Compare before you buy: a five-minute method that beats scrolling reviews

Most online shopping decisions are made backwards. You open a listing, read the top four reviews, notice one angry one, get uncertain, read six more, and buy the thing with the most ratings. Thirty minutes gone, and the decision was made by whichever strangers wrote the most memorable paragraph.

Here is the order that works better, and it takes about five minutes.

Step 1: decide the two things that actually matter

Before you open anything, name the two attributes that would make you return the product. For a breakfast cereal that might be added sugar and fibre. For a sunscreen, SPF and finish. For a t-shirt, fabric weight and fit. Two, not five - a list of five criteria is a list of no criteria, because nothing ever wins on all of them.

If you cannot name the two, you are not ready to shortlist yet. Read one good explainer on the category first - that is what the guides and blog posts here are for.

Step 2: filter, do not browse

Browsing is how you end up with the product with the biggest ad budget. Filtering is how you end up with the product that meets your constraint. Use the category filters to cut by the hard constraints first - brand if you have a preference, price band, pack size, dietary flags - and let the score ordering do the rest.

Aim to get to eight products or fewer. A shortlist you cannot hold in your head is not a shortlist.

Step 3: compare the shortlist side by side

This is the step people skip, and it is the one that does the work. Put three or four products in the compare view and read across the rows rather than down the pages. Differences that are invisible when you read listings one after another are obvious in a single row.

Comparisons on Filtizy are shareable links, which is the underrated part: send it to whoever you are buying with instead of screenshotting four pages.

Step 4: check the real landed price, not the listed price

The same product often sits at four different prices across retailers on the same day, in different pack sizes, with different bank offers attached. The number that matters is the price per unit after the offer you can actually use. There is a whole post on that arithmetic; the short version is: always convert to price per 100 g or per unit before you feel good about a discount.

Step 5: now read reviews, and only for one thing

Reviews are bad at telling you whether a product is good. They are excellent at telling you whether it fails - leaks in transit, sizing runs small, pump breaks in a month, colour differs from the photos. Read them last, and read them only for failure patterns that repeat.

• Sort by most recent, not most helpful - formulations and suppliers change.
• Three people describing the same defect is a pattern; one furious review is a story.
• Ignore reviews about delivery. That is the courier, not the product.
• For apparel, reviews that mention height and usual size are worth ten that say nice fabric.

Worked example: picking a breakfast cereal

The two things that matter: added sugar and fibre, both per 100 g. Everything else - grain type, brand, shape, the athlete on the box - is downstream of those two for this category.

Filter to the category, sort by score, and take the top eight. Drop anything whose first ingredient is sugar or whose fibre is under 5 g per 100 g. That usually leaves three or four, which is a comparison you can read in one screen.

Now compare. You will typically find one option strong on fibre and weak on sugar, one the reverse, and one boring option that is decent at both and cheaper per 100 g than either. The boring option is usually the right answer, and you would never have found it by scrolling - it has no badge, no hero image and no marketing angle.

Total time: about four minutes. Compare that with reading thirty reviews of one product that the algorithm chose for you.

Why this order works

Each step throws away more than it adds, and the cheap filters run first. You spend your attention on four products instead of forty, and the one subjective input - other people opinions - arrives last, when it can only correct the decision rather than drive it.

It also generalises. The same five steps work for a washing machine, a pair of running shoes and a jar of peanut butter - only step one changes, because only the two attributes that matter change.

Save the shortlist when you are done. Prices move, and a saved product with a comparison link is a five-minute decision you never have to make twice.