Skip to the indexed answers
Digital Marketing Index

usdigitalmarketingindex.com SEO

How can a publisher write a fair product comparison without hands-on access to every product?

By saying exactly how each product was assessed, keeping tested findings apart from researched ones, comparing every product on the same criteria, and marking what was not tested as not tested. A comparison built from specifications, documentation, and other people’s published tests can be useful and fair. It becomes unfair the moment it implies experience the writer did not have, or scores products on qualities nobody measured. The reader is owed one thing above all: the ability to tell, for every statement, whether it comes from use or from reading.

Declaring review methodology

Put the method near the top, where a reader sees it before the verdicts.

State:

  • Which products were used in person, for how long, and for what tasks
  • Which were assessed from manufacturer specifications and documentation
  • Which third-party tests or datasets were relied on, by name
  • How products were chosen for inclusion, and whether any maker supplied a product or pays a commission
  • The date of the research

Google’s guidance for review content asks reviewers to “Provide evidence such as visuals, audio, or other links of your own experience” (Google’s guidance on writing high quality reviews). Where that experience exists, show it. Where it does not, the honest equivalent is to say so and to show where the information came from.

Separating firsthand and desk research

Label every claim by its origin. Three labels are enough.

Label Meaning Example wording
Tested The writer used the product and observed this In our use over two weeks, the battery lasted a full workday
Manufacturer-stated From the maker's published specification The manufacturer rates the battery at ten hours
Reported From a named third party's published test A named lab's test measured eight hours

Never convert the second or third into the first. Writing the battery lasts ten hours, with no attribution, turns a claim by the maker into an observation by the writer.

Keep verdict language for what was tested. Feels sturdy, is comfortable to hold, and runs quietly are sensory judgments. They cannot be made from a specification sheet.

Desk research can start from the manufacturer’s own description, a text that every retailer of the product also holds. Making a store’s product page say more than that text is covered in the guide to how retailers can differentiate manufacturer-supplied product descriptions.

Comparable criteria

Fix the criteria before filling in any product, and apply each one to every product.

  1. Choose criteria that matter for the decision the reader is making.
  2. Define each so it can be answered from evidence: a measurement, a yes or no, a documented feature.
  3. Use the same unit and the same source type across products where possible. A lab figure for one product beside a marketing figure for another is not a comparison.
  4. Where sources differ, show the source in the cell.

The same guidance asks reviews to “Discuss the benefits and drawbacks of something, based on your own original research”. Drawbacks belong in the matrix for every product, including the one that pays the highest commission.

A sample matrix follows. The products and values are invented, and the layout is the point.

Criterion Product A Product B Product C
Assessment basis Tested, two weeks Not tested; specifications and one named lab report Not tested; specifications only
Weight 1.2 kg, weighed 1.4 kg, manufacturer-stated 1.1 kg, manufacturer-stated
Battery life 9 hours, tested 8 hours, reported by the lab 10 hours, manufacturer-stated
Noise in use Quiet, tested Not tested Not tested
Warranty 2 years, manufacturer-stated 3 years, manufacturer-stated 1 year, manufacturer-stated
Known drawback Slow charging, tested Heavier than rivals, per specification Short warranty, per specification

Handling unavailable evidence

Gaps are normal. Handle them openly.

  • Write not tested in the cell. Do not leave it blank, and do not fill it with a guess.
  • Do not score what was not measured. An overall rating out of ten that blends tested and untested products gives false precision. Either rate only on criteria with comparable evidence, or give no numeric score.
  • Do not name a winner on a criterion where the winner was not tested and the others were. Say that the specification suggests an advantage that has not been confirmed.
  • Where the maker’s figure and a third party’s figure disagree, show both.
  • If a product could not be assessed well enough to compare, say so and list it separately, or leave it out and say why.

The decision rule for the final recommendation:

Evidence for the recommended product Wording allowed
Tested by the writer We recommend, based on our testing
Not tested, strong documented evidence On paper the strongest for a stated need; we have not tested it
Not tested, thin evidence No recommendation; list its specifications

Revisit the comparison on a schedule. Specifications, prices, and model lines change, and a row that was right at publication can mislead a year later.

This guide covers how to write the comparison once a comparison is the right format. Deciding whether a query calls for a comparison or a single product page is covered in the guide to why some product queries favor comparison articles over product pages.

The standard: a reader should be able to point to any cell and know whether a person measured it, a maker claimed it, or a third party reported it.

Link trail

Reader comments on this answer

Add your comment to this answer

Your email address will not be published. Required fields are marked *