01 / Start with a real decision

We define the person, situation and constraint before choosing products. The best shelter for a sheltered weekend campsite may be a poor choice for an exposed route. The same distinction applies to a desk chair, a training shoe or a sleep tracker.

  • Describe the intended use and the conditions that matter.
  • Identify the tradeoffs a buyer can reasonably accept.
  • Set the comparison criteria before reviewing commissions or choosing a winner.

02 / Name the kind of evidence

Research only means we have examined sources without completing a personal-use test. Hands-on means a documented test exists. Mixed evidence means a page combines direct observations with sourced information. Each claim should make its basis clear.

  • Manufacturer claims stay attributed to the manufacturer.
  • Independent findings retain their source and testing conditions.
  • Personal observations are tied to a real product, method and record.
  • Unknowns stay unknown. A missing result is not a passing result.

03 / Make the limitation useful

We explain what was not tested, where results may not transfer and which details could change the decision. One sample cannot establish the reliability of every unit. A short test cannot demonstrate years of durability.

For nutrition, health monitoring and sleep, consumer observations remain separate from clinical conclusions. We do not turn a wearable estimate or a personal experience into a diagnosis or treatment claim.

04 / Let the evidence decide

Personal-use judgments and research assessments stay distinct. Research-supported choice means applicable external evidence supports the stated use; it does not claim that Righteous Recon personally tested the product. Insufficient evidence is a legitimate outcome.

  • Stays: the documented evidence supports the stated use.
  • Stays with conditions: the recommendation depends on explicit limits.
  • Doesn’t stay: the evidence does not support keeping it for that use.
  • Insufficient evidence: the record does not yet support a conclusion.
  • Research-supported choice: a scoped assessment of external evidence, with conditions and limitations.
  • Not recommended for this use: a scoped external-evidence judgment, not a claim of personal rejection.

Budget, Mid, Premium. Three places each.

Each category has Budget, Mid and Premium tiers, with 1st, 2nd and 3rd place within each tier. Every shortlist applies to a stated use and comparable type of product. Different functions, sizes, fit requirements or compatibility can change which models belong together.

An awarded place needs a dated price for the stated market and configuration, supporting product evidence and an explanation of why it outranks an eligible alternative. Price bands belong to the category; Premium is not automatically better.

We challenge the ordering against uncertainty, price changes and relevant user priorities. When the order is unsupported, the place remains unassigned. A useful product assessment alone does not establish a first-place finish.

One exact model keeps one research record even when it serves several disciplines. Corrections follow the shared record.

Trace the claim to its origin

A manufacturer manual can establish what the manufacturer specifies. An original comparative test may support a performance difference under its conditions. A registry can establish a particular certification; a page describing a standard cannot certify a product. The source must fit the claim.

We check exact model and revision, methods, test conditions, uncertainty, funding and conflicts, and corrections or contrary findings. Several articles repeating the same experiment are one originating evidence stream. Source counts and agreement between AI tools are not substitutes for independent evidence.

This is an editorial application of source-appraisal principles. It is not a clinical systematic review, a validated universal trust score or a promise of absolute certainty.

Spend the effort where it changes the decision

We establish category criteria once, screen models together and preserve source extractions for reuse. Deeper research goes to unresolved claims that could change suitability or the ordering. Routine specification work does not need a new essay for every model.

Before closing an investigation, we record what was searched, check decisive extractions and seek counterevidence. If the remaining question requires unavailable testing or access, we publish only the conclusion the evidence supports—or leave the recommendation open.

Price evidence has a maximum 30-day review window in the current editorial policy, with earlier review for changes. That window does not guarantee prices stay fixed. An expiring comparison needs a new review and successful deployment to update a static page.

The verdict can change

A product revision, a better comparison, a recall or a factual correction may change our conclusion. Significant changes belong in the audit’s update record. Refreshing the layout does not count as new testing.

Product and business-opportunity judgments stay separate in MLM Edition. A useful product does not establish that its associated business is a good decision.

End of record