The price is a decision. The take-home is arithmetic.

Every fee schedule below is an illustrative snapshot of the shape of each platform's pricing, used to compare structures on identical numbers. Platforms revise fees, and regional processing rates and tax handling differ. Check the current published fee page before deciding anything, and treat tax comments here as orientation rather than tax advice.

Start with take-home, not with price

A digital product has no marginal cost worth counting — the second copy costs the same as the millionth, which is essentially nothing. That makes the fee structure the entire cost side of the business, and fee structures differ from each other far more than people expect. Two platforms charging "about ten percent" can leave you three dollars apart on the same sale.

take-home = list price − (list × percentage fee) − fixed per-transaction fee effective fee rate = (list price − take-home) ÷ list price

The second line is the one that matters and the one nobody computes. A fixed per-transaction fee is nearly invisible at $199 and brutal at $9, which is why the same platform can be the cheapest option for one product and the most expensive for another.

The same $49 product, five ways

Spread across the hosted options: $2.40 per sale between the best and worst, or about 5% of the list price. At 200 units that is $480, which is real but is not the difference between a product working and not working. Do not choose a platform on fee rate alone — choose on what it handles for you, then check that the fee is not absurd.

The same comparison at $19

Now run identical structures against a low-ticket price, where the fixed component dominates.

Two conclusions fall out. First, a flat-percentage platform is structurally better at low prices and worse at high ones, and the crossover is somewhere in the $20s for most of these structures. Second, the $0.50 or $0.30 fixed fee is an argument for higher prices and for bundles, but a much weaker one than people claim — combining a $19 and a $49 product into one transaction saves exactly one fixed fee, around $0.50. Bundle for conversion reasons, not fee reasons.

When the paid plan pays for itself

Payhip's paid tier drops the platform cut from 5% to 2%, saving 3% of revenue in exchange for a monthly fee. The break-even is a division, not a judgement call.

break-even revenue = monthly plan fee ÷ percentage saved $29 ÷ 0.03 = $967 of monthly revenue at a $49 list price, that is 20 sales a month

Below 20 sales a month the free plan wins. Above it the paid plan does. The same arithmetic applies to every "lower rate for a monthly fee" offer in this category — run the division before subscribing, and re-run it if volume drops.

VAT, sales tax, and what merchant of record actually buys you

This is the part that quietly justifies most of the fee difference in the table above, and the part most first-time sellers do not know exists until an email arrives.

Selling a digital product to a buyer in the EU or UK generally means value-added tax is due at the buyer's location, from the first sale, with no small-seller threshold for a seller established outside that market. That implies charging the right rate for each of a couple of dozen countries, collecting and retaining evidence of where the buyer was, registering under the relevant simplified scheme, filing periodically, and keeping records for years. In the US, a substantial number of states tax digital goods, each with its own economic-nexus thresholds. None of this is difficult in principle. All of it is a real, recurring, unpaid job.

A merchant of record platform becomes the legal seller. It charges the tax, remits it, files the returns, and carries the compliance obligation. That is what the extra few percent over raw Stripe buys — not payment processing, which Stripe already does, but the tax function and the liability attached to it.

For a first product, the merchant-of-record option is usually correct, and the honest way to think about the fee is as an insurance premium against a compliance obligation you do not want to discover you had. Revisit it when volume makes the arithmetic move, not before.

Tier design: why $19 and $49 can beat a single $29

The common claim is that two tiers always outsell one. That is not true, and the version that is true is more useful.

A single price puts the visitor in a yes-or-no decision, and the default answer to yes-or-no is no. Two prices change the question to which one, which quietly removes zero from the comparison set. The higher tier also anchors: $49 sitting next to $19 makes $19 read as the careful choice rather than as an amount of money. And some buyers were always going to pay $49 — a single $29 price simply hands that difference back to them.

The condition, stated honestly

Two tiers beat one price only if they raise total conversion, shift the mix toward the higher tier, or both. If neither happens, splitting a $29 product into $19 and $49 makes less money. Illustratively, on 1,000 visitors:

Row two is the one worth staring at. Adding a tier with no behavioural effect loses money, because most buyers move to the cheaper option. The gain comes from the visitors who would have bounced at a single price and the ones who happily pay more, and both effects depend on the tiers being genuinely different rather than the same thing with features removed.

Rules that hold up

1. Make the higher tier different in kind, not in quantity. "Everything plus templates, the editable source files, and a walkthrough" converts. "Everything plus 40 more pages" does not.
2. Three tiers is usually the ceiling for a solo product. Beyond that, the comparison becomes work, and work is the enemy of a cold visitor's decision.
3. Put the tier you want most people to buy in the middle, if you run three. The outer options exist largely to define it.
4. Do not create a tier you would be annoyed to deliver. A high tier that includes your time is not a digital product any more; price it as a service on a services page instead.
5. Avoid launch discounts you intend to repeat. A price that is 40% off every month is not a price, it is a $29 product with a $49 sticker, and buyers learn the pattern faster than sellers expect.

Refunds: model them, do not fear them

A clear, generous refund policy raises conversion more than it raises refunds, and refund rates on digital products are usually low single digits. Model it explicitly so it stops being a source of anxiety.

effective net per sale = net × (1 − refund rate) − (refund rate × fee retained on a refunded sale) at $49 on a 6% platform, 3% refunds: $46.05 × 0.97 − (0.03 × $2.95) ≈ $44.58

That is a 3.2% haircut on take-home. Compare it against a half-point swing in conversion, which is worth 25% of revenue at a 2% baseline, and the priority ordering becomes obvious: refund rate is a rounding error next to conversion. Note the second term, though — many processors keep some or all of the fee on a refunded sale, so a refund can cost slightly more than the revenue it returns. Check your platform's actual behaviour rather than assuming a clean reversal.

Chargebacks are a different animal and worth avoiding actively. They typically carry a fixed penalty on top of the reversed amount, and an accumulating chargeback rate can put a payment account at risk. If a buyer is unhappy and asks, refund them — it is cheaper in fees, and dramatically cheaper in risk, than making them dispute it.

The break-even unit math

Digital products have no marginal cost but they are not free to make. The cost is the hours, and pretending those hours are worth nothing is the mistake that makes a product feel successful while being a poor use of a month.

Illustratively: 40 hours to build, write, design, and set up a two-tier product, valued at $30 an hour of opportunity cost, is $1,200 of sunk cost. Blended net at a 70/30 mix across a $19 and a $49 tier on a 5% + $0.50 platform is $26.10 per sale. Apply a 3% refund rate and it is roughly $25.26.

Two thousand cold visitors is the number that reframes the project. It is entirely achievable — one guide that ranks, one post that travels, one newsletter mention — but it is a distribution problem, not a pricing problem, and no amount of tier optimisation substitutes for it. The same conclusion appears in the print-on-demand margin guide from the other direction: the unit economics are usually fine, and the traffic is the constraint.

The corollary is the useful part: because the fixed cost is hours and the marginal cost is zero, a digital product's break-even is a threshold, not a slope. Unit 49 and unit 4,900 are worth the same to you. That is a fundamentally better shape than a physical product, and it is the reason the answer to "is it worth building" is usually about whether you can reach 2,000 people, not about whether $19 or $29 is the right number.

Testing a price without an analytics stack

Most solo products never see enough traffic for a statistically clean A/B test, and pretending otherwise leads to confident decisions made on twelve conversions. The workable substitute:

• Measure revenue per visitor, not conversion rate. A higher price that converts less can still win, and conversion rate alone hides that.
• Change price for a fixed window — two weeks on, two weeks off — rather than splitting traffic, and accept the confound from seasonality and traffic mix instead of pretending it does not exist.
• Only test big changes. $19 against $49 is detectable at small samples. $27 against $29 is not, at any sample size you are likely to reach.
• Ask buyers. Ten replies to "what nearly stopped you buying this?" beat a dashboard with no conversions in it.
• Raise the price before lowering it. Underpricing is the more common error, and it is the one that is invisible because the product still sells.

How to instrument any of that without a tracking stack is its own subject, covered in the landing page guide. The products these numbers were worked against are in the store, and the free templates are on the templates page.

Tools referenced in this guide

• Store — the digital products these fee structures and tiers were worked out against.
• Templates — free templates and starting points.
• Landing page guide — how to convert the traffic a price only matters after you have.
• Services — for the work that is a service rather than a product — price it differently.