The job is small and universal: almost everyone needs one decent photograph of themselves, and almost nobody wants to book a photographer. In 2023 Danny Postma wrapped an image model around that single job and called it HeadshotPro. You upload selfies, wait a couple of hours, and get a set of studio-style headshots. The prices today start at $29 for 30 headshots and run to $59 for 70 in 4K, with team plans from about $25 a person.
The scale is stated on his own site: 196,987 customers and more than 17.9 million headshots generated. A December 2024 breakdown by Starter Story put revenue at around $300,000 a month.
The product is a wrapper, and that is the point
Nothing in HeadshotPro is technically unavailable to anyone else. The underlying image models are public. Starter Story records that the first version was built in about 30 days, and that Postma's interest started when Stable Diffusion arrived in 2022.
What was scarce was not the model. It was the packaging: a specific promise (a LinkedIn-ready headshot), a specific input (your selfies), a specific output (dozens of usable images), a fixed price and no subscription. Every one of those decisions removes a reason to hesitate.
That is the general lesson for anyone looking at AI products in 2026. The advantage is rarely the model. It is choosing one job narrow enough that the result can be judged in five seconds and priced without a conversation. Builtplain made a version of this argument about AI product images for small stores: the value shows up where a specific, repeated, expensive task gets cheaper, not where the technology is most impressive.
Affiliates do the selling
The most copyable part of this business is its affiliate programme. In a case study published by Rewardful, the tool that runs it, Postma says the programme brings in "$50,000+" a month, described as more than 15% of monthly revenue.
His explanation of why it matters is not about the commission at all: "I think without the affiliate program, fewer people would write about us as there is no financial incentive to do so for them. Having this affiliate program makes us a much stronger brand."
Every listicle of AI headshot tools, every YouTube comparison, every blog post about updating your LinkedIn photo, is written by someone deciding which products to feature. An affiliate programme is the difference between being mentioned and being recommended. And the setup cost, in his telling, was trivial: "All I have to do is pass the affiliate's ID to Stripe Checkout... This made integrating it into our business essentially a 5-minute task."
Builtplain has covered what running one of these on Stripe actually costs, including the tools and the disclosure language the FTC accepts. For a consumer product with a one-time price, it is close to the only paid channel that cannot lose money, because the commission is only ever paid out of a completed sale.
The second channel is search, built page by page
Starter Story's breakdown describes around 222,000 monthly visits and roughly $1.35 of revenue per visitor, with the traffic coming from ranking for phrases like "professional headshots" and from more than 200 programmatic pages targeting long-tail variations, including location-based ones.
Programmatic pages are simply a template filled from a list: the same page generated for each city, profession or use case. It is unglamorous, occasionally spammy, and it works when each page genuinely answers the search that produced it.
Combined with the affiliate network, this gives a business with no ad budget two compounding channels: pages that keep ranking, and writers who keep recommending. Both are versions of the same principle, that the channel is the durable asset, not the product.
The founder was not starting from zero
It matters that this was not a first attempt. His own site says his first AI company was acquired for seven figures before HeadshotPro. One published account identifies that product as Headlime, an AI copywriting tool that made $60,000 in a week on a lifetime deal and sold in March 2021 for more than $1 million, and describes an intermediate product, ProfilePicture.AI, that came before the headshot business.
So the sequence is: a copywriting tool, an exit, a consumer AI experiment, then the focused version that worked. The same account dates the HeadshotPro launch to January 2023 and claims $100,000 in the first 14 days, though other reports place the launch in March 2023. Take the exact date as unsettled and the pattern as clear: he had capital, an audience on X, and two products' worth of scar tissue before this one landed.
What the buyer is comparing it to
The pricing works because the alternative is expensive and annoying. A photographer costs several hundred dollars, requires a booking, a location and an afternoon, and produces images weeks later. HeadshotPro's own pricing page leans on that comparison for companies, citing a law firm case study that describes "Saving $15,000+ per year across 40 offices".
This is the calculation to run before building any AI product: not what the technology can do, but what the customer currently pays to get the same outcome, in money and in inconvenience. If the honest answer is that they currently pay nothing and tolerate the problem, the price ceiling will be low no matter how good the output is.
The transferable parts
- Pick a job people already pay a professional for and where the output is easy to judge. Headshots, product photos, transcripts, floor plans.
- Charge once. A $29 one-time price removes the subscription objection entirely for a purchase most people make twice in a decade.
- Add a team plan. The per-person tier is what turns a $29 consumer purchase into a company invoice, and the pricing page leads with corporate logos for exactly that reason.
- Launch the affiliate programme on day one, not once you are big. Reviewers write about products that pay them.
- Build the boring search pages for every variation of the request, and make each one actually useful.
- Ship in weeks. The first version took about a month, in a category where being early mattered more than being complete.
What to be careful about
Three honest caveats. The first is competition: this category now has dozens of near-identical products, because the wrapper is exactly as easy to build as it sounds. Being first with good distribution is doing most of the work, and that door has largely closed for headshots specifically.
The second is cost. No source opened for this article discloses HeadshotPro's compute costs or profit, and image generation is not free. Anyone modelling a similar product should run the arithmetic on serving costs before the revenue arithmetic, in the way Builtplain set out for self-hosting open models.
The third is durability. The revenue figures here are from 2024 reports and from his own posts; nothing verified this run says what the business earns in 2026. Consumer AI products face both model commoditisation and platform features that can absorb the whole use case.
What survives all three caveats is the shape: one job, one price, no subscription, and two channels where other people are paid to do your selling. That is not a technology strategy. It is an old-fashioned commercial one.
Discussion
Sign in with Google or just a name. No email link, no password to remember.