Define one product need, remove rows that fail a basic relevance check, compare the survivors on category-specific evidence, then keep no more than five. Every saved row needs a reason, a known uncertainty, and one next check.
Before pass one: write the comparison sentence
Complete this sentence: “I am comparing ___ for ___, and the detail that matters most is ___.” A useful version might be “I am comparing lightweight zip jackets for mild weather, and measured chest width matters most.”
If you cannot write that sentence, thirty rows are thirty unrelated possibilities. Category, use, and priority create a fair baseline.
I am comparing [product type] for [use], and [one detail] matters most. My maximum shortlist is five.Pass one: relevance, not quality
Scan all rows quickly and remove anything that does not belong in the comparison. Do not zoom into every photo yet.
- Wrong product category or use.
- No identifiable destination or item context.
- Missing the single feature you marked essential.
- Duplicate row with no meaningful difference.
- Price or likely parcel weight clearly outside your working range.
The goal is to reach roughly ten or fewer candidates. A row removed here is not declared bad; it is simply not relevant to this decision.
Pass two: compare evidence on the same fields
Choose three to five fields that match the category. For jackets, that might be chest width, length, lining, closure, and likely bulk. For shoes, use measured fit, shape, sole, heel view, and estimated package size. Do not reward a row merely because it has more text.
| Field | What to record | Avoid writing |
|---|---|---|
| Photos | Which relevant views are present | “Looks perfect” |
| Measurements | Number, method, and comparison item | “True to size” without evidence |
| Source | Destination type and match status | “Trusted” based on platform label |
| Price | Visible option and comparable range | “Best value” without total context |
| Weight | Estimate source and uncertainty | A precise shipping promise |
Pass three: make every survivor earn a sentence
For each remaining candidate, write: “Keep because ___; uncertain about ___; next check ___.” If you cannot finish all three parts, the row is not ready for the final five.
“Keep because the flat chest and length measurements match my comparison jacket; uncertain about fabric weight; next check is a close view of the lining.”
This note is more valuable than a star rating. It records why the row survived and prevents you from reopening the same pages without a purpose.
Give each finalist a clear reason to stay
Your five rows can fill different roles instead of being squeezed into one score:
- Best evidenced: the most important details are visible.
- Best fit candidate: measurements align with your baseline.
- Lighter or simpler option: useful when parcel weight matters.
- Alternative style: meaningfully different, not just another color.
- Needs one answer: promising only if a critical unknown is resolved.
Roles make tradeoffs visible. A single numeric score can hide the difference between missing evidence and a genuine mismatch.
When every tab starts to look the same
After many similar rows, it is easy to accept vague labels, forget which page showed a measurement, or favor the last result you opened. At that point, the notes are no longer reliable.
Pause when you cannot describe the difference between two open tabs. Close anything that did not enter the shortlist, take a short break, then resume from the written notes rather than browser history.
Use a stop rule
Stop when five or fewer comparable candidates remain, each has a reason and an uncertainty, and a new search result would need to replace an existing row. If a critical question is unanswered for all candidates, change the query to answer that question—do not simply add more products.