The Rack Scan: A Brand System for Thrift Store Racks
Flip Johnson, Summit Listing Engine house writer ·

It's 8:15 p.m. I've got 90 minutes before the store closes and about 300 items on the rack in front of me. If I stop to look up every label, I'll leave with two items and a sore thumb.
This is the real brand problem. It isn't knowing which brands are worth money. It's making a few hundred decisions in an evening without turning each one into research.
So I stopped treating brands as trivia and started treating them as a machine. Here's how it works.
The big picture: five stages, one job
The machine has one job: get a yes or no on each item in seconds, and spend real thinking time only on the few that need it.
It has five stages, and each one filters out more items than the last:
- Brand tiers. A short list you've memorized, sorted by how much you trust each brand.
- The five-second scan. You read the rack with your hands and eyes, not your phone.
- The comp gate. You check sold prices, but only for items that earned it.
- The math gate. You run cost, fees, shipping and time against a minimum profit.
- The sales log. Every sale feeds back into the tiers, so the list gets smarter.
Most resellers only do stage 1, and they do it badly. They carry a giant list in their head, buy on a hunch, and wonder why bins fill up with brands that were hot once.
The stages connect. A good tier list makes the scan fast. A fast scan means you can afford to comp the borderline items. The math gate keeps you honest. And the log keeps the whole thing from going stale.
Why a brand name alone isn't a decision
A label tells you who made the item. It doesn't tell you whether anyone wants it, in this size, in this condition, right now.
I learned this on a stack of mall-brand jeans. They had a name I'd seen sell, and they cost $4 each. I bought three.
One sold for $14 after seven weeks. The other two went to the donation bag. That's $12 spent, about $14 earned before fees and shipping, and a couple of hours I won't get back.
The brand was real. The sell-through wasn't. Sell-through is simply the share of what you list that actually sells, and it matters more than the name on the tag.
Stage 1: Build tiers you can actually remember
My first brand list ran past 140 names. I could recall maybe 20 of them in the store, and the rest were decoration.
So I cut it down to three tiers.
Tier 1: Grab. Brands that have sold for me quickly and at a margin I like. If the condition passes, they go in the cart without a phone check, as long as the price is under my cap (more on that in stage 4).
Tier 2: Check. Brands that sometimes pay and sometimes don't. The brand isn't enough. Style, size and condition decide it, so these get a comp check.
Tier 3: Skip. Labels that show up everywhere and sell for what they cost. I don't need to feel bad about leaving these on the rack. That's what the rack is for.
Keep each tier short. Ten to fifteen names in Tier 1 is plenty. If you can't recite it while walking an aisle, it's too long.
Sort by category, not just by name
It helps to group your list by what you're looking at on the rack, because that's how you walk the store. Here are examples of the kinds of labels that tend to land in a reseller's upper tiers. Your own tiers should come from your own sales, not mine:
- Outerwear and outdoor: Patagonia, The North Face, Arc'teryx
- Workwear and denim: Carhartt, Levi's
- Activewear: Lululemon
- Wool and heritage: Pendleton, Ralph Lauren
- Boots and shoes: Red Wing, Dr. Martens, Birkenstock, Allen Edmonds
- Bags: Coach, Dooney & Bourke, Kate Spade
- Accessories: Ray-Ban frames, leather belts, silk scarves
This list is a starting point, not a verdict. A brand that moves fast for one seller can sit for another, depending on size range, price point and what you list it next to.
Add a wildcard slot
Some of your best flips won't be on any list. The label is unknown to you, but the material is good: real leather, wool, silk, cashmere, or heavy construction you can feel.
So I keep a fourth slot called Wildcards. Material, weight and build earn an item a comp check even when I don't recognize the name. It's a small door left open, so the list doesn't become a wall.
Stage 2: The five-second scan
This is where the time savings live. Here's the math.
If I pull out my phone for every item, I spend about 90 seconds each. Across 300 items, that's 27,000 seconds, or 7.5 hours. I don't have 7.5 hours. I have 90 minutes.
With a five-second scan, those same 300 items take about 25 minutes. Then I comp roughly 15 borderline items at 90 seconds each, which adds about 22 minutes. The whole rack takes under an hour, and I still have time to check the shoe wall.
Here's what the scan looks at, in order:
- Label. Is it a Tier 1, Tier 2 or Wildcard name? If it's Tier 3 or nothing, move on.
- Fabric. Read the content tag. Natural fibers and heavy, well-made fabric get a second look. Thin, pilled polyester usually doesn't.
- Condition. Check the armpits, collar, cuffs, hem and knees. Hold it to the light for stains and holes. If it needs a repair or a stain fight I haven't won before, it goes back.
- Size and season. Sizes in the middle of the range tend to move easier for me than the extremes. A heavy coat in the wrong season isn't dead, but it's frozen cash for months.
The order matters. The cheapest checks come first, so most items die at step 1 and cost me two seconds.
The one-hand rule
I scan with one hand on the hanger and my phone in my pocket. If I haven't said yes or maybe by the time I've checked the tag, the item goes back.
It feels rude to the rack. It's also how I get through the whole store before closing.
Stage 3: The comp gate
Comps are recently sold listings for the same or very similar item. They tell you what buyers actually paid, not what sellers hope for.
I only run comps on Tier 2 items, Wildcards, and any Tier 1 item that's unusual for its brand. That's the whole point of the gate. It protects your phone time for items where it can change the answer.
When I comp, I look for three things:
- Sold, not active. Asking prices are wishes. Sold prices are data.
- Same style and size. A brand's best-selling jacket and its forgettable one carry the same label and very different prices.
- How many sold. One sale at $80 is an anecdote. Eight sales between $35 and $50 is a price.
If I can't find at least a few sold examples, I treat the item as unproven and pay accordingly, or skip it.
Stage 4: The math gate
This is the stage that stops good-looking flips from being quietly bad ones.
The formula is simple. Expected sold price, minus fees, minus shipping, minus what I paid. Whatever's left is profit, and then I ask how long the item will take me.
Note that these are my own planning numbers. For fees, I assume roughly 15% of the sale price as a rough placeholder. Fees differ by platform and change over time, so plug in what your platform charges today.
A good flip (my example). A fleece jacket costs $6. Comps say it sells around $40. Fees at 15% come to $6. Shipping is about $8. Profit is $40 minus $6 minus $8 minus $6, which is $20.
Photos, measurements, listing and packing take me about 25 minutes total. That's $20 for 25 minutes, or about $48 per hour. That's a good night shift.
A bad flip (my example). A basic blouse costs $5. It sells around $14. Fees are about $2.10, and shipping is about $5. Profit is $1.90.
At 25 minutes of work, that's about $4.56 per hour, plus months of shelf space. The label looked fine. The math says no.
Set a buy cap per tier
To save time, I turned the math into caps. For Tier 1, I'll pay up to a set amount without checking. For Tier 2, I'll pay only after a comp. For everything else, the cap is whatever leaves me at least $15 in profit.
Why $15? Because below that, a flip doesn't pay for the evening it costs me. Pick a minimum that works for your time, write it down, and stick to it.
And weight matters. A heavy coat or boots can eat a big share of the sale in shipping, so the same label can pass at one weight and fail at another.
Stage 5: The sales log
This is the stage nobody wants to do, and the one that makes the machine improve.
For every item I sell, I log four things: brand, cost, sold price, and days to sell. That's it. It takes about 30 seconds per sale.
Once a month, I look at the numbers by brand. A brand that keeps selling fast moves up a tier. A brand that sits for 90 days moves down or gets cut.
I had a small script, put together with some AI help, that sorts my sheet by brand and days to sell. You don't need one. A plain spreadsheet and a few minutes of sorting does the same job.
Twenty sales from a brand is a decent sample. Two sales is a story you're telling yourself.
Where the machine breaks
No system is perfect. Here's where this one leaks.
Counterfeits cluster around hot brands. The more in-demand the label, the more fakes exist, especially in bags and sneakers. A Tier 1 name is not proof an item is real. Check tags, stitching, hardware and materials, and if the item still feels off, walk away. A $12 pass beats a refund and a lost buyer.
Trends move. A brand that sold in a hurry two seasons ago can go quiet. This is why the log matters, and why the list needs a monthly check.
Your bins are a bottleneck too. A perfect brand strategy still fails if you've got ten bins waiting to be listed. If your backlog is already full, tighten the math gate and buy less until it clears. Buying faster than you list isn't sourcing. It's freezing cash.
Your first evening with the machine
You don't need to build all five stages tonight. Here's a plan that fits in one evening:
- Write a Tier 1 list of 10 brands you've already sold or seen sell. Put it in your phone notes.
- Write a Tier 3 list of the labels you keep passing over. Naming them out loud makes them easier to skip.
- Next trip, scan with your hands first and only pull out your phone for Tier 2 and Wildcards.
- Before you buy, run the profit math. If it's under your minimum, put it back.
- Start the log. Four columns. Five minutes.
It won't be perfect on the first night. Mine wasn't. But by the third trip you'll be faster, and by the tenth you'll know which names deserve cart space.
The short version
A brand list isn't a treasure map. It's a filter.
Tier it, scan it, comp the maybes, run the math, and log the results. Each stage saves time for the next.
Buy the label that sells, not the label that sounds good. Then go list something.
Flip Johnson is the house writer for the Summit Listing Engine blog: a pen name for posts written by the Summit Listing Engine team, not an independent reseller or customer review.
Working out what a sale really nets? Try our free reseller calculators.