How to sell art online: originals, prints, commissions

There is a specific kind of evening an artist loses: four hours spent fighting a website builder that treats a 40x60cm oil painting as a "product variant," ending with a shop page that makes the work look like a stock photo of a sofa.

Learning how to sell art online is really three separate skills that get bundled into one scary task. Deciding what you sell. Pricing it without resentment. And shipping it so it arrives intact and you still make money. The website is the easiest part, and it is the part everyone starts with.

So let's take them in the order that actually matters.

Decide what you are selling before you build anything

Most artists end up selling some mix of four things, and each one behaves like a different business.

The mistake is launching with all four. Originals and open prints pull your shop in opposite directions — one wants scarcity, slow consideration and a gallery feel, the other wants volume, clear sizing and fast checkout.

A workable first shape: originals as the heart of the shop, one tight range of prints in two or three sizes, and commissions behind an enquiry rather than a buy button. On mAutomate the original is set up as a single-quantity product and the prints as size and edition variants of the same artwork, so one upload produces both without you maintaining two listings.

Editions need a decision you cannot reverse: how many, and what happens when they sell out. Write the edition size on the listing, keep a log of which number went to whom, and never quietly extend an edition. Collectors who buy editions check.

Pricing: the part where artists undercharge

Pricing art online is not a formula, but there are two anchors worth holding.

The first is your own floor. Add up what the piece cost you — materials, framing, the hours, the fee your payment processor takes, the shipping you are about to absorb — and treat that total as a price you refuse to go under. Not a target. A floor.

The second is consistency. The same size and medium should cost roughly the same across your shop, whether or not you happened to love painting it. Many artists price by size per square inch or centimeter for exactly this reason: it is defensible, it scales, and it stops you from charging less for a piece you are tired of.

Prints are a different calculation. A $45 print is not a cheap original; it is a product that has to carry its own packaging, postage and breakage into a small margin. Price the print off its true landed cost, then check it against what comparable artists in your medium charge — not against your originals.

Two things not to do. Do not run frequent sales on originals; you train collectors to wait and you devalue every piece already hanging in someone's home. And do not price low "to get started." The first buyers set your market, and raising prices on an audience that found you at half the number is genuinely hard.

If a price makes you slightly uncomfortable to say out loud, it is probably about right. If it makes you resent the sale when it comes, it is too low.

Photograph the work properly, or nothing else matters

Online, your photograph is the artwork. A badly shot painting does not read as a badly shot painting — it reads as a worse painting.

What actually moves the needle:

• Flat, even, indirect daylight. Near a large window on an overcast day beats any light kit you have not learned yet. No direct sun, no overhead bulbs mixing in.
• Square to the canvas. Camera centered, lens parallel to the surface. Tilt shows up as keystoning and makes the piece look amateur.
• Color-check against the real thing. Hold the painting beside the screen and correct until they match. Buyers remember the screen color and they will tell you when it is wrong.
• One scale shot. The piece on a wall, ideally with a chair or a door frame in view. Dimension lines in text do not communicate size; a sofa does.
• Two detail crops. Edge or texture, and a signature. These sell texture-heavy work more than the full view does.
• Honest condition shots for older pieces. A visible repair or a soft corner mentioned up front prevents a return later.

For prints, add a framed mockup — but say clearly whether the frame is included. The most common print complaint is a buyer who expected a frame.

The shipping problem nobody warns you about

Here is what the tutorials skip: shipping a stretched canvas is not like shipping a mug, and the real cost is dimensional, not weight-based. A large canvas is light and enormous, so couriers bill you on volume. A 100x80cm piece can cost more to send than a box of books three times its weight.

The consequences are concrete.

Quote shipping before you publish the price. Get a real quote for your largest piece to your furthest likely buyer, both domestically and internationally. Then decide whether you ship that size internationally at all. "Originals ship within the country only; prints ship worldwide" is a perfectly respectable policy and it saves you from a $200 loss on a sale you were happy about.

Unstretched and rolled changes the math. Shipping an unstretched canvas in a tube costs a fraction of a crated stretched piece. It also means the buyer pays a framer. Offer the choice, state it plainly, and price both.

Glass is a decision, not a default. Framed work behind glass is heavy, fragile, and often refused by cheap courier insurance. Acrylic glazing or unframed-with-mount solves most of it.

Insurance and the condition record. Photograph the piece, the packing stages and the sealed box. Keep the receipt. Insurance claims on art without a documented pre-ship condition rarely go well.

Customs paperwork is on you. Cross-border art sales need a declared value and a commodity description. Under-declaring to help a buyer with duty is fraud and voids your insurance.

Build returns into the price. Even a 3% return and reprint rate on prints has to be funded by the other 97%.

Set your packaging standard once — tube for rolled, corner protectors and double-wall for stretched — buy those materials in bulk, and stop making the decision per order.

What AI can draft, and what it must not touch

This is worth being specific about, because the useful line is not where most of the arguments about AI and art put it.

It can reasonably draft: your product listings and titles, the material and dimension details, the "about the work" paragraph from your own notes, your artist statement's first pass, the collection pages grouping a series, the shipping and returns pages, FAQ answers about sizing and framing, the social posts announcing a drop, and the on-page SEO that makes a piece findable. That is the admin that eats studio time, and mAutomate's store builder for artists is pointed squarely at it: upload the image, get a draft listing with title, story and suggested print sizes, edit it, approve it.

Its support agent also handles the repeat questions — print sizes, framing, shipping timelines, commission availability — and passes you the enquiries that are actually real.

It cannot make the work, and it should not try. Nothing generated goes into your portfolio or your shop as art. That is not a technical limit, it is the whole basis of what a collector is buying.

It cannot hold your pricing judgment. A model can tell you what prints in your category tend to sell for. It does not know your reputation, your waiting list, the gallery that takes 50%, or that you are about to be in a show. Pricing is where your market knowledge lives, and handing it to a tool is how artists end up anchored low forever.

It cannot write the story of a piece. It can tidy your sentences about it. The difference shows immediately: generated artwork descriptions read like interior-design captions, and collectors who read a lot of them learn the sound.

It cannot be the relationship. Collectors buy again because of you. Let the AI answer "what size is this framed" and keep "I loved this piece, what else do you have" for yourself. That is exactly why nothing mAutomate drafts is published until you have approved it.

A sane first month

1. Shoot ten pieces properly in one session. Same light, same setup.
2. Get real shipping quotes for your largest and smallest pieces before pricing anything.
3. Publish ten originals, one print range in three sizes, and a commissions enquiry page.
4. Write your edition policy and your returns policy in plain language.
5. Tell your existing audience once, by email, with your three strongest images.

That is a shop, and it is roughly one weekend plus one evening of edits. mAutomate's trial runs 14 days with no card required, which is enough to put the ten pieces up and see how they look on a real page; pricing starts at $29 a month for one store, and a custom domain arrives on the Grow plan. If you also sell prints at volume, the margin arithmetic in print on demand store setup applies to you more than this post does. Otherwise, get started and spend the saved evenings painting.