How We Test and Review Amazon Seller Software
Last updated
Every claim on this site should be checkable against the page it came from. This is how we get there — what we verify, how, what our ratings mean, and, just as usefully, what we do not do. If any of it does not match what you find on our pages, that is a fault worth reporting.
Verification means opening the vendor’s page ourselves
When we state a price, a plan tier, or a discount, we have looked at the vendor’s own live page in a browser at the time of writing. Not from memory, not from a competitor’s summary, not from a coupon aggregator. Then we record the date we did it, and put that date next to the claim.
Our SmartScout coupon page is the clearest worked example. The code was confirmed on the live partner checkout on 1 August 2026, and the plan tiers were re-confirmed on 2 August. Both dates appear on the page, because a price claim without a date is a claim you cannot judge. Software pricing moves quietly, and a figure that was right in March is not automatically right now.
We check tables in a real browser, not through a fetcher
This sounds like a technical detail. It costs people money.
We found that automated page-fetching misreads vendor pricing comparison tables — the kind with a column per plan and ticks down the rows. The text comes back, the alignment does not, and a feature ends up attributed to the wrong tier. So anything that depends on which plan a feature sits on gets read in a real browser, off the rendered page.
Tier claims are where buyers lose money — somebody pays for the cheap plan and finds the tool they came for starts one tier up.
We read the vendor’s terms when a claim depends on them
If a claim rests on what a vendor’s terms permit, we read the terms rather than guess.
The stacking question is the one that matters on SmartScout. Several well-ranked pages tell readers to combine the 25% coupon with annual billing and report the total as 42%, 45%, even 50% off. It does not work. SmartScout’s Terms of Use state that discount offers may not be combined with other discounts unless explicitly stated, and their checkout carries a specific rejection state for coupons valid only on monthly plans. So those pages are advertising a price nobody can actually buy. We say 25% and explain why, which is the less impressive number and the true one.
We say so when a vendor contradicts itself
Vendors are not always consistent across their own site, and when they are not we report the contradiction rather than pick the version that suits us.
SmartScout’s affiliate-programme page advertises a “7-day free trial”. Its pricing page offers a 7-day money-back guarantee. Those are different products: one costs you nothing, the other takes your card and refunds you if you ask. There is no free trial, and we say so plainly even though the more marketable claim was sitting there ready to copy.
We publish corrections, and we say when we were wrong
We get things wrong. When we do, we fix the page and leave a note saying what changed, rather than editing quietly and hoping nobody noticed.
Two live examples. We previously described SmartScout’s Share of Voice as a Business-tier feature; it is Enterprise-only, and the page now says so. And our Rank Maker review carries a section headed “A correction to our own earlier version” — we had placed the historical keyword suite on Enterprise when it starts at Essentials. That correction is the difference between booking a sales call and paying $97 at a checkout, which is exactly why it is on the page instead of in a changelog nobody reads.
How our ratings work
Ratings run 0 to 5, in half points. They are editorial judgement, not a computed score. There is no weighted rubric behind them, no percentages assigned to features and support and value. Anyone publishing one has usually reverse-engineered it from the number they already wanted.
What the numbers mean in practice:
| Rating | What it means |
|---|---|
| 4.5 | Best in its lane. We would buy it ourselves for the job it does, and the caveats are about fit rather than quality. |
| 4 | Good, with a real limitation we name — a gap in coverage, a tier that puts the useful part out of reach, or a workflow that fights you. |
| 3.5 | Works, but narrowly. Usually a supporting feature rather than a reason to buy a subscription, or a tool that only pays off in a specific situation we describe. |
Nothing on the site currently sits below 3.5. Read that as the limit of what we cover rather than proof everything is good — we write about tools worth considering, and when something is the wrong buy we say so in the copy instead of dressing it up as a number. Most pages carry no rating at all, because a pricing guide or a feature explainer is not a verdict.
A rating has never moved because of a commission, and it never will.
Affiliate links, and the limit we put on them
We earn a commission on some links. It is disclosed in our disclaimer, and it comes with a hard limit: one affiliate link per article, placed early, and never more. That is not a guideline we try to honour — it is a check that runs against every page and fails if a second link appears.
Pages that bury eight identical links through the body are optimising for clicks, not for the reader. One link, near the top, where you can take it or ignore it.
A commission never changes a verdict, a rating, or a recommendation. The most useful thing we publish is often “do not buy this”, and that only stays useful if we are willing to lose the sale.
What we do not do
The limits matter as much as the method.
- We do not publish original screenshots of every tool’s interface. Some pages have them and some do not. Where a page has none, that is the honest state of the page rather than an oversight we are hiding.
- We do not claim long-running hands-on trials we have not run. No “we tested this for six months” unless we did.
- We do not imply we pay for every tool we hold an account on. Some of our accounts are partner or comped accounts provided because we are an affiliate. That gets us into the product to verify what it does; it does not make us a paying customer, and we will not describe ourselves as one. Our SmartScout account is a comped Business plan, which is how we were able to confirm what Ad Spy shows on a non-Enterprise tier.
- We do not republish vendor marketing statistics as our own findings. Where we cite a vendor’s own number or claim, it is attributed in the sentence — “SmartScout’s own Rank Maker page says”, “their pricing page lists” — so you can see whose measurement it is.
- We do not invent coupon codes. We publish codes we have seen apply at a checkout. A large share of the codes circulating in this niche were never issued by anyone.
How often we re-check, and what triggers an update
Deal and pricing pages are re-checked most often, because they go stale fastest. Reviews get re-checked when something underneath them moves.
An update is triggered when a vendor changes pricing or plan structure, when a code stops applying at checkout, when a feature moves between tiers, when a reader tells us something is wrong, or when we find our own error. The updated date on every page reflects the last time it was checked, not the last time a comma changed.
We do not run an automated price monitor. Verification is done by hand, which is why the date stamp is the honest signal: if it is old, treat the number as old.
Tell us we got something wrong
If a price has moved, a code has stopped working, or a tier claim does not match what you see at checkout, tell us. Corrections get made on the page, and where the error was material we say what changed rather than removing the evidence.
More about who we are and why we cover this category is on our about page.