Most academic market-entry projects fail because they confuse research for thinking. The team reads twelve reports, builds a PESTLE, runs a Porter’s Five Forces, color-codes a SWOT, and arrives at a recommendation that any competent operator could have produced from intuition in a single afternoon. The artifact looks rigorous. It is not.
The project I led for Workana at MBA was an attempt to do the opposite — to spend less time on frameworks and more time on real customer conversations, and then write a recommendation sharp enough that the company’s growth team could pick it up and execute without translation. The brief was deepening Workana’s position in Buenos Aires, an existing market where the company had real but not dominant share. The output was, by my count, three opinionated moves. The company actually used two of them. That’s the receipt.
§ 01 — The setup
Workana is the largest freelance work marketplace in Latin America — millions of registered freelancers, hundreds of thousands of active clients, a meaningful chunk of the region’s white-collar gig economy moving through the platform every quarter. The pitch is straightforward: connect Latin American freelance talent to clients in LATAM, the United States, and Europe, in Spanish-and-Portuguese-native workflows, at price points that beat the US-defaults of Upwork and Fiverr.
Buenos Aires is one of Workana’s earliest and densest supply markets. Argentina’s freelancer-per-capita numbers are extraordinary — a combination of macroeconomic instability that pushes people to dollar-denominated remote work, a strong cultural tradition of independent creative and technical work, and a labor market where formal employment has been getting more expensive for decades. ILO and World Bank data place independent and informal work at well over a third of total Argentine employment in recent years, and the white-collar slice of that — designers, developers, marketers, translators, accountants — has been growing faster than the formal economy.
That should be the easiest market in Latin America to win. So why was it merely doing well, rather than dominantly winning?
§ 02 — The framework trap
The way most teams would have approached this brief — and the way I watched several peer teams in the same cohort actually approach it — is the framework trap. PESTLE the country. Porter the industry. SWOT the company. End up with a recommendation that says some version of “double down on what’s working, mitigate the FX risk, invest in brand.”
None of those sentences are wrong. None of them are useful, either. A framework can rule out bad answers, but it almost never produces the specific answer that wins. The specific answer comes from somewhere else: from talking to enough customers that you can hear the constraint they don’t know how to articulate.
Steve Blank has been writing this for twenty years. The customer development literature has been writing it for forty. And yet: every cohort, every project, every team retreats to the framework when the framework feels safer than the conversation. The framework is a way to look like you’ve done work. A conversation is actually doing it.
§ 03 — What we did instead
We allocated a week of the semester — out of fourteen — to frameworks. We allocated four weeks to interviews. Supply-side and demand-side, both. Forty-plus structured conversations, conducted in Spanish where appropriate, transcribed and tagged.
The supply-side interviews — freelancers — were the easier half. People with strong opinions about the platforms they earn money on are not a scarce resource. The themes that emerged:
- Currency and payment friction was the single most-cited pain point. Not commission rates. Not algorithm fairness. The specific mechanics of moving USD from a foreign client to a usable Argentine peso (or, more often, into a stable dollar-denominated savings vehicle).
- Brand fatigue. Workana’s brand was strong but felt mature. Younger freelancers — the next generation of supply — were drifting toward newer competitors with sharper positioning, even when the underlying product was inferior.
- Discovery in oversaturated categories. Argentine designers and developers competing against thousands of peers had real, specific complaints about how surfacing worked, and how a new freelancer establishes credibility.
The demand-side interviews — clients — were the harder half, because clients are less reachable and less talkative. But they were where the real insight was hiding:
- Local Argentine businesses hiring on Workana had a fundamentally different use case than the foreign clients who’d built Workana’s brand. They were not arbitraging currency. They were not trying to find cheap labor. They were managing a small business that needed a designer for a specific weekend project and didn’t want to hire a full-time employee for it.
- The platform was, in their telling, over-engineered for the long-engagement, foreign-client use case and under-engineered for the short, local-client use case that was growing faster.
- Trust signals were misaligned. The signals Workana surfaced — ratings, completion percentage, response time — were calibrated for the foreign-client trust model. Local clients had different trust priors and were looking for different cues, none of which the platform surfaced.
None of this was on a PESTLE. None of it was in a published report. It was in forty conversations, and you could not have written the recommendation without them.
§ 04 — The recommendation
The recommendation was three opinionated moves. Not twelve. Not “doubling down on what works.” Three.
Move one: build a fast-lane product for local short-engagement work. Strip the onboarding friction. Surface trust signals calibrated to local-client priors (verified phone, neighborhood, prior Argentine clients) rather than the international defaults. Lean into a category of work — small-business design and copywriting — where local supply and local demand were both growing and the existing UX was a poor fit.
Move two: a payments-first brand campaign aimed at the supply side. Not “we are the biggest marketplace.” Not “we have the most jobs.” Specifically: “we make moving USD into your life easier than any of our competitors do.” This is the actual constraint freelancers face in Argentina. It is what they would tell their friends about, if the platform actually solved it. The marketing dollars belong here, not on undifferentiated brand spend.
Move three: reposition the next generation of supply. Workana’s brand felt mature to a 22-year-old designer entering the market. A targeted creator-program — partnerships with university programs, sponsored creative competitions, an actual content presence in the channels where young Argentine creatives spend their time — would have done more for ten-year supply pipeline than another six months of generic acquisition spend.
The deck was twenty-something slides. The interview appendix was bigger than the deck.
§ 05 — What got built
I’m not the company’s growth team and I will not claim credit for product decisions I don’t own. But two of the three moves — the local-fast-lane direction and the payments-as-brand framing — showed up in their roadmap and communications in the months that followed. Whether our project moved that needle or whether their team was already there, I can’t be sure. What I can say is that the recommendations were specific enough to be usable, which is the bar an academic project rarely clears.
The third move — the next-generation supply repositioning — is harder, slower, and a different kind of investment. I still think it’s the highest-leverage of the three. The companies that win twenty-year platform games are not the ones with the best current supply. They’re the ones whose current supply recommends them to the next generation.
§ 06 — How we ran the interviews
If the punchline of this project is “do more interviews and fewer frameworks,” the natural follow-up is “how do you run forty interviews without producing forty pages of useless transcript?” Here is what worked.
We split the team by language and side of the marketplace. Native Spanish speakers ran the Argentine supply-side and local demand-side calls. Non-native speakers ran the international-client demand-side calls. This sounds obvious; teams violate it constantly. A freelancer interviewed in their third language gives you a third-language version of their opinions. The signal density is thinner than the transcript length suggests.
We wrote two interview guides, not one. Supply-side and demand-side have different priors. A single guide forces you to ask both populations the same questions, which produces parallel-but-shallow comparisons. Two guides let each conversation go where the participant’s expertise actually lived. We re-cut and synthesized later, in analysis, not during the conversation.
We banned hypothetical questions. “Would you use a feature that…” is the worst question in product research. People are terrible at predicting their own future behavior. We replaced every hypothetical with a behavior question: “Tell me about the last time you…” The transcripts got dramatically more useful overnight. Erika Hall has written this advice for years. It is still the single highest-leverage interview rule I know.
We asked about the last bad experience, not the average experience. Average experiences are forgettable and produce neutral commentary. The last time a freelancer didn’t get paid on time, the last time a client got a deliverable they couldn’t use — those memories are crisp, specific, and full of the constraints we needed to surface. We led with the bad memory and let the average emerge from the surrounding context.
We synthesized weekly, not at the end. Forty interviews analyzed at the end is forty interviews you have largely forgotten. Five interviews per week, synthesized into a one-page memo every Friday, kept the pattern visible the whole way through. By week four, we knew which themes were holding up and which were noise. By week six, the recommendation drafted itself.
§ 07 — Why the third move matters more than the other two
I noted earlier that the third recommendation — repositioning for the next generation of supply — was the slowest, hardest, and in my view the highest-leverage of the three. It deserves a paragraph more than it got.
Two-sided marketplaces fail in one specific way: the supply ages out faster than the demand renews. The freelancers who built Workana’s brand in 2012-2018 are now in their thirties and forties, often graduating off the platform into senior in-house roles or higher-end agency work. The freelancers who are becoming the supply now — 22- to 28-year-olds entering the market — have a different reference set. They didn’t grow up with Workana. They grew up watching designers on Instagram, developers on YouTube, and copywriters on Substack. The platforms they associate with “where serious work lives” are different.
A marketplace that doesn’t actively recruit the next generation of supply is not stable; it is decaying with a delay. The decay shows up in five years as thinning supply in the categories you care about. By then, repositioning is much more expensive than it would have been to invest in early. The companies that get this — Etsy in the 2010s, Substack today, even LinkedIn’s slow shift to creator content — make the investment before they need to. The companies that don’t, get out-positioned by a newer competitor whose only real advantage was that they spoke to a 22-year-old like a 22-year-old.
This is the kind of recommendation that does not show up in a framework. It shows up in an interview where a 23-year-old illustrator tells you, completely unprompted, that Workana feels like “where my older sister gets her work.” That is a sentence I would not have heard if I had spent the semester on a PESTLE.
§ 08 — The take
Every market-entry project — academic, consulting, in-house — is one good interview away from being a real recommendation instead of a generic one. Most of them never have that interview, because the conversation feels less productive than the framework, and the framework is reassuring to the team that hasn’t done the conversation yet.
If you’re running a market-entry or growth project right now, do the following experiment:
- Take your current recommendation.
- Imagine handing it to an operator at the target company.
- Ask yourself: could they have written this themselves, in an afternoon, from intuition?
If the answer is yes, your recommendation is not adding value. The work you did to produce it was useful for you — frameworks teach the analyst — but the artifact you produced is decorative.
The way to fix that is not more frameworks. It is more conversations, with the people whose constraints you are claiming to understand. Forty interviews is not a lot. It is a semester. It is the difference between a deck the operator throws away and a deck the operator uses.
Real client. Real recommendations. Mostly real-world useful. That’s the bar.
Receipts, not vibes.