Blog

In-Aisle Math & Sourcing Workflow

Rapid Scanning vs. Deep Checks: How to Source a Thrift Aisle in 15 Minutes

November 11, 2026

Source a thrift aisle in 15 minutes. Rapid Scan one photo into Picks, then spend a deep check only on rows that can clear your pay-up-to price.

A thrift aisle is a timer. Color-tag day, a lunch break, an estate sale that opened at nine and will be picked through by nine-thirty. The failure mode is the same in all three: you fall in love with the third item, research it like a term paper, and never see the rest of the rack.

The fix is a two-speed route. Rapid Scan is the fast pass: one photo, the row drops into Picks, and you keep walking. A deep check is what you do when a row might actually clear your pay-up-to price. You stop, open it, read the range and the live asking prices, and then buy, negotiate, or pass. ResellFinder is built around that split. Fifteen minutes is enough for one aisle if the speeds stay in the right order.

Set the math before you park

Do not build your profit rules between hangers. In Settings, before the first photo, confirm the starting point or your own version of it:

• Target profit $15
• Platform fees 15%
• Shipping allowance $7
• Prep allowance $2

Those four numbers are the pay-up-to formula: resale × (1 − fee rate) − shipping − prep − target profit, floored at $0. On a scan, the resale input is the lower of the low end, the likely figure, and the median of active eBay asking prices when those exist. You want that choice already true when the first row hits Picks.

Turn Keep scanning on if you want the camera to stay open between snaps. Leave it off if you would rather see Picks after each photo. For a 15-minute aisle, on is the better default. You are collecting candidates, not judging them one thumb at a time.

Sign in before you start the pass. The camera can open without an account. Research cannot. A snap that never analyzes is a souvenir, not a pick.

What each speed is for

Rapid Scan answers “is this worth a second look?” One photo of the whole item. Research runs in the background. The row shows up in Picks with a preliminary range and a buy ceiling. You do not confirm a title, write a description, or build a listing. Listing copy is a job for after you own the item, in whatever tool you use to post. It is not this pass.

A deep check answers “does the tag beat the ceiling?” You open one Picks row. You read pay-up-to, the low / likely / high range, and the confidence. If eBay listings are there, you read them as asking prices for the same or similar items, not as sold comps. You look at the tag again. Then you decide.

The mistake is deep-checking everything, and the other mistake is buying off the thumbnail without opening the row. Rapid Scan fills a list. The list is not a purchase order.

Picks is the carry step, a session list rather than an inventory. At the end of the aisle, open it and split the rows:

$70 × 0.85 − $24 = $35.50. A worksheet that used the Walkman’s $85 likely figure alone would say $85 × 0.85 − $24 = $48.25. The phone uses $70. Save the keeper into the active Haul Locker if you are on Picker or above, and enter what you paid. Leave the candle holder in the session. Unsaved rows are not a death pile until you buy them.

Put the useful detail in the one photo, or in a second snap when the first one missed it: the whole item, the maker mark or model, the barcode or care tag, and the chip, stain, or crack. Skip faces, mail, and anything you would not want stored with the find.

A separate confirm flow — up to four guided photos, then a title, category, condition, completeness, function, and a quality mark from 1 to 5 — is specified and turned off in this release. There is no Scan item mode on the camera right now. The deep check you can do today is the Picks row. If the label was blurry, take another Rapid Scan of the label. Do not wait for a mode the shutter does not offer.

A 15-minute pass, timed

This is one clothing aisle and the endcap of housewares. Twelve candidates. You will not deep-check twelve.

Minute 0–1. Walk it once without the camera. Note the dense spots: denim, coats, the endcap with lamps and frames. You are choosing where the twelve photos will come from, not shopping yet.

Minutes 1–8. Rapid Scan, one photo each. Center the item. Include a label when it is on the front. Skip faces, mail, and anything you would not want stored with the find. With Keep scanning on, the camera is ready for the next snap.

A reasonable twelve for this aisle:

1. Corduroy jacket, tag $14.99
2. Sweater with a small hole, tag $5.99
3. Plain tee, tag $3.99
4. Leather belt, tag $4.99
5. Board game, sealed, tag $6.99
6. Stack of DVDs, $1.99 each
7. Toaster, tag $8.99
8. Picture frame, tag $3.49
9. Jeans, tag $7.99
10. Blender base with no jar, tag $6.99
11. Wool coat, tag $24.99
12. Pyrex bowl, tag $9.99

You are not doing math in these seven minutes. You are making a list while the estimates land.

Minutes 8–12. Stand still and open Picks. Sort with your eyes, not with a new research project. The only rows that earn a deep check are the ones whose pay-up-to might beat the tag, plus anything with low confidence that you still believe is the real item. Everything else is a pass you do not need to touch again.

Here is how a set of results can come back, using the starting Settings. The ceiling uses the cautious input, not the flattering one.

The jacket ceiling is $55 × 0.85 − $24 = $22.75. The tag is $14.99, so $7.76 of room remains against the $15 target. That is the deep check.

The wool coat is $48 × 0.85 − $24 = $16.80. The tag is $24.99, which is $8.19 over. You can ask for $16. You do not pay $24.99 and call it close.

The sweater’s $22 input is $22 × 0.85 − $24, which is negative, so the ceiling is $0. A hole and a $5.99 tag do not get a second photo.

Minutes 12–15. One real deep check, then checkout or the next aisle. Open the jacket. Confirm the range is about the jacket you are holding, not a costume or a kid’s size the title guessed wrong. Glance at confidence. If it is low, photograph the brand mark and run that snap, or walk. If the listings are similar jackets asking around $64, remember that those are unsold asks and that the ceiling already refused to rise to meet them. Pay $14.99. It is under $22.75.

If you are on Picker or above, save the jacket to the active Haul Locker and enter what you paid. Thrifter can make this decision and cannot save the find into Haul. That is a reason to upgrade when the route is weekly, not a reason to skip the ceiling on a free plan.

Then leave the aisle. The other eleven rows can expire in Picks. Unsaved Rapid Scan rows are a working session, not an inventory.

What a deep check includes, and what it leaves out

Ninety seconds is a full deep check. Spend it like this.

Pay-up-to, first. Compare it with the tag before you fall into the photos. If the tag is over and the seller does not negotiate, you are done.

Range and confidence, second. Low, likely, and high. A wide spread or a low confidence mark means the photo was not enough. One extra detail — the label, the size, the chip in the rim — is the retry. Do not “pick the high number” to settle the doubt.

Asking prices, third. Median of live listings, same or similar items. Filter condition if the list is a mix of new and for-parts. The list does not alter the visual range. It can lower the ceiling when the median is the smallest input. It cannot raise the ceiling when sellers are ambitious.

Your own eyes, the whole time. Stains, missing buttons, a zipper that will cost more than the $2 prep allowance, a plug you cannot test. The formula assumes the allowances you set. A repair you can already see is a reason to pass or to raise the prep allowance and watch the ceiling fall. It is not a reason to hope.

Leave these out of the 15 minutes: title writing, background removal, cross-posting, authentication lectures, and a sold-comp rabbit hole. If the item is worth owning, those jobs happen at a table later. If it is not under the ceiling, they are how you miss the next aisle.

How many snaps a plan actually covers

A 12-photo aisle is one trip, not a business model. Match the plan to the route so the meter does not die on the third rack.

Fifteen free scans is one careful Saturday, or one 12-photo aisle plus a little room to reshoot the jacket label. A weekly route burns through Thrifter on purpose. Picker is the plan that adds Haul Lockers and enough scans to work more than one store. One-time packs are $3 for 50 more AI scans or 25 more comps when a month runs hot. Packs do not renew. Unused pack credits stay until you use them. The monthly allowance still resets.

You do not need a deeper plan to learn the two speeds. You need the scans you will actually take, and you need Haul once you want the trip scored after the items sell.

The pass fails in predictable ways

Opening every row. Picks is a queue. If you deep-check the tee, the frame, and the DVDs, you have rebuilt the slow trip and named it Rapid Scan.

Buying because the likely number looks exciting. The jacket’s likely figure was $70. The ceiling used $55 and said $22.75. The likely number is allowed to be interesting. It is not allowed to be the budget when a lower input exists.

Saving the whole session. Save the keeper. Let the passes die. A locker full of $0-ceiling sweaters will make the haul look busy and the payback look worse, because investment counts what you bought, and you should not have bought them.

Waiting on eBay for every snap. Listings are optional context. If the ceiling is $0 before they arrive, walk. If the tag is far under a cautious ceiling and confidence is decent, you can buy and let the listings finish in the car. If the tag is within a dollar or two of the ceiling, wait for the median. That is the lamp problem: a $6 sticker can fit a $10 ceiling and miss a $4 ceiling once softer asks come in.

Run the next aisle on a clock

Fifteen minutes, twelve photos, one or two deep checks, one purchase that is under the pay-up-to price. That is a sourced aisle. The rest is walking.

Get ResellFinder for iPhone, turn Keep scanning on, and give one aisle a clock the next time you go out. Allowances and Haul, which starts on Picker, are on the plans page.

Related reading

• The Pay-Up-To Formula
• Active Asking Prices vs. Sold Comps
• Did Your Sourcing Trip Pay for Itself?

Estimates only. Rapid Scan returns a cautious range and a pay-up-to price, not an appraisal and not sold comps. eBay rows are asking prices for the same or similar items. If research fails, the price stays blank.