They Killed the Billable Hour. Then They Built the Rest of the Agency Around It.
BaD Mktg's Ian Buck and Yotam Dor on ditching the billable hour, building their own AI tools, and running an agency with no junior bench.
BaD Mktg Co-Founders, Yotam Dor and Ian Buck
Anyone who's worked with an agency has felt the billable hour at some point: the slow creep of hours added to an invoice, the sense that time on the clock matters more than the work itself. It's a structure built around getting paid for effort rather than results. Ian Buck and Yotam Dor spent years inside that system before they walked away from it entirely.
BaD Mktg, the indie shop they founded two years ago, charges project fees instead of hours. That single decision, they say, is the one everything else traces back to: the two-person team, the AI tools built in-house, the freelancer network standing in for a junior bench. Get the incentive right, and the rest of the agency follows from it.
Two years in, BaD Mktg has grown from $1.5 million in revenue to roughly $2 million, still with only two full-time employees. We asked Buck and Dor to walk us through what running an agency this lean actually looks like, from the tools they built themselves to the assumption about indie shops that turned out to be wrong.
A Different Kind of Lean
The billable hour wasn't an isolated fix. Buck and Dor say the shift to project fees was the decision that shaped everything else: the size of the team, the tools they built, the way they staff a project. “No padding, no clock-watching, no meeting that runs long turning into a line item," Buck says. “Everything else about how we built BaD Mktg follows from that one decision."
That includes staying at two full-time employees, with no junior bench underneath them. It's a structure that sounds thin on paper, but Buck and Dor say the trade-off wasn't the one they expected. “The assumption going in was that being lean meant constantly underwater, no room to think," Buck says. Instead, offloading the admin and first-draft work to their AI tools freed up time for the part of the job they'd actually gone independent to do: the strategy, the creative, the thinking.
It also changes what a client relationship looks like day to day. At a larger agency, a small issue can sit with a junior contact before it works its way up to someone who can act on it. At BaD Mktg, there's no queue to sit in. “It's myself or Yot, every time, so we catch the thing while it's still a five-minute fix instead of a fire drill," Dor says. Fewer layers, in their telling, means fewer surprises: nothing gets the chance to get worse before someone senior is looking at it.
The Tools They Built Themselves
Most agencies talk about using AI to move faster. BaD Mktg built its own tools to do it, and drew a hard line around what those tools are allowed to touch.
Goodie, an AI chief of staff running on Claude, came out of a simpler problem: two people running a full agency's worth of client relationships, buried in the connective tissue of the job. “We were drowning in the stuff that has nothing to do with the actual work," Buck says, “pulling together form data, calls, texts, emails, status reports." That's the layer Goodie absorbed almost immediately: anything administrative or aggregative, anything that amounts to gathering information and putting it in one place.
What Goodie doesn't touch is anything with a client's reputation or money attached to it. Whether a client is actually satisfied or just being polite, when to push back on a brief, whether a creative idea is actually right: Buck and Dor say those calls stay with them. “That needs to come from us," Buck says.
Researcher, their second in-house tool, works earlier in the process, turning a brief into a shaped, opinionated starting point in hours instead of days. But Dor is clear about where its usefulness ends. “It doesn't have taste," he says. Researcher can generate ideas faster than two people working alone ever could, but deciding which idea is actually right for a given client, in a given moment, and having the judgment to kill eight good ones for the one great one, is still entirely on them.
A Bench Without a Roster
Without a junior layer, the natural question is who catches the mistakes. Buck and Dor's answer is that the production process is the actual safeguard, not a layer of oversight sitting between the work and the client.
Their freelancer network isn't a cold list of names assembled to fill a gap. It's built on relationships going back to the agency Buck and Dor ran together starting in 2012, more than a decade of watching the same creators and crews deliver, project after project. “The junior-bench model assumes you need layers to catch mistakes," Buck says. “We'd argue you need the opposite: a bench of proven talent, already vetted, so most mistakes never get the chance to happen in the first place."
Clients know exactly what they're getting, according to Dor: Buck and Dor themselves as the constant, senior presence on every project, and then whichever creators or crew fit that specific brief, not whoever happened to be available that month. “The project decides who's right for it, full stop," he says.
What That Model Actually Produces
The clearest example of how BaD Mktg's process plays out is a recent project for Kaseya, the global IT management and cybersecurity company. The agency was brought in to open Day 2 of Kaseya Connect, the company's flagship customer conference, and the brief could easily have landed as a standard tech sizzle reel: product shots, UI walkthroughs, the usual innovation language.
Instead, BaD Mktg built a cinematic film around a simpler idea: that the invisible work IT teams and MSPs do deserves to be seen. The people on screen are Kaseya's own engineers and leaders, not actors, and the film's closing line, “That's why we build," became the thing that outlasted the keynote.
What started as a single conference opener has since become part of Kaseya's brand platform. That closing line now anchors the company's latest release campaign, and the original film has been recut into a suite of ads and follow-up pieces. You can watch it here or here.
What the Margin Actually Proves
BaD Mktg hit $1.5 million in revenue by its second year, with a 38 percent profit margin, The Globe and Mail reported in April. Buck and Dor say revenue has since grown to roughly $2 million, with margin settling at 36 percent. On paper, that dip might read as a warning sign. They don't see it that way.
“This is a volatile industry: client budgets shift, projects move quarters, priorities change overnight," Buck says. Riding some fluctuation year to year isn't a red flag in his telling, it's just what running an agency looks like.
What strikes them more is where that margin sits relative to the industry, not the small shift within it. Most agencies run single-digit margins, and even at 36 percent, BaD Mktg isn't close to that range. “That gap is the real story," Buck says. “It's the unlock we didn't fully appreciate until we saw where we sat against industry norms." Growing revenue by roughly $500,000 while staying at two full-time employees is, to them, the actual proof point: better tools can scale an agency further than adding headcount ever could.
What Two Years Actually Gets You
Ask Buck and Dor what assumption about running an indie shop turned out to be wrong, and they push back on the premise. “It's less 'wrong' and more ‘we had no idea how right,'" Buck says. Figuring things out as they went was always part of the plan, given how fast the underlying technology was moving.
What they didn't see coming was building their own tools. Researcher and Goodie weren't bought or licensed: they were built in-house, running on models that didn't exist when BaD Mktg launched two years ago. “If you'd told us at launch that we'd have our own, working AI chief of staff within two years, we'd have said that's a five-year roadmap for a company with an engineering team," Buck says, “not two people."
It leaves them with a question they can't fully answer yet. If two years gets an agency this far, run by two people who left the industry they knew to build something smaller, what does the next two look like?
