HeadshotPro: $100K ARR in 14 Days, Then Up to $300K MRR — Examining the AI Headshot Business

Note: Yen conversions in this article use a rough $1 = ¥150 rate.

Plenty of generative-AI stories brag about initial launch speed, but few publish hard numbers two years later, complete with a breakdown of acquisition channels. HeadshotPro, run solo by Dutch-born developer Danny Postma, hit $100K ARR just 14 days after launching in March 2023, and has since sustained $300K MRR (about ¥45M a month) through 2024–2025. He’s also disclosed efficiency numbers down to the channel level: over $50K a month from affiliates alone (15%+ of total revenue), 222,000 monthly visits, and $1.35 in revenue per visit. As material for dissecting a solo business’s revenue structure, this is top-tier.

The Published Numbers

What the Business Is, and Who Built It

HeadshotPro generates AI business headshots from a user’s selfies. Postma started in web marketing before moving into development, and his earlier product was ProfilePicture.AI, a general-purpose AI icon generator. Building the initial version of HeadshotPro took about 30 hours. That speed came not from inventing anything from scratch, but from redeploying the technical foundation he’d already built for his earlier product. The stack is Python + Stable Diffusion + DreamBooth. Generated images get automatically quality-checked with LLaVa, and Codeformer removes artifacts (generation glitches). This pipeline, filtering out “obviously AI” failed photos without human review, is why one person can handle 190,000 customers and 14 million photos.

It shares the same technical foundation and launched around the same time as PhotoAI, but by narrowing its focus to a single use case, business headshots, it captured B2B demand from companies wanting a uniform look across an entire team of employees.

One-Time Purchase, Starting at $29

Pricing is $29 (about ¥4,350) for one session, or $69 for three. It’s a one-time purchase, not a subscription. Headshots are a use case with thin repeat demand, and going monthly would mean fighting churn. He deliberately chose a pricing model that doesn’t accumulate recurring MRR, matching the nature of the use case. Flip that around, and the reality behind $300K MRR is that he keeps winning $300K worth of new purchases every single month, and what sustains that is the acquisition structure below.

A Three-Layer Acquisition Structure

What generated the initial speed was a viral launch on X (Twitter). Publishing to the audience Postma had built up over years produced the 14-day, $100K-ARR curve.

The first pillar of sustained growth is SEO. He ranks in the top 10 for “professional headshots” and runs over 200 programmatic pages targeting location-based keywords like city names, plus a blog targeting high-volume search terms.

The second pillar, affiliates, is the core of this case. Postma used to struggle with manually paying out commissions, a clunky dashboard, and referral-link management, but after switching to Rewardful (Stripe-integrated, about a 5-minute setup), he systematized the referral program and now pulls in over $50K a month (about ¥7.5M), 15%+ of total revenue from this one channel. In a solo operation, channel-management overhead itself becomes the bottleneck, so tool choice, far from a trivial matter, is organizational design in itself. Postma says (paraphrasing), “Without the affiliate program, far fewer people would have written about us,” and rather than just paying out and moving on, he keeps investing in nurturing top referrers into “superstar partners,” with the stated goal of becoming “the number-one affiliate program in the industry.” What’s more, the links affiliates place also function as backlinks, boosting SEO ranking too, a doubly-reinforcing structure where affiliates and SEO strengthen each other.

Why This Structure Worked

Even with the same technology, narrowing the use case changes both price point and customer base. Against PhotoAI’s general-purpose portraits, HeadshotPro specialized in business use, becoming a clear substitute for the existing expense of a professional photo studio session. In generative-AI applications, revenue is decided less by differences in model performance and more by the design choice of exactly whose spending, on what line item, you’re replacing.

In his own reflections, Postma emphasizes pivoting in response to market feedback, prioritizing product quality over a rushed launch, and the fact that SEO is what sustains long-term growth. There’s a consistent division of labor: viral growth ignites the fire, and SEO plus affiliates keep it burning.

Affiliates can function as a “solo sales force.” Because it’s purely performance-based, it scales with zero upfront investment, and unlike personal-brand-driven output like Shogo Log’s YouTube-and-sponsorships approach, it borrows other people’s reach instead of consuming the founder’s own time. Running $3.6M ARR alone is built on stacking up these “channels that don’t consume my own time.”

What to Discount When Reading This

Nothing about the 14-day, $100K-ARR launch speed is reproducible, since it was a one-time result of Postma’s years of built-up social-media audience colliding with AI-boom demand. An unknown developer launching the identical product wouldn’t draw the same curve. Also, while one-time purchase pricing has nothing to do with churn, it also means that if acquisition stops, revenue takes a direct hit starting the very next month. With no subscription-based inertia revenue cushioning it, dependence on sustained channels like SEO and affiliates is higher than it appears. On top of that, the technical foundation for AI headshots is shared infrastructure (the Stable Diffusion family), so there’s no real barrier to entry in the feature itself. The moat here is on the acquisition-asset side, a $50K-a-month affiliate network and search rankings, while output quality is a differentiator that erodes over time.

What Japanese Builders Can Take Away

What generalizes: (a) building systematic channels (affiliates, SEO) during a boom’s window and converting them into acquisition that outlasts the boom. (B) matching one-time purchase pricing to a use case with thin repeat demand. And (c) automating quality checks to remove the ceiling on how much one person can process. What doesn’t reproduce is the launch speed itself, $100K in 14 days starting from zero audience runs on completely different preconditions. When reading the numbers, treat the initial speed as a footnote. The real substance is the systematization that converted a boom’s tailwind into sustained channels.

Also Worth Reading

• PhotoAI — the same technology, a different focus
• TypingMind — another pattern of early AI-boom speed