Five features matter most for an independent bookshop: fast ISBN lookup, special orders tied to a customer and deposit, a separate record for used copies with their buy-in cost, supplier ordering, and reports that show sell-through by title. AI reorder suggestions are a useful sixth, but only as drafts you approve.
A general retail till can cover some of these, and some bookshop systems cover all of them. The way to choose is to run your own awkward cases through a live demo, and the sections below give you the cases, a scorecard and the vendor facts I could verify.
Why a bookshop is not a general shop
A clothes shop sells items that come in sizes. A bookshop sells titles, and each title has several editions, formats and states. The same novel may exist as a hardback, a paperback, a large-print edition and three used copies in different condition, each with its own price and cost. A customer may also come in not to buy something on your shelves but to ask you to find it, which creates a promise that sits outside normal stock.
That is why four features carry the decision. They are listed here in the order in which they bite when missing:
- Special orders. A request attached to a named customer, with a deposit, a supplier, a promised date and a status.
- Used-book buy-ins. Recording what you paid, what condition the copy is in, and a price band, usually at speed across a counter with someone waiting.
- ISBN lookup. Scanning or typing a code and getting title, author, publisher, edition and list price without retyping.
- Supplier ordering. Building an order for a wholesaler or publisher from what sold and what customers asked for.
Reports and AI come after these. A beautiful sales dashboard is no help if a customer's special order disappeared into a note field.
What the vendors say they do
I checked three systems that come up when owners compare. These are from each vendor's own pages in October 2026; anything not listed below I could not verify and you should ask about it.
| System | Price shown | Book-trade features listed | Not stated on the page |
|---|---|---|---|
| Bookmanager | Not on the page I read | Special orders, used books, ISBN lookup, ordering from wholesalers | How each feature works, and its cost |
| Square for Retail | Free at $0, Plus at $49 per location a month, Premium by quote | Purchase orders and vendor profiles, barcode labels, low-stock alerts, customer profiles, sell-through tracking and slow-stock identification; barcode labels and multi-location stock on Plus | Special orders, layaway, ISBN metadata lookup, a used-book buy-in workflow |
| Lightspeed Retail | Basic $89, Core $149, Plus $289 a month; annual billing saves up to 18% | Purchase-order sync, a mobile scanner app for counts, loyalty linked to customer accounts | Special orders, ISBN lookup, used books |
Read the last column carefully. "Not stated on the page" is not the same as "cannot do". It means you have a question to ask before you pay. Lightspeed, for instance, may handle special orders through a workaround or an add-on, but the page I read did not say so. Square's page covers purchase orders, vendor profiles and sell-through tracking, which suit reordering, but it says nothing about special orders, ISBN metadata or buying used stock from customers.
Bookmanager's page names exactly the features this tutorial is about, but gives no detail on how they work, which is a reason to ask for a trial rather than a reason to rule it out. Prices on all three pages can change, and extra fees such as payment processing and hardware are separate from the plan figure.
Special orders: the feature to test first
A special order has a life cycle, and the till must follow it. If it cannot, you rebuild the cycle on a clipboard, which is how customers end up waiting three weeks without a call.
The record a special order needs
Here is an illustrative, filled-in record for one customer request. Your system should hold each of these fields, or let you add them:
| Field | Example entry |
|---|---|
| Customer | [first name], mobile saved, OK to text |
| Title and edition | Paperback, ISBN scanned, not the hardback |
| Quantity | 2 copies |
| Deposit | $10 taken, balance due on collection |
| Supplier | Main wholesaler, ordered Tuesday |
| Expected date | Friday, with a note if it slips |
| Status | Ordered, Received, Customer notified, Collected |
| Held until | 14 days after the notification |
The three that most systems miss are the status changes, the held-until date and linking the received copy back to the customer. When the box arrives, the person unpacking it should scan the book and see "special order for [first name], call them", rather than shelving it. Ask any vendor to show that moment.
A demo script for special orders
1. Create a special order for 2 copies of a title I choose,
taking a $10 deposit by card.
2. Order it from a supplier. Show me where the order is recorded.
3. Receive the book. Does the till tell the person unpacking
that it belongs to a customer?
4. Send the customer a text from the till. Does it keep a record?
5. Collect the book and take the balance. Does the deposit
come off the total automatically?
6. Now cancel a different special order. Where does the
deposit refund go?Six steps take ten minutes and expose most gaps. Do the cancellation too; refunds of deposits are where small shops tend to find that their till treats a deposit as ordinary revenue.
Used books: buy-ins at the counter
Used stock is where margins can be best and bookkeeping worst. A customer places a bag of books on the counter, and in a few minutes you must decide which to take, what to pay, and how to price each copy, with a queue behind them.
A worked example with made-up numbers: someone brings in 40 books. You take 26. You decide cash is 25% of the intended selling price and shop credit is 40%, a policy of your own. Your 26 books are priced at an average of $8, so intended sales are $208. Cash offer: $52. Credit offer: about $83. If you sell all 26 at average price you gain $156 on a cash buy-in, or $125 on credit. Realistically, some will not sell, so you keep a sell-through figure per section rather than assume everything moves.
What the till has to do here:
- Record each copy with its cost, even if only as a batch, so you can see margin by section.
- Offer price bands, so condition grading becomes a quick tap rather than a typed price.
- Handle shop credit as a customer balance rather than a gift-card workaround.
- Cope with books that have no ISBN, such as older editions and ex-library copies, without forcing a fake code.
- Keep used and new copies of the same title linked in search, so staff see both.
A demo script for buy-ins
1. Buy 5 used books from a customer: 3 with ISBNs, 1 with a
worn barcode, 1 with none. Pay half in cash, half in credit.
2. Price them using condition bands.
3. Sell one of the used copies to a different customer.
4. Show the margin on that sale using the buy-in cost.
5. Show the customer's remaining credit balance.If the vendor says used-book handling needs a manual workaround, ask what the workaround is, and whether you can see its cost in a report. A mess at the counter becomes a mess at year end.
ISBN lookup and the editions problem
Scanning an ISBN should return the title, author, publisher, format and list price. Even where it does, the work is not done, because the same work exists in several editions, and an old code may point to an out-of-print version. Test it with three titles:
- A current bestseller, to see how quickly it returns metadata.
- A title with a paperback and hardback edition, to see that you can tell them apart on screen.
- An older book with a ten-digit code, to see whether it converts to the longer form.
Cover images and descriptions matter if you also sell online. If that is your plan, taking a local shop online with AI covers what to do with product listings once your records are clean.
AI reorders: what to expect and what to refuse
AI reordering means software looks at sales history and proposes what to buy. Square's retail page lists sell-through tracking and slow-moving stock identification, which are the building blocks. Some platforms add AI to turn these into suggested orders. Shopify's old Stocky app stopped on 31 August 2026, with purchase orders moving into Shopify admin and its Sidekick assistant suggesting reorders. If a vendor says "AI reordering", ask what it actually looks at.
Books make forecasting harder than most goods. A title can sit untouched for a year and then sell eight copies after a radio mention. A local author's event creates a spike no history predicts. Series and school reading lists create steady demand that history does predict well. So AI is useful for the steady middle of the shelves and unreliable at the edges.
A prompt that keeps AI in a supporting role
If your till exports sales as a spreadsheet, you can use a general AI assistant to draft an order, without giving it any customer names. Remove names, phone numbers and special-order details first. Then try:
Here is a CSV of the last 12 months of sales by title for a
bookshop: title, format, supplier, copies sold, copies in stock,
last sold date. Suggest a reorder list for the main wholesaler.
Rules: never suggest more than 2 copies of a title that sold
fewer than 3 in the year; flag any title with 0 stock and 3 or
more sales as urgent; list titles that have not sold in 6 months
separately as candidates to return. Show your working in a table.An illustrative reply might list 14 titles to reorder, 3 marked urgent and 9 return candidates. What you would fix: it will not know that two of the urgent titles are a book club's pick for next month, or that a return candidate is a signed copy you are keeping. That knowledge sits with the person who knows the shelves, so the output is a draft, never an order that goes out unchecked. How small shops use AI to forecast stock and reorder goes into the method, and the comparison of AI inventory tools shows what else is on the market.
The call that keeps special-order customers coming back
A special order only earns its keep if the customer collects and comes back. The notification matters as much as the ordering. Here is a short message to save as a template in the till or on a phone. The square brackets are yours to fill in:
Hi [first name], your copy of [title] has arrived at
[shop name]. We will hold it for 14 days. Balance due on
collection: $[amount]. Reply here if you would like us to
keep it a little longer. [your name]Two details are worth copying. The hold period is stated, so you are not storing orders for ever, and the balance is stated, so there is no surprise at the till. If the till cannot send the message or log that it was sent, a shared note on the order does the same job, as long as the person who unpacks the box is the one who checks it.
A mistake that shows up in the first month
An illustrative case: a shop with a general retail till records special orders as a note on the customer's profile. A busy Saturday delivery arrives, a part-time assistant shelves three books in the right section, and nothing connects them to the notes. Two weeks later a customer comes in annoyed that they heard nothing. The book was on the shelf the whole time. The fix is not to blame the assistant. It is to make the received item announce itself, or to put a coloured band on every special-order book at the point of ordering so it never reaches the shelf.
Reading the numbers that decide next month's orders
Once stock flows through the till properly, three reports are worth opening every week, however small the shop.
- Sell-through by section. Copies sold divided by copies received over a period. Suppose your travel section received 60 copies this quarter and sold 21: that is 35%, and it tells you to buy fewer, not to discount.
- Slow stock. Titles on the shelf with no sale in six months. These are returns or display changes, and square-footage is the cost of ignoring them.
- Special-order age. Open orders by days since placement. Anything past its promised date is a phone call today.
If the system you are testing cannot produce these without exporting to a spreadsheet, score it down on reports. Square's retail page lists sell-through tracking and slow-moving stock identification; for others, ask for the report to be shown on screen using your own data.
A quick sum on used-book margin
Another illustrative calculation: a shop buys 300 used books across a month for an average of $2.50 each, a total of $750. It sells 210 at an average of $8, which is $1,680. Gross margin on the books sold is $1,680 minus the cost of the same 210 books at $2.50, or $525, so $1,155. The 90 unsold books still carry $225 of cost. Without cost per copy or per batch, you would see $1,680 of used sales and no idea whether buying or pricing was the problem. That is the value of a till that records buy-in cost.
A scorecard to compare two or three systems
Score each system from 0 to 3 on every row after the demo, then weight by what matters to your shop. These weights are examples; change them.
| Criterion | Weight | What a 3 looks like |
|---|---|---|
| Special orders | x3 | All six demo steps work without a workaround |
| Used-book buy-ins | x2 | Cost recorded per copy or batch, credit balances, no ISBN needed |
| ISBN lookup | x2 | Edition-level data returned in a couple of seconds |
| Supplier ordering | x2 | Orders built from sales and special orders, then sent or exported |
| Reports | x1 | Sell-through and margin by section without exporting |
| AI suggestions | x1 | Explains what it looked at and lets you edit the draft |
| Cost and exit | x2 | All fees in writing; data export without a charge |
A system that scores three on special orders and used books but nothing on AI is a better bookshop till than a polished system that scores the reverse. For a more general vendor checklist, use the AI software vendor scorecard alongside this one.
Three bookshops, three choices
A small shop that is mostly new stock
Say a two-person shop sells 90% new books and takes perhaps ten special orders a week. The special-order record matters most, and used books almost not at all. A general retail system with a working special-order feature (test it, as the vendor pages above do not list one) may be enough, provided ISBN scanning behaves. Free or low-cost plans such as Square's make sense for a trial, as long as you check the editions test.
A shop with a large used section
A shop where half the floor is used stock lives or dies on buy-in speed. Test the used-book demo script above first and rank systems on it. A till that treats every used copy as a one-off item with a hand-typed price will cost you minutes at every buy-in.
A shop with events and a school trade
School orders and event sales behave like bulk special orders, with deadlines. Ask whether orders can be grouped under one account, invoiced and partly paid. A system that is fine for walk-ins may fall apart when a school wants 30 copies and an invoice.
Switching without losing your stock records
Moving to a new till is the riskiest moment. Before you commit, ask how existing stock is imported: from a spreadsheet, from a supplier file or typed afresh. Keep the old system running read-only for a month. Count a section by hand, say crime paperbacks, and compare it with the new system's figure for the same shelf. If they differ by more than a few copies, find out why before you move the next section.
Also export your special-order list and check that every open order exists in the new system with its deposit. A lost deposit is a customer and a conversation you cannot recover.
A short list to take into the first demo
- Run the special-order and buy-in scripts with your own books, not the vendor's sample data.
- Ask for every cost in writing: plan, payment processing, hardware, any add-on for books.
- Ask what the AI looks at, and whether you can switch it off.
- Ask how to export everything if you leave.
- Take a week's trial with real transactions before committing.
After you have chosen and used it for three months, review it on the same scorecard. Shops change; a till that fitted a quiet January may not fit a school-order September.
Questions independent booksellers ask when comparing tills
Can a general retail POS run a bookshop?
Often yes for a small shop, if you can live without book-trade features. Check that it handles ISBN scanning, customer special orders with a deposit and notes, and a second stock type for used copies. If you have to fake those with item notes, you will outgrow it quickly.
What is the difference between a special order and a back order?
A special order is a book a customer asks you to get for them, tied to their name and usually a deposit. A back order is stock you ordered for the shelf that has not arrived. Mixing the two means customers wait without a phone call.
Does AI reordering work for books?
Only partly. Sales history helps with steady sellers and series, but a single title can spike after a review or an event. Treat AI suggestions as a draft order that someone who knows the shelves approves, and keep your special-order list separate from it.
How should I price used books at the till?
Set a few price bands by condition and format rather than pricing each copy from scratch, and record the buy-in cost for every copy. Without a cost you cannot see whether the used section earns more per square foot than new stock.
Should I trust a vendor's feature list?
Use it as a prompt for questions, not as proof. Ask for a live demo using your own awkward cases: a special order with a part-payment, a used copy with no ISBN and a title that has three editions. If the demo skips them, ask why.
Further reads
- How to Create Outfit and Styling Posts for a Boutique With AI — Social posts for a shop, with ideas that carry over to shelf picks.
- Hardware Store POS Systems: Which Features Matter and Where AI Fits — A feature checklist for an independent hardware store POS: SKU depth, trade accounts, special orders and reorders, plus what AI really does in each.
- Music Store POS System: Serial Numbers, Rentals, Repairs and AI — A checklist for choosing a music shop POS: serial-tracked stock, rentals, repair tickets, trade-ins, a test script and where AI really helps.
- Virtual Receptionist for a Small Bookstore: AI or Human Answering? — AI or human phone cover for a bookshop: monthly costs, what each does with stock checks, special orders and events calls, and a worked sum.
- Bookshop Event Ideas: Author Nights, Book Clubs and Signings With AI — A filled-in event plan, a publisher outreach email, ticket-page copy and a promotion run for author nights, signings and book clubs.
- Comic Book Store Software: Pull Lists, Back Issues and AI Listings — Set up pull lists, list back issues from phone photos and send a weekly new-release email, with AI drafting the words and you checking the facts.
- Record Store Software: Catalogue Used Vinyl and Track Stock With AI — A one-row-per-copy sheet, a shared grading standard, AI-drafted listings and a pricing check you do yourself, worked through on a 40-record box.
- AI Tools and AI Development: The Complete 2026 Guide — the AI hub, including every tutorial in the AI-for-business series.
Sources: Bookmanager, Square for Retail and Lightspeed Retail vendor pages, checked October 2026. Features not listed there are described as questions to ask, not facts.