A star average is a summary of opinions, not a description of how a product will work in your home. Reviews become more useful when you read them with a particular question: will this model perform the task I need, under conditions similar to mine?
Write down two or three uncertainties before opening the reviews. For a desk lamp, they might be clamp opening, switch access and cable reach. This keeps the reading focused. Without a question, it is easy to spend half an hour collecting praise and complaints that have little bearing on your purchase.
Check whether each useful review concerns the same model and variation. Retailers sometimes display comments about different sizes or generations together. A complaint that a container is too small may concern a smaller option than the one in your basket. A photograph can help establish context, but it still does not replace the dimensions in the specification.
Separate product behavior from transaction service. A late parcel tells you something about that delivery experience, while a hinge that repeatedly loosens raises a product question. Neither should be ignored, but they belong in different parts of your comparison. Also notice how long the reviewer used the item. An unboxing reaction offers different evidence from a detailed account of routine use.
Read several sources rather than treating one retailer's rating as the whole picture. Look for specific observations, explained limitations and disclosures of incentives where provided. A verified-purchase badge or a very confident tone should not end your assessment. Review systems can contain misleading material, and an individual account can be sincere while still describing circumstances unlike yours.
Suppose, hypothetically, that several owners mention a short cable. The useful next step is to measure the route from your planned lamp position to the outlet and verify the published cable length. You have converted an opinion into a compatibility check. If the manufacturer does not provide the length, ask for it. Do not infer a measurement from how frustrated the reviewers sound.
Use recurring complaints as investigation prompts, not automatic verdicts. Some differences are matters of preference; others expose a missing requirement. When you finish, record which questions were answered and which remain open. If an unresolved issue could make the item unusable, look for another option or obtain clarification before buying. The goal is a defensible decision, not certainty from a crowd or an endless search for a product nobody has criticized.