Weeding by feel is the fastest way to stall a collection project and burn out staff. Someone pulls a dusty hardcover. Someone else insists it’s a hidden gem. The cart sits in the workroom for three weeks while everyone argues about literary merit, and then it gets wheeled back into the stacks untouched because nobody has the energy to fight. Circulation data doesn’t end those debates entirely, but it shrinks them down to a manageable size. When the numbers say a book hasn’t left the shelf in four years, the conversation shifts from “should we keep it” to “is there a specific, articulable reason to override the data.” That’s a faster, less personal conversation, and it produces a weeding routine you can actually repeat every year without dread.
1. Pull the Right Circulation Report Before Touching a Shelf
The report you need isn’t the one your ILS calls “weeding list.” Those canned reports often bundle in items by age alone, ignoring whether the title circulated heavily three years ago and still moves during summer reading. Start by requesting a custom item-level extract from whoever runs your ILS reporting. If that’s you, even better. The fields that matter: barcode, title, author, call number, shelving location, collection code, item status, total checkouts (lifetime, not just the current fiscal year), last checkout date, date added, copy count, and replacement price. If your system tracks in-house use via scans, include that too, though treat it as supplementary. In-house use data is notoriously inconsistent across branches and staff habits.
Ask for a window that covers items added at least three years ago, and ideally five. A three-year window catches the titles that never found an audience. A five-year window gives you more confidence with slower-moving nonfiction topics where a single checkout every couple of years might still justify shelf space in a smaller collection. If your ILS allows it, run the report against all statuses except “discard” and “billed,” so you’re looking at everything theoretically available to patrons. Export the results into a spreadsheet. The goal is a single file you can sort, filter, and annotate without logging back into the ILS every time you have a question. This report is the foundation of every decision that follows, so get the fields right on the first pull. Re-running it mid-project because you forgot copy count or last checkout date will cost you a morning and a chunk of staff goodwill.
2. Set Cutoff Rules That Match Your Building and Budget
Before anyone scans a barcode, write down the thresholds that will flag an item for weeding consideration. These are not commands to delete. They are triggers for review, and they need to be clear enough that a shelver working a Sunday shift can apply them without texting a supervisor. Start with the defaults, then adjust for your specific building size, hold queue pressure, and budget reality. The point is consistency. If two staff members look at the same report and arrive at different weeding piles, your thresholds aren’t specific enough.
Default thresholds by collection
For adult fiction, a reasonable starting threshold is zero checkouts in the last 36 months, or fewer than two checkouts in 60 months with a last checkout date more than 24 months old. Adult nonfiction benefits from a longer leash: zero checkouts in 48 months, or fewer than three in 72 months, especially for topics like health, finance, and technology where currency matters more than raw usage. Children’s and teen collections move faster and show wear sooner. I’d flag anything with zero checkouts in 24 months and fewer than three in 48 months. Media — DVDs, audiobooks, music CDs — should follow the fiction thresholds but with an additional condition: pull anything with a last checkout more than 18 months old and fewer than five lifetime checkouts, because media formats degrade and patron expectations for availability are higher. These numbers aren’t sacred. They’re starting points that produce manageable cart lists without gutting a collection on the first pass.
Budget-linked exceptions
A tight materials budget changes the math. If your replacement budget is frozen or cut, you can’t afford to keep marginal titles on the shelf hoping they’ll circulate someday. Lower the nonfiction threshold to 36 months of zero checkouts. Lower media to 18 months regardless of lifetime checkouts. Every title you keep that doesn’t move is a title you’re not buying that would. The shelf space is the real budget line, and when money is short, you need that space working harder.
The reverse exception applies when replacement costs are high. A specialized art book that cost $85 and circulates once every three years might still earn its keep if replacing it with anything comparable would eat half a month’s nonfiction budget. Set a replacement-price floor — say, $50 — above which you raise the circulation threshold by 12 months. That $85 art book needs to have zero checkouts in 60 months, not 48, before it hits the list. The same logic applies to multi-volume sets where discarding one volume breaks the set’s usefulness. Flag those on the report before anyone pulls them, and route them to a separate review pile.
3. Clean the Report Before It Becomes a Cart List
The raw report will include items that look like weeding candidates but aren’t. Pulling them wastes time and creates unnecessary hold requests when patrons notice gaps. Before you print shelf lists or assign cart sections, filter out items with active holds. If a patron is waiting for it, the book stays, regardless of its checkout history. Next, check for on-order replacements. If your acquisitions department already has a new copy in the pipeline with the same ISBN, the old copy can go, but track it separately so you’re not counting it as a deselection that reduces collection size.
Remove duplicate records. An item barcode that appears twice in the report — once as a circulating copy and once as a reference or storage copy — needs a cataloguing cleanup, not a weeding decision. Flag it for technical services and move on. Missing items should be pulled from the report entirely. If the ILS status says “missing” and the last checkout date is three years old, the item is already gone. Marking it as weeded in the system is a cataloguing task, not a shelf task. Don’t let it clutter your cart list.
The trickiest filter is for floating collections and shared branches. If your system floats materials between locations, an item with zero checkouts at your branch might have circulated steadily at another branch before landing on your shelf. The total checkout count will catch some of these, but not all. Sort the report by copy count and look for titles where your branch holds the only copy with low circulation. If the system-wide copy count is high and all copies show low use, the title is a legitimate weeding candidate. If your copy is the only low performer in a system full of active copies, hold off. The problem might be your shelving location or local demand, not the title itself.
4. Manage the Edge Cases Without Letting Them Stall the Project
Edge cases are where weeding projects die. A staff member pulls a list of 400 items, notices a local author’s book on it, and suddenly the whole project pauses while everyone debates the merits of keeping “important” but uncirculated titles. The solution isn’t to ignore edge cases. It’s to contain them in a process that has a hard time limit and a hard cap on volume.
Quick review pile
Set a rule before the project starts: no more than 5% of the weeding list gets pulled for human review. On a 400-item list, that’s 20 titles. Staff can flag anything they want, but once the review pile hits 20, nothing else gets added unless something else comes off. The review itself gets 30 minutes, not a committee meeting spread across three weeks. Two people scan the pile, make a yes-or-no decision on each title, and move on. Any title that survives the review pile gets a note in the item record explaining the override — “local author,” “award winner,” “curriculum tie” — so the next weeding cycle doesn’t re-litigate the same decision.
Local history and reference
Local history and reference materials need their own thresholds, because zero circulation often doesn’t mean zero use. A city directory from 1987 may never check out but gets consulted at the desk twice a month. For local history, the weeding trigger should be condition and duplication, not circulation. Keep one copy in the best condition. If a digital surrogate exists and is accessible, you can consider discarding duplicates. For reference, in-house use data is the closest proxy you have, but it’s imperfect. If your system doesn’t track in-house use reliably, apply a longer window: no recorded circulation and no staff recollection of use in 60 months. Anything flagged under that rule still goes through the quick review pile, but the default should be retention unless the item is superseded by a newer edition or available digitally in a stable format.
5. Turn the Clean List into a Staff Workflow
A clean, sorted list is useless if it sits in a shared drive while everyone waits for a slow Tuesday that never arrives. Break the list into shelf-location chunks — adult fiction A–D, adult nonfiction 000–399, children’s picture books, and so on — and assign each chunk to a specific person with a specific cart limit and a specific deadline. A good cart limit is one shelf section per shift, roughly 30–50 items depending on how dense your stacks are. More than that and the puller starts scanning barcodes without looking at condition, which defeats the purpose of having a human involved at all.
Give each shift a printed list sorted in shelf order. The staff member walks the section, pulls the items on the list, checks condition, and scans each barcode into a project-specific status in the ILS — “weeding review” or whatever your system allows. Items that are in good condition and might still circulate get a different status and route to a separate cart for possible relocation or display. Items that are damaged, dated, or worn get moved directly to discard processing. The key is that no cart stays in the workroom longer than two weeks. Set a completion window for the entire project: 14 calendar days from first pull to last discard. If a section isn’t finished in that window, the remaining items stay on the shelf and get revisited in the next cycle. A half-finished weeding project is worse than no weeding at all, because it trains staff to ignore carts and trains patrons to expect inconsistency.
6. Record What You Pulled and Check the Results Six Months Later
The final step is the one most libraries skip, and it’s the one that makes the next weeding cycle faster and more defensible. Log the deselection counts by collection: how many items pulled from adult fiction, adult nonfiction, children’s, teen, media. Note the thresholds you used and any exceptions applied. Store that log somewhere accessible, not in a staff member’s email archive.
Six months after the weeding project finishes, run a circulation report on the items that remained in the weeded sections. Compare the post-weed circulation to the same six-month period from the previous year. The goal isn’t to prove that weeding caused circulation to rise — too many variables affect that number — but to check whether the remaining items are performing better on average. If the average circulation per item in the weeded section increased, your thresholds were reasonable. If it stayed flat or dropped, your thresholds might have been too aggressive, or the collection has deeper demand problems that weeding alone won’t fix.
This follow-up report also gives you ammunition when a board member or administrator asks whether the weeding project was worth the staff time. A spreadsheet showing that the weeded adult nonfiction section saw a 15% increase in average circulation per item over the next six months is a concrete answer. A shrug and a vague sense that the shelves look better is not.
Weeding by circulation data doesn’t remove judgment from collection management. It moves judgment to the right place: the exceptions, not the routine decisions. The routine decisions — the 95% of items that haven’t circulated in years and have no holds, no replacement on order, and no local relevance — should be fast and mechanical. Save your energy for the edge cases, and set a timer when you do. The materials budget, the shelf space, and the patrons looking for something to read will all be better off.