Custom Puzzles: A Print on Demand Product Guide

TL;DR: A custom puzzle lives or dies on image resolution at piece scale, because an adult puzzle piece is only about three quarters of an inch across and a phone photo runs out of detail long before it fills a 20 inch board. Printful, Printify and Gooten print puzzles on chipboard with a glossy face, boxed in tins or card, while Gelato's published catalogue does not list them and Amazon Merch does not make them. Piece count is a product decision, not a spec line: 30 pieces is a toy and 1,000 is somebody's quiet week. Margins are unforgiving, because a heavy boxed item that turns up crushed or blurry is a refund and a reship.

Puzzles get added to print-on-demand shops for a reason that has nothing to do with puzzles. They are a gift. Somebody wants to hand a person a photograph they cannot hang on a wall, and a box of 500 pieces turns that photograph into two hours with it. Real purchase motive, peaks at Christmas, not especially price sensitive.

Then there is the other half of the traffic for this phrase. If you are about to buy one and want to know whether these things are any good, the answer is yes, with a condition. Print-on-demand puzzles are cut from laminated chipboard with a glossy printed face, the construction the mass-market brands use, and the pieces interlock properly. What varies is not the board. It is the picture somebody put on it, and a puzzle made from a low-resolution photo is a bad product however good the factory is. Nobody in the supply chain will stop it being printed.

Resolution at piece scale is the whole game

Here is the part that catches people, and it is worth being concrete about rather than repeating a DPI rule.

Take the dimensions suppliers publish. Printify's personalised puzzle runs from 10 by 7.99 inches at 110 pieces up to 28.50 by 19.02 inches at 1,014. Gooten's are 10 by 14 inches at 252 pieces and 16 by 20 at 500. Divide board area by piece count and something useful falls out: across almost the entire adult range, one piece stays around three quarters of an inch square. More pieces does not mean smaller pieces. It mostly means a bigger board.

That matters because of how a puzzle gets looked at. A poster is viewed whole, from a few feet away, and your eye forgives a lot. A puzzle is held one piece at a time, at reading distance, for hours, and the assembler is hunting for detail inside that three quarter inch square. Soft edges that vanish on a wall are the only thing they can see.

Now do the arithmetic on a phone photo. A 12 megapixel image is 4032 pixels along its long side, and Printful asks for at least 300 DPI for the best result. Spread 4032 pixels over the 14 inch edge of a 252-piece board and you get about 288 pixels per inch, basically on spec. Spread the same pixels over the 20 inch edge of a 500-piece board and you are down to roughly 200. Same photo, same phone. One puzzle is sharp and one is soft, and all that changed was the piece count the customer picked.

So if you sell your own artwork, work at final board size from the start instead of scaling up later, and the same relationship between physical dimensions and pixel counts that governs standard art print sizes applies here unchanged. If you accept customer photos, the resolution problem becomes somebody else's camera roll, which gets its own section below.

Printify's guidance adds the rest: 300 DPI at final print size, sRGB colour, PNG or JPEG. Treat it as a floor, not a target.

Piece count is a product decision, not a detail

Sellers treat piece count as a dropdown. It is closer to deciding which product you sell.

A 30 piece board is a preschool toy. The buyer is a parent or grandparent, the image wants one bold subject, and the price expectation sits near a picture book. A 1,014 piece board is an evening, or several, for an adult who already owns puzzles and has opinions about them. They notice piece fit, board thickness and whether the image has enough variation to be assemblable. Not the same customer, and a listing written for one reads badly to the other.

The middle is where most sales happen. Somewhere around 252 to 500 pieces is the safe gift default, because it finishes in a sitting and intimidates nobody.

If you carry only two options, 252 and 500 cover the gift market. Add 1,000 when your artwork rewards the time, which most photographs do not.

Who actually prints puzzles right now

I checked the live supplier pages while writing this. Catalogues move, so confirm before building a listing around any of it, and read the current price sheet in your own account.

A few notes on what I could and could not confirm.

Board thickness. None of the three publishes a thickness figure on the pages I could read. If that matters to your positioning, and on a 1,000 piece listing aimed at hobbyists it does, order a sample and measure it rather than quoting a forum number.

Edge to edge printing. No supplier page states this in those words. Gooten publishes templates defining the printing area and the bleed, which is the practical answer: design into the template, run your artwork past the trim, let the die cut fall where it falls.

What comes in the box. Printify and Gooten both include a reference image, a lid print in Printify's case and a printed guide in Gooten's. That sounds trivial and it is not: a hobbyist assembling a 1,000 piece board with nothing to look at writes a one star review.

Gelato. Its catalogue help page lists wall art, apparel, phone cases, mugs, tote bags, calendars, cards and photo books. No puzzle category appears. The main catalogue page blocked my request, so check your dashboard if you sell through Gelato, but do not plan a puzzle line around it.

Connecting any of these to a shop works like any other product, and that side is covered in print on demand companies for Etsy.

The photo upload question changes everything

Search interest in "custom photo puzzle" is a large slice of the demand for this product, and that phrase describes a different business from selling your own designs.

Sell your own artwork and the transaction is clean. You uploaded a file you control, you know it is sharp, and a complaint is about the product. When the buyer supplies the image, you have taken on a job, and three new problems arrive with it.

The listing has to collect a file. Etsy gives you a personalisation field, but it takes text, not a photograph, so most sellers ask for the image by message or an upload link once the order lands. That is a manual step inside what was supposed to be an automated product. For one-off commissions, a private listing on Etsy keeps the price and the brief attached to the right buyer.

Customer messages become part of the product. You will spend real time asking for bigger files, explaining crops, and telling people a screenshot of an Instagram post will not survive 16 by 20 inches. Write that conversation once and reuse it. State the minimum pixel dimensions in plain language before the customer picks a piece count, and your Etsy message to buyers is the obvious place for it.

Refund exposure moves. A personalised item is not resellable, so a dispute is a total loss rather than a return to stock. Worse, the usual cause of dissatisfaction is the thing you warned them about, a soft image. Either you reject weak files before printing, which costs sales and goodwill, or you print them and absorb some unhappy buyers. Pick a policy, publish it, price it in.

My read: photo puzzles are the more lucrative half of this market and the more operationally annoying half. Want puzzles to behave like passive print on demand? Sell your own art. Want the higher ticket? A small service business comes attached.

Niches that sell, and what drives them

Puzzle demand splits fairly cleanly into two motives, and confusing them is how listings end up mispriced.

The pattern underneath is simple. Gift niches are bought by someone other than the person who opens the box, so packaging and listing photos matter more than difficulty. Hobby niches are bought by the assembler, which makes the artwork itself the product and the reviews brutally specific.

Where to sell, and what I could actually verify

Etsy. The natural fit. Gift intent, personalisation expected, a buyer already in the mood to pay for something specific. Every supplier above connects to it, and this is where the photo puzzle business lives.

Your own store. Worth it once you have traffic, and the only route where you control the upload flow.

Redbubble. It does carry jigsaw puzzles, and its shop pages are indexed for several piece counts including 1,000. The limits of that claim: Redbubble returned a 403 to my request, so I am going on indexed listings rather than a page I read, and I could not confirm current sizes, board specs or artist margin. Check your own dashboard first. Redbubble is a pure artwork channel, so customer photo uploads are not part of it.

Amazon Merch on Demand. Not an option. The public product type guides I could read cover apparel plus phone cases, PopSockets, totes and throw pillows. No puzzle appears in any of them. The Merch portal sits behind a login, so I could not pull an official catalogue page, but nothing public suggests puzzles exist there. Amazon still matters for reading demand, since third party sellers list plenty of custom puzzles, but you will not be printing through Merch.

Pricing and margin, in plain terms

There are no base costs or margin figures here, because supplier pricing changes and a stale number is worse than none. Open the live price sheet in your own account before setting a retail price. The shape of the economics is what travels, and it does not move.

A puzzle is a heavy, boxed, rigid item. Three things follow.

Shipping is a real line in your maths rather than a rounding error, and it gets worse on international orders and larger boards. Free shipping offers that work fine on a t-shirt can quietly eat a puzzle's entire margin.

Damage in transit is a genuine cost centre. A crushed corner draws a complaint even when every piece inside is perfect, and you will usually reship instead of arguing.

And a print fault is total. A t-shirt with dull colour gets worn anyway. A puzzle with a blurry face gets photographed and refunded. Your quality floor sits higher here than on apparel, which is the argument for sampling each board size before you list it.

The upside is a friendly retail ceiling. Gift buyers compare a personalised puzzle against other gifts, not against a supermarket puzzle, and that comparison works in your favour. Hobbyists are the opposite: they know what a 1,000 piece board from a known brand costs and they will measure you against it. Price the gift side on occasion and the hobby side on quality. The framework in how to price print on demand products holds, with shipping weighted heavier than usual.

The downsides nobody puts in the product description

Missing pieces. The defining failure mode here. A puzzle with 499 of 500 pieces is not 99.8 percent of a product, it is zero percent of one, and the customer only finds out two hours in. Suppliers differ on how they handle it and you will eat some.

Box damage. Rigid boxes crush and tins dent. Neither affects the puzzle and both affect the review, especially when the item was a gift handed over in its packaging.

Colour on board. Glossy lamination is kind to saturation, but board is not paper, and deep blacks and subtle gradients read differently than on your monitor. Sample your own artwork before deciding your colours are fine.

A busy image is genuinely harder to assemble. Sellers underestimate this. Large areas of flat single colour, a plain sky or a solid background, are miserable to work through, and so are very repetitive patterns. Hobbyists say so in reviews, by name. Artwork with varied detail spread evenly across the board is not only prettier, it is a better product, and it earns better ratings.

Personalised returns are not returns. You cannot restock a puzzle of somebody's dog.

Intellectual property. Puzzles attract fan art and character illustration the way posters do, and the stakes are identical. If you are putting a phrase, a name or a recognisable property on one, read print on demand trademark before you design rather than after a takedown.

Validate before you commit design time

The principle is boring and it is the only thing separating a shop that sells from a shop that does not: confirm real buyers exist for a niche, and see how many sellers already serve them, before you make anything.

Search volume alone is a trap. Plenty of puzzle phrases get searched constantly and are so thoroughly served that a new listing lands on page nine forever. The opposite trap is a niche with no competition because nobody wants it. You want money changing hands with room left for another seller, ideally with an angle nobody has covered.

One thing worth saying clearly, because it gets muddled. Studying what already sells and then making something of your own aimed at the buyer those listings are pulling in, in the register they respond to, is legitimate and it works. Copying somebody's file is not. The gap between those two is where the durable shops live.

Doing this properly means reading four marketplaces, and by hand that is an afternoon per idea.

The tool I make, since it is relevant here

Plain disclosure before this goes further. I own Trendlytic, I built the thing myself, and I have an obvious interest in you liking it. Read the next paragraph in that light.

What it does is narrow. Hand it a phrase. It reads four marketplaces in the same pass, those being Etsy, Redbubble, TeePublic and Amazon Merch, then reports two things back: whether money is moving in that phrase, and how many sellers already stand between you and it. The second number is usually the one that changes a decision. There is also a one-click lookup that sends a phrase to the live USPTO register, included on every plan, which I offer as a starting screen rather than legal guidance, because I am not a lawyer and neither is a database. It costs from $9 monthly, with a trial that runs before any money moves, and the signup page covers the rest.

It has nothing to do with printing. It will not tell you whether your customer's photo is sharp enough, and no tool will.

FAQ

Are custom photo puzzles good quality?

The board is. These puzzles are cut from laminated chipboard with a glossy printed face, the same basic construction as the ones you buy in a shop, and the pieces interlock normally. The variable is the image. A photo with enough resolution for the board size gives you a genuinely nice product, and a low-resolution phone snapshot stretched over 500 pieces gives you a soft, disappointing one.

What resolution do you need for a custom photo puzzle?

Printful recommends at least 300 DPI at the final print size and Printify says the same. In practice a 16 by 20 inch board wants an image around 4800 by 6000 pixels. A standard 12 megapixel phone photo is 4032 pixels on its long side, comfortable for a small board and clearly short for a large one. Check pixel dimensions, not file size in megabytes.

How many pieces should a custom puzzle be?

For a gift, 252 or 500. A 500 piece board finishes in an evening and is the count most buyers picture when they think of a puzzle. Drop to 30 or 110 for young children, and only offer 1,000 or more when your artwork has enough detail to be worth the hours.

Can you sell print on demand puzzles on Etsy?

Yes. Printful, Printify and Gooten all list puzzles and all connect to Etsy, where the gift and personalisation demand concentrates. The friction is collecting the customer's photograph, since Etsy's personalisation box takes text rather than an image file, so most sellers request it by message once the order comes through.

Does Redbubble sell jigsaw puzzles?

It does, and its shop pages are indexed for several piece counts up to 1,000. I could not load the page to verify current specifications, because the site returned a 403 to my request, so confirm the available sizes and your margin in your own Redbubble dashboard. Redbubble is an artwork channel, so it does not support customer photo uploads.

What happens if a puzzle piece is missing?

You deal with it, usually by reshipping. A missing piece makes the whole product worthless, and the buyer discovers it hours into assembly, long after any normal complaint window feels reasonable. Supplier policies differ, so find out what yours covers before you list, and treat a small replacement rate as a cost of the category.

Worth adding to your shop?

Puzzles are a better gift product than most people expect and a worse passive product than most people hope. Sell your own artwork and you get a clean, high-ticket listing that rewards detailed work. Chase the photo puzzle keyword and you get more money with a service job attached.

Which side would you take: your own designs on a 1,000 piece board for hobbyists, or customer photos at 500 pieces for the gift market?