Deep Research: $400 in Startup Capital, Price per Project From $700 to $6,000 — the Year It Changed What It Was Actually Selling

Dollar figures in this article come with a rough yen equivalent at 1 USD = 150 JPY.

Running customer interviews and delivering the transcript used to sell for $400-700 (about ¥60,000-100,000). The same person, with the same research capability, now charges $4,000-6,000 (about ¥600,000-900,000). This is the story of Abhishek Kumar (originally from Jamshedpur, India, now based in the UK) and the market-research agency “Deep Research” he launched in January 2023.

Startup capital: $400 (about ¥60,000), the cost of a simple website built on Umso. As he puts it, “up until I landed my first client as an agency, I hadn’t spent more than $400 total.” As of the Starter Story interview, average monthly revenue is $10,000 (about ¥1.5M), with the best month, October 2023, hitting $15,000 (about ¥2.25M). That’s roughly $120,000/year (about ¥18M), with a team of 3 including himself.

The numbers behind the ramp-up

The $150/month customer acquisition cost figure says a lot about the nature of this business. He isn’t using ads. What he’s paying for is mostly his own time, the $150 covers community dues and tool costs.

What the business actually does

He sells customer research for startups. But the content has changed across three stages over the first year.

Stage one was purely conducting customer-interview calls on someone else’s behalf. $400-700 per project. He did this as a freelancer for about 2 years.

Stage two, starting with the agency pivot in January 2023, changed the deliverable from “interview transcripts” to a comprehensive report including “buyer personas, a messaging guide, positioning, and feature and marketing recommendations.” Price: $4,000-6,000.

Stage three, roughly three months before the interview, expanded scope to UI/UX suggestions and GTM (go-to-market) strategy, becoming a package over $10,000.

Who he interviews, how he designs questions, and his interviewing skill itself didn’t dramatically change from stage one. What changed was what he was delivering the research as an answer to.

The turning point: deciding to sell “a decision,” not a deliverable

What he describes as his turning point is reframing the research not as “a service delivered” but as “a business solution.” The before/after shows up directly in price: roughly a 6-10x jump, from $400-700 to $4,000-6,000, same person, same time spent.

Another decision overlapped at the same moment. In January 2023, Kumar quit his job, with zero external clients at that point. What let him take that leap was a relationship that had continued for roughly two years since his freelance days: his first client, an agency called “Ignore No More,” and its principal, Sophia. His revenue foundation had exactly one company on it, and that one company also doubled as his first proof point and referral source.

The mechanics behind the price increase getting accepted

The reason a 6-10x price increase got accepted by clients comes down to buyer-side accounting, not willpower.

The point of comparison changed. “Conducting 10 customer interviews on your behalf” gets compared to a freelancer’s hourly rate. Nobody pays $6,000 for something a $50/hour person finishes in 10 hours. But “the conclusion on what your positioning and messaging should be” gets compared to the cost of getting that decision wrong. If it’s a decision that reshapes six months of development direction or where ad budget flows, $6,000 is small. Shift what you’re selling from labor to a conclusion, and the yardstick for price shifts from cost to avoided loss.

It’s placed right before the buyer’s decision point. Including personas, messaging, positioning, and feature/marketing recommendations in the deliverable means eliminating the step where the client has to think through the results themselves after receiving them. Research that doesn’t fill that gap needs someone internally to reinterpret it before it’s usable, and its value gets discounted accordingly.

The distribution channel was already where the buyers were. Kumar deeply engaged with paid, selective communities, Trends (Dru Riley), MicroConf, Indie Worldwide, Ramen Club, Lenny’s Newsletter, and On Deck’s fellowship. These are places where founders themselves show up, and places where mass-blast sales emails are actively disliked. In his own words: “in small communities, people talk. I regularly end up on calls with new clients referred by past clients,” and “in a good community, money comes in without you having to sell.” The 400+ individual emails/DMs and 50+ calls per year were spent building relationships in exactly these settings.

What didn’t work, and what he avoided

He deliberately didn’t adopt mass-blast cold email. It runs against the nature of what an agency is, in his view. Similarly, the website was never a primary acquisition channel. The $400 site functioned as a credibility anchor, not an acquisition engine.

His comment on media exposure is equally cool-headed. “A magazine feature has never once earned me even $1,000 more.” Newsletter ads were launched too early to measure their effect, and he dropped podcast editing work once he realized he didn’t enjoy post-production. He also temporarily moved back to India at one point to lower his burn rate.

His view on hiring runs the opposite direction. “Good people are worth paying for. It took a long time to convince myself to pay premium people, but they’re worth it.” Bringing on a researcher and a VA with 10+ years of experience led to an immediate improvement in results. In a bootstrapped business raising its price point, the ceiling on price ultimately comes down to the quality of the deliverable. That’s the order of operations here.

What’s replicable, and what isn’t

What’s replicable is the reframing of what’s being sold, and the substance behind it. Stop selling the process (interviews, tabulation, document production) and instead place the client’s actual decision at the end point of the deliverable, which means including recommendations in the deliverable itself. This transfers beyond research to design, development, accounting, or any outsourced service, and requires no capital. The $400 startup cost proves as much.

What’s harder to replicate is the foundation that was in place before he quit. At the point he went independent, he already had one client that had continued for two years. That single client simultaneously carried his revenue floor, his proof point, and his referral source. Attempt the same price increase with that foundation absent, and you create a period where revenue drops to zero the moment you raise prices. Kumar also started a podcast during the pandemic and built a base of over 130,000 followers on TikTok, laying groundwork within the community well in advance. Anyone can join a paid community, but the time it takes to become someone people actually talk about isn’t something money can buy.

And the scale ceiling should be looked at honestly too. $10,000/month across 3 people works out to roughly $3,300 per person per month. A $10,000 package consumes a large amount of the provider’s own time, leaving little room to grow without adding headcount. His stated plan to increase his acquisition budget 15x and expand into newsletter ads, podcast appearances, and public content suggests that word-of-mouth within a community, the most efficient channel he has, is likely to hit a ceiling on scale sooner rather than later.

Related reading

• MENTA’s sale to Lancers — a contrasting design that turns an individual’s skill into something tradable
• The food truck that launched in 4 months — a domestic example that also faced the relationship between price and available working hours, on a small startup budget