How to price a digital product (with the calculator math shown)

A digital product's price comes from three dials rather than from your costs: the band the market already pays for that product type, a multiplier for who buys it, and a multiplier for how complete it is. Multiply, snap the result to a 9-ending ladder, then subtract the platform fee before you believe any revenue number.

Most pricing advice for digital products is a vibe dressed up as a method. Charge more. Know your worth. Helpful in the way a motivational poster is helpful. When I had to price nineteen products for this catalog, I wanted arithmetic I could defend, so I built a model. It runs the free digital product pricing calculator, and this post opens it up.

One scoping note. This is the model-based approach: defensible defaults from market data, usable before you have any buyers. The survey-based alternative, asking real buyers four questions and finding where the curves cross, is in pricing with Van Westendorp. The model gets you a launch price in five minutes. The survey validates it once you have an audience to ask.

Dial one: the market band for your product type

Every product type has a range the market already accepts, visible in what actually lists across the major storefronts. Trimmed of outliers, the 2026 bands the calculator uses are printables $3-$12, prompt packs $9-$39, e-books and guides $9-$49, templates $19-$79, software and toolkits $19-$149, courses $29-$299.

These are norms, not laws. A strong brand can leave the band, and a few famous outliers price ten times above it. Leaving the band should be a decision you make on purpose with evidence, not a default you back into because a round number felt right. The band is where the burden of proof sits.

Dial two: who buys it

The audience multiplier moves the final price more than the product itself does. Hobbyists buy from an entertainment budget, so multiply the band by 0.7. Professionals buy from a tools budget, so multiply by 1.0. Businesses buy from an ROI calculation, so multiply by 1.6.

The business premium is the one people resist and the one the evidence supports most strongly. A team lead expensing your $79 template is not spending their own money, and they are comparing the price against an employee-hour. Pricing a product that is useful to a team at hobbyist levels does not win volume. It costs margin, and it quietly costs credibility, because a price that low reads as a signal about the product rather than a favour to the buyer.

Here is what the two dials look like when you run every combination through the engine. This is the whole grid, at standard depth:

Read across any row and the same artifact doubles in price without changing at all. That is the part worth sitting with. Most people agonising over their price are actually undecided about their buyer, and no amount of thinking about the product will resolve it.

Dial three: depth, honestly assessed

A starter product that solves one narrow job multiplies by 0.8. A standard, complete-for-its-category product is the 1.0 baseline. A comprehensive flagship multiplies by 1.4.

The discipline here is honesty, because every maker believes their product is comprehensive. The test I use: would a buyer who got it free still need to buy something else to finish the job? If yes, it is not the flagship multiplier.

The depth dial widens the band rather than just shifting it, which is easiest to see on one row. Software toolkit, sold to professionals:

The anchor is the geometric mean of the multiplied band rather than the arithmetic one, because price bands are ratio-scaled. Averaging $19 and $149 arithmetically gives $84, which is far too close to the top. Geometrically it gives $53, which snaps to $49 and is where comparable products actually cluster.

Then snap to the ladder

Multiply the three dials and you get an ugly number. Snap it to 9, 12, 15, 19, 24, 29, 39, 49 and upward, with gaps that widen as prices rise.

Two things are going on. The 9-ending itself has held up in field experiments for two decades: Anderson and Simester's catalog tests found 9-endings raised demand, and in one case an item sold more units at $39 than at $34. Selling more at a higher price is an unusual result, which is why the finding stuck.

The widening gaps are the less famous half and they matter more for a catalog. Buyers perceive price differences as ratios, so $19 to $24 reads as a real step while $119 to $124 reads as noise. A ladder with even dollar gaps gives you tiers your buyer cannot distinguish at the top and cannot afford to skip at the bottom.

Then subtract the fees, before you celebrate

Whatever price the model gives you, the platform takes its cut before you see revenue, and the cut varies more than most people expect. On a $29 sale that is roughly 3.4% through Stripe's published rate of 2.9% + 30¢, about 15.7% on a direct Gumroad sale once processing is added to the 10% + $0.50, and a flat 30% on a marketplace-attributed one. Revenue projections that skip this step overstate income by a sale in every seven on some platforms.

The platform fee comparison shows the net on your exact price across all of them, and the full itemization for the platform I use is in Gumroad fees in 2026.

Then sanity-check the volume. IRP Commerce's market data put average ecommerce conversion at 2.23% in August 2026, and that average is lifted by returning buyers on established stores. A plan built on 5% cold conversion is a plan built on a typo. Run the revenue table at 0.5%, 1% and 2%, and check that the pessimistic row still clears your effort bar.

The dial nobody turns

Here is what I think the pricing genre gets backwards.

Almost every guide treats price as the output of a calculation about the product. Time invested, features included, competitors surveyed. The calculator above does that too, because it has to start somewhere. But look again at the matrix: moving one row down changes the anchor by about 60%, and moving one column right changes it by about 60% as well. The columns are audience. The rows are product type.

Which means the highest-leverage pricing decision is not what to charge. It is who to build the next version for, and that decision is usually made unconsciously, in the language of the landing page, months before anyone thinks about a number. A toolkit written in hobbyist language cannot be rescued by a professional price. The price just becomes the thing buyers argue with.

So the practical move is to run the model twice. Once for the product you have, once for the same product described for the audience one column to the right. If the gap is large and the work to close it is small, the price was never the problem you were solving.

If you are modeling that with real numbers rather than by feel, the AI CFO Toolkit has the unit-economics and scenario sheets I use for it.

The whole method in one line

Price band, times audience, times depth, snapped to the ladder, minus fees, sanity-checked at 0.5% conversion. Five minutes of arithmetic that most sellers skip, and the cheapest competitive advantage in the business. Run it in the pricing calculator, then go and validate it against real buyers with the survey method.