BuyUp

How We Review Products

We believe you deserve to know exactly how our reviews are produced. This page explains our data sources, analysis process, scoring methodology, and the limitations of our approach.

Our Data Sources

Every review we publish draws on several sources of information:

  • Manufacturer specifications — the official specs sheet for each product, including dimensions, weight, battery life, connectivity, and technical details
  • Amazon owner reviews — we analyse ratings and review trends from verified purchasers to understand real-world performance over time
  • Competitor comparison — each product is evaluated against its direct rivals in the same category and price bracket
  • Feature analysis — we examine the feature set and determine how it stacks up against what buyers actually need at each price point

AI-Assisted Analysis

Our reviews are produced using AI-assisted analysis. We use large language models (currently Claude by Anthropic) to synthesise product data, owner feedback, and competitive positioning into structured, readable reviews.

We're transparent about this because we think it matters. AI analysis lets us cover more products more consistently than a small team could manage manually, but it has trade-offs. The AI hasn't held the product, felt its build quality, or used it for a week. What it can do is process large amounts of owner feedback and specification data to identify patterns, strengths, and weaknesses that might take hours of manual research to uncover.

Scoring Methodology

Each product receives an overall score out of 10, broken down across four dimensions:

  • Performance (0–10) — how well the product does its primary job, based on specs and owner feedback
  • Value (0–10) — price relative to what you get, compared against competitors at similar and different price points
  • Design (0–10) — build quality, aesthetics, ergonomics, and portability where relevant
  • Features (0–10) — the breadth and usefulness of the feature set for the target audience

Scores are calibrated honestly. A 7 is genuinely good, 8 or above is excellent, and anything below 6 means we've identified significant shortcomings. We don't cluster scores at 8–9 to make everything look appealing.

What We Don't Do

In the interest of honesty, here's what our reviews don't include:

  • We haven't physically tested these products in a lab or at home
  • We don't have long-term durability data beyond what owners report
  • We can't verify every manufacturer claim independently
  • Our price data may not reflect real-time changes or regional variations

If you need hands-on testing, publications like Which?, Wirecutter, or TechRadar run physical test labs. Our strength is in thorough data analysis and clear, structured comparisons that save you research time.

How We Make Money

BuyUp earns revenue through Amazon affiliate links. When you click through to Amazon and make a purchase, we receive a small commission at no extra cost to you. This commission does not influence our scores, rankings, or editorial content. Products that don't meet our standards receive low scores regardless of affiliate potential.

Updates and Corrections

Reviews are regenerated periodically to reflect new pricing data, updated owner feedback, and any changes from manufacturers. If you spot an error or have feedback about a specific review, please get in touch.