数字产品应该定价多少？我询问了 AI、研究了市场，并改变了 pricing 的思考方式

When I recently finished building a paid version of my small product, CutDone (where users create and customize printable goal lists), I ran into a question that looked simple at first: How much should I charge?

Should it be $2.99? $3.99? $4.99? $5.99?

I originally thought pricing was mostly a matter of intuition. Pick a number that feels reasonable and move on. But after discussing it with AI and looking at similar products across the web, I realized that pricing is much deeper than I had thought. And surprisingly mathematical.

CutDone Is Not a Zero-Cost Product

It would be wrong to say a digital product costs nothing. There are real fixed costs: time spent developing it, domain costs, infrastructure, maintenance, testing, and future customer support. Most importantly, there is my own time.

However, once the product is built, the economics change. The cost of serving one additional customer becomes very small. Whether 10 people use it or 10,000, the server cost doesn't scale linearly the way it would for a physical product.

So there are two different costs: - Fixed cost: building and maintaining the product. - Marginal cost: the cost of serving one more customer (close to zero for digital goods).

For digital products, pricing isn't about covering ingredient costs like a restaurant. It is about behavior.

The Formula Is Simple

The basic formula is straightforward: $$\text{Revenue} = \text{Price} \times \text{Number of Buyers}$$

The difficult part is that the number of buyers changes when the price changes. Lower the price, and more people may buy. Raise it, and fewer people will buy. The real goal is not to maximize price, nor is it to maximize the number of buyers. It is to maximize Price × Buyers.

A Simple Example

Imagine 1,000 people reach the payment screen. Depending on customer sensitivity, the optimal revenue point shifts. If a price of $4.99 yields a 3.0% conversion rate, it generates $149.70. If lowering it to $3.99 pushes conversion to 3.5%, it generates $139.65. Same product, different revenue.

One calculation became immediately useful to me: Suppose I am charging $4.99 and consider lowering it to $3.99. How many more customers do I need just to make the same revenue?

The answer is about 25%. If 100 people buy at $4.99, I need roughly 125 people to buy at $3.99 just to stay even. This changes how you think about discounts. Cheap does not automatically mean smart; sometimes you are simply giving away revenue.

Looking at the Market

I checked comparable products across Etsy and Gumroad. Simple printable trackers often sell around $2.99 to $3.99, while more polished planners sit around $4.99 to $5.99. However, many products listed higher are almost permanently discounted into the lower range due to fierce competition.

CutDone is slightly different because the user doesn't just buy a static PDF. They type their own goals, choose styles, preview the result, and see their finished design before deciding to pay. That creates a much stronger moment of purchase.

Why $4.99 Made Sense for Me

$4.99 landed in a sweet spot: 1. It stays within the normal range for digital printables. 2. It remains low enough to be an impulse purchase. 3. It is high enough that the product doesn't feel cheap. 4. It leaves room to experiment. If conversion is poor, I can always test $3.99 later. Starting low means I might never know if people would have happily paid $4.99.

Can AI Help Find the Best Price?

AI can help calculate an optimal price, but only after real users generate real data. If you test a few price points and track paywall views, checkout starts, and completed purchases, AI can model the demand curve and estimate the revenue-maximizing price. Without data, asking AI for the "perfect price" is just guesswork.

Price Competition Is a Trap

If your only advantage is being cheaper, someone else can always undercut you. $4.99 becomes $3.99, then $2.99, then $1.99. Eventually, everyone fights over nothing.

I'd rather improve the product—make the design better, the experience simpler, and the result worth keeping. If users genuinely like it, I don't need to win a race to the bottom.

This article was originally published on Ordinary Man Trying.