First pass
Remove category mismatches, titles that do not describe the item, and listings with too little visible information to compare.
Practical method · from search to shortlist
The goal is not to collect the most links. It is to leave with a small set of candidates, a reason for each one, and a clear next check.
Two minutes · reduce noise
Start with a plain target statement that includes the product type, intended use and one or two constraints. “Black low-top shoes for everyday wear, with a clear size chart” is more useful than “good shoes.” The point is not to write a perfect brief; it is to make unrelated results easy to dismiss.
Choose the closest product category next. Category-first browsing gives every candidate the same first questions. Shoes can be compared on shape, sole and sizing; hoodies on measurements, fabric and print placement; bags on scale, hardware and interior structure.
Five minutes · enough choice, less clutter
Remove category mismatches, titles that do not describe the item, and listings with too little visible information to compare.
Compare the same details across every remaining candidate: useful angles, measurements, material clues, source relevance and price context.
Keep only candidates that are meaningfully different or better. Several near-duplicates do not create a stronger decision.
Stop opening new tabs when: the next result does not add a new option, better evidence or a useful price comparison. More browsing is not automatically more research.
Three minutes · make the shortlist explain itself
Record the strongest visible advantage, such as a usable measurement chart, clearer close-ups, a better match for the target or lower packaging risk.
Name the missing fact instead of writing “check later”: exact chest width, outsole view, bag dimensions, fabric detail or packed weight.
Decide before you are attached to the item: wrong measurement, poor shape, crooked placement, missing hardware view or a source page that no longer matches.
Candidate: Zip hoodie, dark grey
Keep reason: Clear garment measurements and useful front, back and fabric photos.
Open question: Is the chest width measured flat, and does the cuff match the listing image?
Reject trigger: Remove if the warehouse measurement differs materially from the size chart or the print placement is visibly off.
Before ordering and after arrival
Write the exact photo or measurement that would confirm the reason you saved the candidate. A request is useful when it answers one unresolved question.
Check the real item against the listing, the saved advantage and the reject trigger. Do not lower the standard simply because time has already been spent.
Two minutes · value after packaging
Do not calculate value from the displayed item price alone. Note whether the item is likely light, dense, bulky, fragile or dependent on protective packaging. Shoes with boxes, padded jackets and structured bags may behave differently from thin clothing even when their item prices are similar.
You do not need to invent a precise shipping number. Record what is known, what is estimated and what still needs official confirmation. If the candidate only looks attractive under the most optimistic shipping assumption, it is not yet a strong final choice.
The useful output
A finished session leaves fewer links and better questions.
Reading path