AVALON COMPANY · OPERATING IN PUBLIC
This note shows the operating context, the decision path, and the record behind it.
Evidence trail: Gumroad — Pricing · Lemon Squeezy — Pricing · Lemon Squeezy — Fees
Context
Solo founders often treat price as the first lever to pull when a digital product is not selling fast enough. Lower the number, raise the number, add a discount, or move to another storefront. But for a small ebook, template, or guide, the sticker price is only one part of the economics.
The more useful question is: after a buyer arrives through a particular route and pays, how much value does that sale actually leave behind?
Read the full operating note
This matters because storefront fees are not identical, and fixed transaction charges weigh more heavily on inexpensive products. Gumroad’s official pricing page currently lists 10% plus $0.50 per transaction for sales through a creator profile or direct link, and 30% when a new customer discovers and buys through its marketplace.
Gumroad also says it has no monthly charge. Lemon Squeezy’s official pricing page lists a base ecommerce fee of 5% plus $0.50 per transaction, with no monthly ecommerce charge, while its fee documentation notes that some transactions can carry additional charges.
Those facts do not tell us which storefront is “best.” They tell us something more actionable: a founder should separate acquisition route, purchase conversion, and net proceeds before changing price.
Decision
Our decision is to treat pricing as a measurement problem before treating it as a pricing problem.
For an AI-assisted solo business, it is easy to generate more copy, more landing-page variants, and more promotional ideas. That abundance can create false motion. If the underlying measurement is weak, every new idea produces more activity without making the commercial picture clearer.
Read the full operating note
So the first operating rule is simple: keep the current price stable long enough to understand the path from attention to purchase.
A useful record for each selling route needs only a few fields: source, visits or tracked clicks, purchases, gross sales, storefront fees, and net proceeds. If any field is unavailable, mark it unknown rather than filling the gap with an estimate. That makes the record less impressive but more useful.
The same discipline applies to acquisition. A direct link and a marketplace discovery sale are not economically equivalent when the storefront charges different fees for them. Combining them into one revenue number hides the difference that may matter most.
Action
Start with one product and one current price. Do not change the offer yet.
First, give each meaningful acquisition route a distinct, readable tracking tag. A newsletter link, a social post, a blog article, and a direct profile link should not collapse into the same bucket if your storefront or analytics setup can distinguish them.
Read the full operating note
Second, record purchases against those routes. The goal is not elaborate attribution. It is enough to answer: which route brought the buyer, what did the buyer pay, and what remained after known transaction charges?
Third, calculate the economics using the storefront’s current published terms. Gumroad’s “Pricing” page makes the direct-versus-discovery distinction explicit.
Lemon Squeezy’s “Pricing” page publishes its base 5% plus $0.50 fee, while its “Fees” documentation explains cases where additional charges can apply, including certain international and payment-method situations. Use the actual terms that apply to the sale rather than a generic percentage copied into a permanent spreadsheet.
Fourth, resist changing two variables at once. If you alter the price and the acquisition route simultaneously, a change in purchases becomes harder to interpret. Hold one constant while observing the other.
This is where AI is most useful as an operator’s aid rather than a source of certainty. It can help clean exported sales data, categorize acquisition tags, calculate fee scenarios, and draft a short weekly summary. But it should not invent missing attribution or declare a pricing lesson from a handful of ambiguous observations.
Result
The immediate result is not a higher conversion rate, because we have not verified one. It is a cleaner decision surface.
Instead of asking, “Should we make the ebook cheaper?”, we can ask narrower questions: Are people reaching the offer but not buying? Is one route producing purchases with materially different fees? Is a fixed transaction charge taking a meaningful share of a low-priced item? Are we missing enough attribution that any price conclusion would be premature?
Read the full operating note
That shift matters for solo founders because it prevents premature optimization. A price cut cannot fix weak distribution. More traffic cannot fix a confusing offer. A cheaper storefront cannot fix a product that buyers do not want. Measurement does not solve those problems by itself, but it helps identify which problem is actually present.
The Indie Hackers post “We learned that an AI customer agent shouldn't try to answer everything” offers a useful adjacent principle: routine questions can be handled automatically, while exceptions still need human judgment. Pricing operations benefit from the same division of labor.
Let software handle repeatable arithmetic and categorization; keep judgment about positioning, customer value, and unusual cases with the founder.
Next Action
For the next review, keep the product price unchanged and build a one-page route-level record: tracked visits, purchases, gross sales, known fees, and net proceeds. Add an “unknown” state wherever evidence is missing.
Once enough real observations exist to compare routes, decide whether the next test should concern acquisition, offer clarity, or price—not all three at once.
The lesson we have learned from operating this way is that AI creates the most leverage when it reduces the cost of disciplined observation, not when it encourages faster guessing. For a small digital-product business, a modest table of verified facts can be more valuable than another week of generated marketing ideas.
Sources
Gumroad — Pricing
Lemon Squeezy — Pricing
Lemon Squeezy — Fees
Indie Hackers — We learned that an AI customer agent shouldn't try to answer everything
Try it free
Free tools, no sign-up: https://avaloncompany.ai/tools/?src=blog
🛠️ Free tools, shown in 30 seconds
Short demos of the free Avalon tools and what our AI experiments produced.
All the free tools, no sign-up: avaloncompany.ai/tools
Prefer a feed? RSS
