How we score merchants in 2026: inside the 340-hour testing standard

A review score is only useful if you understand how it was produced. At SiftPrime we publish a single number next to every merchant we rank — and that number carries real weight for readers deciding where to spend their money.

This article exists so you never have to take that number on faith. Below is the exact process behind it: how long we test, what we measure, how the weights are set, and what happens when a merchant disagrees with us.

Why most review scores are not trustworthy

Many comparison sites rank merchants by who pays the highest affiliate commission. The score is dressed up with stars and percentages, but the ranking was decided in a spreadsheet long before anyone opened an account.

We start from the opposite direction. A merchant cannot appear on SiftPrime until it has passed our intake review, completed a hands-on trial, and been scored against the same fixed rubric we apply to every competitor in its category.

The 340-hour standard, step by step

The 340 hours in our name is not a marketing round number — it is the total reviewer time that goes into building a single category ranking: opening trial accounts, running the same tasks across every competitor, logging evidence, and having a second reviewer verify the result before anything is published.

1. Hands-on trial accounts

Our reviewers open real, paid accounts using their own credentials. We use the product the way a genuine customer would — for a minimum of 30 days and, for merchants with subscriptions, at least two full billing cycles. A free trial alone never qualifies.

2. The weighted scoring model

Each merchant is graded across the same five dimensions we use site-wide, described in full on our review methodology page. The weights are locked in before testing begins, so no result can be quietly nudged toward a desired outcome.

Each criterion is scored from 0 to 10 in half-point increments, then combined into a single overall score out of 10. A merchant scoring 9.0 or above earns our Exceptional verdict and is eligible for an Editor's Choice badge; below 6.0 is Not Recommended. The full verdict bands are on the methodology page linked above.

3. Independent verification

Before a score is final, a second reviewer who did not run the trial re-checks the raw notes, receipts and screenshots. If the evidence does not support a claim, the claim is removed — even when the score drops as a result.

How we handle bias and conflicts of interest

SiftPrime earns revenue through affiliate partnerships, which means we can be paid when a reader signs up with a merchant we recommend. That is exactly why the firewalls below matter.

• Reviewers never see commission rates while testing.
• Editorial has the final say on every score; commercial considerations cannot change a rating.
• Sponsored content is always labelled and never counts as a review.
• Any reviewer with a personal or financial tie to a merchant is removed from that test.

You can read the complete disclosure framework, including how scores are kept separate from affiliate links, on our how we make money and review methodology pages.

What happens after publication

Scores are living documents. We re-check every merchant at least once a quarter, and immediately if pricing, ownership or terms change. When a score moves, we publish a dated note explaining what changed and why. Older versions stay on the record — nothing is quietly edited away.

Frequently asked questions

Related reading