like if we didn’t at least give it a shot, it would nag at me for the rest of my life.”Todd Kaufman on starting Test Double
The proposal formed over those drives and dinners. Diagnose root causes rather than sell patches. Finish an engagement without making the client dependent. Reward quality, not the frantic appearance of progress. Build a more equitable place for the people doing the work. In 2011, Kaufman and Searls founded Test Double, remotely, with no prior experience running a business and a name borrowed from testing vocabulary.
The company grew by word of mouth, by its standing in the Ruby and Rails community and by the public teaching of its people. Kaufman did his share of the teaching. His talks ranged from enacting change to finding joy at work and demystifying the business of consulting. He wrote about the traps in buying custom software and later made a permanent remote-work policy sound pleasingly final: do not waste time commuting to an office. Ever.
Joy, in Kaufman’s telling, was never office decoration. His 2015 talk on the subject began with an unfriendly warning: starting a business may be the most risky and stress-ridden route to joy imaginable. The useful part was the distinction underneath. A better working life could be designed through autonomy, meaningful craft and room to improve, without requiring everyone to become a founder. Entrepreneurship was his route to changing the rules, not a universal prescription. The idea helps explain why remote work at Test Double preceded the pandemic by nine years. Location flexibility was not an emergency concession. It was an operating expression of trust.
This trust came with expectations. Consultants were supposed to do more than receive a ticket and return code. They needed to understand business goals, communicate across disciplines, challenge an unhelpful request and transfer capability to the client’s own team. Kaufman has said the word consultancy captures that aspiration better than agency: delivery matters, but so does the perspective that leaves a team more capable than it was before. The distinction is demanding because it makes judgment part of the deliverable and candor part of the job.
The company changes owners
Growth produced a problem more interesting than growth itself. As Test Double became more profitable and passed 100 consultants, its success depended less on two founders and more on the expanding team. Kaufman and Searls said they felt an inequity between who generated the value and who received it. Many leaders would answer that discomfort with a bonus scheme or an eloquent memo. In April 2020, they sold Test Double to an employee stock ownership plan.
Employees receive a stake over time without buying shares. The plan is a long-term retirement benefit, but its symbolic force is immediate: “ownership” is no longer merely a request to behave responsibly on someone else’s behalf. Five years after the conversion, Kaufman wrote that he was fully confident they had made the right call and that he looked forward to seeing shareholder balances increase. Corporate culture is full of abstract nouns. Equity becomes wonderfully concrete when it appears on a statement.
The ESOP also fits Kaufman’s preference for a company that can outlast its founders. It distributes reward without requiring the familiar venture story of explosive growth and an exit. Test Double still has hierarchy, clients and commercial pressure; legal structure is not magic. But the ownership decision aligns the economic result more closely with the company’s stated belief that great teams make great software.
There is an elegant tension in a consulting firm choosing this structure. Clients pay for the judgment of particular people, yet conventional ownership often treats those people mainly as a cost against the founders’ equity. Test Double’s answer was to make the balance sheet acknowledge what the client relationship already knew. The person pairing on a difficult upgrade, coaching a team through a release or asking the awkward product question is not adjacent to the value. That person is producing it. Shared ownership turns that observation into policy, one vesting year at a time.
From seats filled to outcomes changed
The firm kept widening its field of view. In November 2023 it acquired Pathfinder Product, bringing in more than a dozen product-management consultants. The move joined product judgment to Test Double’s engineering work. It also answered a recurring failure in software: a technically competent team can deliver exactly what was requested and still leave the business with the wrong thing.
By 2026, Kaufman was applying that distinction to artificial intelligence. His position was neither denial nor costume change. AI can accelerate unfamiliar codebase work and compress the time needed to produce a prototype. It cannot guarantee that the prototype deserves to exist, that a pattern fits the context or that a quick success will survive real scale. His neatest formulation is economic: outputs have become cheap; being wrong has not.
That view pushed Test Double to refine its mission. “Improving the way the world builds software” had served for nearly a decade, but Kaufman had come to dislike one word: builds. A line of code that does not advance an organization’s goals is waste, however elegant it looks in review. The revised mission focuses on improving software’s impact, and the vision supplies a countable horizon: by 2030, achieve 1,000 client outcomes through people, process and technology consulting.
“We’re not going to completely change who we are to chase a dollar.”Todd Kaufman on the AI consulting rush
A measurable outcome forces harder questions at the beginning. What problem is the client actually facing? What should change in the business? How will anyone know? These are plain questions, which is why organizations are talented at avoiding them. A request for more engineering capacity is easier to process than a conversation about whether another feature deserves to be built.
The framework around the work
Kaufman’s public manner mixes candor with a taste for the gently ridiculous. His official Test Double dossier calls him Agent 001. It lists understanding the Enterprise JavaBeans 2.0 specification among his noteworthy achievements and a failed attempt to guide more than 60 consultants through a 2008 downsizing among the others. In a RailsConf introduction, Searls told the audience that Kaufman likes beer, running and pugs. This is not a founder persona burnished to a chrome finish. It is closer to a capable colleague’s record, mistakes left in.
The failure belongs there because Kaufman’s subject is not personal infallibility. It is the design of conditions in which people can exercise judgment. Remote work offers autonomy. Cash restraint protects choices. Consultants who can speak plainly about business goals reduce the distance between promise and delivery. An ESOP shares the result. A measurable mission helps prevent motion from impersonating progress.
The theft-worthy part of his playbook is therefore not a single ritual. It is a habit of looking one layer above the visible problem. A troubled codebase may expose a testing problem; the testing problem may expose a deadline incentive; the deadline incentive may expose a sales promise; the sales promise may expose a business that profits from beginning work before it understands the work. Repairing the nearest symptom is quicker. Repairing the framework is how the symptom stops returning in a nicer font.
The child in Marysville wanted to bend the computer to his will. The founder in Westerville has spent 15 years asking whose will software serves, who pays when it goes wrong and who should benefit when it goes right. The superpower survived. It simply acquired governance.