ToolShedTested
Research protocol

How We Evaluate Tools

Maintained by Morgan Hale · Updated August 27, 2026

Current publication status: ToolShed Tested is research-led. No current guide is presented as hands-on tested. We use evidence labels to state what supports each recommendation, and we do not treat a numerical editorial preference as a measured laboratory result.

The Five-Step Research Protocol

1. Define the buyer and task

Every guide starts with a specific decision: the job, expected frequency, material or capacity requirements, budget tier, and existing battery platform. A product can be a strong tool and still be the wrong recommendation for that buyer.

2. Lock the exact model

We match the manufacturer, model number, tool-only or kit configuration, included battery and charger, voltage platform, and current package. If an exact Amazon product match is not verified, the buying link remains a clearly labeled search rather than a guessed product page.

3. Check primary sources

Manufacturer product pages, manuals, warranty policies, and safety documents are the preferred sources for specifications and coverage terms. Retailer pages provide current package and availability context. Manufacturer performance claims remain attributed and are not rewritten as our own measurements.

4. Compare on one consistent decision matrix

Products in a guide are compared on the same category-relevant factors. Typical factors include capacity, controls, dimensions, compatibility, included equipment, warranty, ownership cost, and task fit. We use ranked recommendations and explicit tradeoffs instead of presenting unsupported decimal scores as scientific measurements.

5. Publish limits and corrections

Every research-backed page states that it is not a physical test. Material factual errors are corrected and recorded. Readers can send a model, specification, source, or warranty correction to hello@toolshedtested.com.

Active Evidence Labels

Physical Testing Gate

ToolShed Tested does not apply hands-on, workshop-tested, or long-term-tested labels by default. A future page may use one only when it provides the exact model, acquisition method, task, materials, test conditions, repetitions where relevant, recorded observations, limitations, and original evidence files or photographs. A product image, owner review, or manufacturer demonstration does not satisfy this gate.

What Owner Feedback Can and Cannot Do

Recurring verified-owner reports may identify patterns worth considering, such as common package confusion or repeated durability complaints. They are contextual evidence, not our own experience and not proof that every unit has the same outcome.

See the Evidence

The Evidence Desk reports current source coverage and links to source-led studies. The editorial policy explains independence and corrections.