How to Choose Amazon Products Worth Reviewing (and Which to Skip)
Most Amazon affiliate reviews fail before the first sentence is written. Not because the writing is bad, but because the product was a poor choice: too cheap to earn anything, too expensive to convert, out of stock by the time the post ranks, or returned so often the commission never sticks.
Product selection is the part of affiliate publishing nobody wants to do, because it feels like admin. It is also the part with the most leverage. A mediocre review of the right product outperforms a great review of the wrong one.
This is the filter we use. It runs in the order you should actually apply it: cheapest checks first, so you can discard bad candidates before spending real time.
Start with the commission, not the product
Amazon pays different rates in different categories, and the spread is large. Some categories pay a low single-digit percentage. Others pay several times that. The product can take identical effort to review; the payout is not identical.
So the first question is never "is this a good product?" It is "what does Amazon pay for this category right now?"
Check the current commission table in your Associates account before you commit to a niche. The rates change, sometimes without much notice, and a category that made sense two years ago may not today. If you already publish in a low-rate category, the fix is usually not to leave it, but to look for adjacent higher-rate categories your audience would also buy from.
Find the price band that converts
Cheap products and expensive products both underperform, for opposite reasons.
Very cheap items (under roughly $20) pay pennies per sale. You would need hundreds of conversions to earn what a single mid-priced sale returns. Readers also tend not to research cheap purchases, so the search demand for reviews is thin.
Very expensive items (several hundred dollars and up) pay well per sale but convert poorly. Buyers research for weeks, compare across multiple sites, and often buy elsewhere or not at all. They also return more.
The productive band for most niches sits between those extremes: products people will research before buying, but will decide on within a day or two. The exact range depends on your category. If you are not sure, start in the middle and let your own click and conversion data pull you up or down.
One thing that makes this band more forgiving: Amazon credits you for anything the reader buys in the session after clicking your link, not just the product you reviewed. A mid-priced review that brings in a buyer who then adds three other things to the cart is a good outcome. That effect fades as the price goes up and buyers get more deliberate.
Use reviews as a demand signal
A product's review count tells you two things at once: whether people are buying it, and whether there is enough public information for you to write a credible review without owning it.
A product with a few dozen reviews may be excellent, but you have no evidence anyone wants it. A product with thousands of reviews has proven demand and a large body of real buyer feedback to draw on: recurring complaints, the use cases people mention, the alternatives they compared it to.
Rating matters less than you would think, as long as it is not bad. A 4.3 with two thousand reviews is a stronger candidate than a 4.9 with fifteen. The exception is anything below about 4.0: you can still write about it, but it belongs in a comparison as the product to avoid, not as a standalone recommendation.
Check the return rate, because returns cancel commissions
This is the one most people never think about. If a buyer returns the product, Amazon reverses the commission. A category with a high return rate can look profitable in your reporting for a few weeks and then quietly deflate.
Amazon does not publish per-product return rates, but you can read the signal. Look at the one- and two-star reviews. If they cluster around sizing, fit, color accuracy, or "not as described," returns are high. Clothing, shoes, and anything that depends on personal fit or taste are the usual offenders. Products where the complaints are about durability over time are much safer, because the return window has usually closed by then.
Prefer products that will still exist in a year
Content takes months to rank. If the product is discontinued, out of stock, or replaced by a new model before your post reaches page one, you have built a review for a dead link.
Warning signs:
- Model numbers that change yearly, where last year's version is already marked "older model"
- Third-party sellers with low inventory and no Amazon fulfillment
- Seasonal or trend-driven products with a short sales window
- Anything currently trending on social media, which usually means the peak has already passed
The safest choices are established products from brands that update slowly and keep older versions available. Boring is good. Boring ranks for years.
Confirm someone is actually searching
A product can pass every check above and still be worthless to review because nobody types its name into a search engine.
Before you write, look for evidence of search intent:
- Search the product name plus "review" and see what ranks. If the results are all retailer pages and forum threads, there is room. If they are all established review sites with deep content, you will need something they do not have.
- Check whether "best [product type]" queries exist and what already ranks for them. A single review is easier to rank if there is a broader buyer-guide term you can also target.
- Read the "customers also compared" and "frequently bought together" sections on the product page. These tell you which comparison articles are worth writing, and a comparison of two in-demand products often outranks a review of either one alone.
If you cannot find any evidence that people search for the product or its category, skip it, no matter how good it looks on the other checks.
Which products to skip
Some categories fail this filter so consistently that it is worth naming them:
- Anything you would be embarrassed to recommend. Reader trust is the whole asset. Once it is spent on a bad product, your other recommendations stop working too.
- Products with restricted or excluded commissions. Some categories and product types are excluded from the Associates program or pay nothing. Check before you write.
- Consumables under a few dollars. Low price, low margin, and buyers do not research them.
- Gift cards, subscriptions, and digital goods. These typically earn nothing or next to nothing.
- Products with few reviews and no brand recognition. Not enough demand, not enough public information, and too much risk of the listing disappearing.
- Fit-dependent items. Clothing, footwear, jewelry, and anything sized. The return rate eats the commission.
- Items where Amazon is not the natural place to buy. If your reader would obviously buy this at a specialist retailer, they will, and you get nothing.
A five-question filter
If you want this as a checklist, ask these in order and stop at the first "no":
- Does the category pay a commission rate I am willing to work for?
- Is the price in a band where readers research but still decide quickly?
- Are there enough reviews to prove demand and give me real buyer feedback to write from?
- Are returns likely to be low, based on what the negative reviews complain about?
- Will this product still be available, and will people still search for it, a year from now?
Five yeses means it is worth writing. Anything less, and your time is better spent on the next candidate.
Where the tooling helps
Running this filter by hand across a hundred products is slow, which is why most people skip it and pick products by gut. AffSERP's product picker searches Amazon and shows each candidate's price, rating and review count side by side, so you can shortlist before you commit to writing, then generates the review, comparison or buyer guide from the products you keep. The judgement calls above are still yours; the tool removes the copy-paste.
Pick the product carefully, and the article has a chance. Pick it carelessly, and no amount of writing will save it.
Shortlist products, then write from them
Search Amazon, compare price, rating and review count, and generate reviews, comparisons and buyer guides from the products you keep.
Start with free creditsKeep reading
