Blog

Competitor Comparisons

Google Lens vs. ResellFinder: Why Visual Identification Isn't Enough for Resellers

December 2, 2026

Google Lens can name a thrift find. It does not compute a pay-up-to price after fees, shipping, prep, and your profit target. Here is how to use a name without overpaying.

Google Lens is good at a question resellers ask constantly: what is this? Point it at a jacket, a bowl, or a small appliance and it will often return a name, a similar image, and a row of shopping pages. That feels like pricing, because the shopping pages have dollar signs.

It is identification, plus whatever prices happen to be attached to those pages. A product title and a retail offer are not a thrift buy decision. ResellFinder starts from a photo too, then does the part Lens is not built to do: a cautious low / likely / high resale range, and a pay-up-to price after your fee rate, shipping, prep, and profit target. The name of the item is an input to that math. The name is not the math.

What you actually get back from a visual search

Lens, and visual search like it, matches a photo against images on the web. When the match is strong, you get a useful label: the brand, the line, sometimes the decade. When the match is weak, you get a neighbor — the same shape, a different year, a reprint, a “inspired by” product that borrowed the silhouette.

The shopping cards under that label are a mixture. Some are current retail for a new item. Some are ads. Some are other people’s marketplace asks. Some are articles. Almost none of them know your fee rate, your shipping allowance, or the profit you require before you will tie up cash. They also do not know that your item has a faded cuff and a replaced button.

So a Lens result can be the right first step and the wrong last step. “This is a Levi’s trucker” is progress. “A store is advertising a new one at $128” is a different object in a different condition on a different kind of price. Paying $18 in a thrift store because a new jacket is $128 is how the afternoon gets expensive. The $128 is not your resale. It is not your net. It is not your ceiling.

The dollars a name still does not compute

Say the visual search is right, and the item in your hand really is that jacket. You still need a spending limit. ResellFinder’s formula, and the formula you can run on a calculator, is:

Buy ceiling = resale × (1 − fee rate) − shipping − prep − target profit

Starting Settings are $15 target profit, 15% fees, $7 shipping, and $2 prep. Anything under zero becomes a $0 ceiling.

You still have to supply a resale figure. Lens did not give you a cautious reseller range. It gave you a name and a pile of pages. If you grab $128 off the retail card and stuff it into the formula, you get a nonsense ceiling:

• $128 × 0.85 = $108.80
• $108.80 − $24 = $84.80

An $18 thrift tag looks like a gift against an $84 ceiling. It is a gift only if a used jacket will actually bring $128, which is the new-item ask you just borrowed. That is the error. The formula is only as careful as the resale you feed it.

Put a cautious used range in instead. Low $55, likely $70, high $88, and a median of live eBay asking prices at $64 for the same or similar jackets. ResellFinder uses the lower of the low end, the likely figure, and that ask median. The input is $55.

• $55 × 0.85 = $46.75
• $46.75 − $7 − $2 − $15 = $22.75 pay-up-to

Now the $14.99 tag fits, and a $28 tag does not. The $128 retail card never entered the limit. High, at $88, did not enter it either. The live asks at $64 did not raise it. Identification told you which jacket to think about. The ceiling told you what to pay.

A second thrift classic: Lens names a DVD, and a retailer still lists a new copy at $14.99. Your disc is used, the case is cracked, and your shipping allowance is $7. A cautious resale of $20 at the starting Settings is a $0 ceiling, because $20 × 0.85 − $24 is negative. The visual search was correct about the title. The title’s retail price was irrelevant. The formula is the part that keeps the $1.99 bin from becoming a trunk full of media you cannot ship profitably.

A mug makes the same split in one glance. Lens returns an Etsy ask at $45, a Pinterest photo, an out-of-stock retailer page, and a blog post from a few years ago. Those are matches. They are not a buy price. ResellFinder on the same mug:

$32 × 0.85 − $24 = $3.20. A likely-only worksheet on $38 would say $38 × 0.85 − $24 = $8.30. Neither number is $12, and neither number is the $45 Etsy ask. Fifteen Lens searches in one aisle become fifteen tabs. Fifteen Rapid Scans become one Picks list you can accept or pass at the bench by the door.

Lens is the better general identifier when you truly do not know the name. ResellFinder is the spending limit after you do.

Where a name helps, and where it should stop

Use visual search, or your own eyes, for the label.

• Brand, model, artist, or pattern, when the mark is readable
• A reissue versus an older version, when the photos on the result page show the difference
• A reminder of what the complete item includes, so you notice the missing lid or the missing charger

Stop there. Do not let the result’s price become the likely resale. Do not average a new retail price with a used ask and call it a comp. Do not treat “similar images” as sell-through. A popular photo is not evidence that used copies are clearing.

Then open the sourcing app for the buy. In ResellFinder, one photo returns the cautious range and the pay-up-to price. Optional eBay listings underneath are live asks for the same or similar items. The search uses text. The app does not post your photo to eBay. You can filter those asks and look at the median, and you should remember the rule from the asking-price article: the median can pull the ceiling down, and it cannot pull the ceiling up. Sold comps are not what that list is.

If the scan fails, the price stays blank. That is a better outcome than borrowing $128 from a shopping card to fill the silence. Take another photo of the label, or pass.

Condition is the part a web match smears

Visual search is attracted to clean catalog photos. Your item is the one with the flaw. The flaw is often the whole margin.

A bowl with a chip, a toaster with a frayed cord, a jacket with a cigarette burn on the cuff: the name can still be right. The resale a catalog photo implies is for a better example than yours. ResellFinder’s range is a cautious visual estimate, and confidence drops when the photo does not support a tight call. Default quality sits in the middle. If you mark the item worse than that, the numbers you are buying against should move down. A sourcing limit that ignores damage is a retail price in a thrift disguise.

The prep allowance is the other half of condition. The starting $2 covers light cleaning and a small supply, not a new zipper and a tailor. If you can see a repair that will cost $15, either pass or put that $15 into prep and let the ceiling drop by $15. On the jacket, a $22.75 ceiling with an extra $15 of prep becomes $7.75. The $14.99 tag no longer fits. Lens will not do that subtraction. It does not know your tailor.

Privacy and what the photo is for

A Lens search sends the image to Google so the match can run. That is the product. Read Google’s current privacy terms if you want the retention details, because they are Google’s, and they change.

A ResellFinder photo is for the estimate and for your account. Item photos saved to the account are private. They are not a public listing, and they are not used as live marketplace proof. New photos are kept up to 90 days unless you delete the find or the account sooner. The eBay check does not upload that photo as the search. If you would not want a particular scene stored, keep it out of the frame: other people’s faces, addresses, paperwork.

Neither behavior makes the other app “the private one” in every sense. They do different work with the picture. One work is web-scale identification. The other is an account-backed sourcing record and a spending limit. Choose on purpose when the item is ordinary housewares. Choose more carefully when the photo would capture something you do not want in a search index.

A two-step aisle habit that uses both

You do not have to boycott a tool that is good at names. You do have to stop it before it sets a price.

1. If you cannot tell what the item is, use visual search or read the mark yourself until you have a name you trust.
2. Photograph it in ResellFinder. Read the pay-up-to price before you read the likely number out loud.
3. Compare the ceiling with the tag. Under the ceiling, with confidence you can live with, buy and record what you paid. Over the ceiling, negotiate to the ceiling or put it back.
4. If listings are present, treat them as asking prices. Let a lower median tighten the ceiling. Ignore a higher one.
5. Save keepers to a Haul Locker when you are on Picker or above, and score the trip after they sell. A correct name does not tell you whether the Saturday paid for itself. Payback does.

Rapid Scan is the version of step 2 that keeps you moving: one photo, a row in Picks, the camera ready for the next item. Deep-check the rows whose ceilings might beat the tag. The name you already learned from Lens makes those rows easier to trust, because you can see whether the estimate is talking about the same object. It still does not replace the $22.75.

Identification is the start of the buy

Lens answers “what am I holding?” ResellFinder answers “what is the most I can pay and still keep $15 after a 15% fee, $7 shipping, and $2 prep, using the cautious end of the range?” Those questions belong in sequence. The shopping-card price belongs in neither one, unless you are buying that new item from that store.

Get ResellFinder for iPhone and run the next named find through the ceiling before it goes in the cart. Thrifter is free, with 15 AI scans a month. Haul Lockers start on Picker. The list is on the plans page.

Related reading

• The Pay-Up-To Formula
• Active Asking Prices vs. Sold Comps
• Rapid Scanning vs. Deep Checks

Estimates only. ResellFinder’s range and pay-up-to price are cautious estimates, not an appraisal, not sold comps, and not a brand authentication. Live eBay rows are asking prices for the same or similar items. If research fails, the price stays blank. Google Lens is a separate product with its own pricing and privacy terms.