Define one question, keep three candidates at most, copy the same fields into a small comparison note and mark missing facts as unknown. Remove mismatches first, then use evidence quality, fit for your use and uncertainty as the tie-break order. More rows do not automatically create a better decision.

Make every candidate answer the same question

A useful Pikobuy product comparison begins with more than “Which one is best?” Include the category, intended use and the two or three conditions that matter most. For example: “Which lightweight jacket has the clearest chest measurements, a visible closure and a reasonable packed-weight estimate?”

A row that answers another question may still be interesting, but it does not belong in the same comparison. Mixing unlike products creates false trade-offs and lets an attractive feature hide a basic mismatch.

Write the question above the shortlistIf the question changes, start a new comparison instead of silently changing the rules halfway through.

Use a three-candidate limit

Three candidates are usually enough to reveal the real trade-off without creating a second spreadsheet. If you have ten tabs, use a fast removal pass first: wrong category, unavailable required option, unusable measurements, missing source context or an important contradiction.

Too wide

Ten loosely related rows create more fields, more missing data and more chances to remember the wrong detail.

Decision-sized

Three comparable rows let you inspect the same photo, size and source evidence side by side.

The limit is not a claim that three options are always enough. It is a control: compare three now, and replace a weak candidate only when a new row clearly answers the same question better.

Normalize the fields before judging

Copy the evidence into your own words. Do not let different page layouts make one candidate look more complete than another.

01

Identity

  • Category and product type
  • Exact color or option
  • Size or dimensions needed
  • Intended use
02

Visual evidence

  • Useful photo angles
  • Measurement images
  • Visible construction details
  • Variant consistency
03

Size evidence

  • Chart units and type
  • Reference-item comparison
  • Category-critical dimensions
  • Stated tolerance
04

Source context

  • Destination type
  • Current title and option
  • Link behavior
  • Information date
05

Cost context

  • Reference price
  • Correct variant price
  • Estimated packed weight
  • Known exclusions
06

Unknowns

  • Missing value
  • Unclear label
  • Contradictory evidence
  • Fact that needs rechecking

Unknown is a result, not an empty box

Leaving a field blank makes it easy to forget whether you checked it. Write “unknown” and add a short reason: no chart, unreadable ruler, source page changed, variant not identified or packed weight not provided.

Unknowns matter differently. A missing decorative detail may not change the decision. An unknown shoe length, connector type or bag dimension might. Mark each unknown as decision-critical or acceptable for this use.

Do not reward confident-looking pagesA polished layout can still omit the fact that matters. Score the visible evidence, not the presentation.

Use a small evidence score, not a fake quality score

A 0–3 score can keep your comparison consistent, but it should describe the support for a field—not the product’s overall quality or the reliability of a seller.

  1. 0 — no usable evidence

    The field is absent, unreadable or unrelated to the chosen variant.

  2. 1 — weak evidence

    A claim or image exists, but the method, variant or context is unclear.

  3. 2 — useful evidence

    The field is specific enough to support the comparison, with a visible limitation.

  4. 3 — strong evidence for this decision

    Multiple consistent details answer the exact comparison question and the remaining limitation is recorded.

Score only important fields. Adding points across dozens of minor details can make a candidate win mathematically while still failing the one condition you actually need.

Match price and photos to the exact variant

The lowest visible price may belong to another size, option, bundle or quantity. Before recording a price, identify the variant that your comparison uses and confirm that the current page still associates that value with it.

Do the same for photos and measurements. If the photo set shows a large size but the price note uses a smaller one, record that distinction. A complete-looking row can still combine facts from different variants.

Better comparison note

“Reference price appears to apply to the black, standard option. The measurement photo does not identify its size, so fit evidence remains separate.”

Use a fixed tie-break order

  1. Remove hard mismatches

    Wrong category, missing required function or unavailable necessary option ends the comparison for that row.

  2. Prefer decision-critical evidence

    Clear measurements, relevant photos and identifiable variants matter before decorative detail.

  3. Prefer fewer serious unknowns

    Count only unknowns capable of changing your choice.

  4. Check practical fit

    Consider your size, use, weight tolerance and ability to recheck the destination.

  5. Use price last

    Compare price only after the candidates still answer the same question and the price refers to the intended option.

This order prevents a small reference-price difference from overriding a major evidence gap.

Finish with an explainable decision note

Three-candidate decision
Question
The category, use and critical conditions
Candidate A
Best evidence and largest unknown
Candidate B
Best evidence and largest unknown
Candidate C
Best evidence and largest unknown
Choice
Keep one, pause one, remove one
Reason
The first tie-break rule that separated them

If you cannot explain the choice in two sentences, the candidates may not be comparable or the question may still be too broad.

Know when to stop comparing

Stop when one candidate satisfies the hard requirements, has usable evidence for the decision-critical fields and no new row is likely to change the choice without adding information you cannot verify. Also stop when every candidate has the same unresolved critical gap; the next action is to find better evidence, not more of the same rows.

A shortlist is a temporary research note, not a promise that the page will remain unchanged. Recheck the live destination before taking any outside action.