Most tool roundups are commission-funded lists with adjectives. The tools section collects everything that has cleared this bar. This page describes what happens before a tool gets written about here, so you can check whether I followed my own process. It applies to software only. Anything touching tax, visas or foreign account reporting is held to the sourcing rules in the editorial policy instead, because a wrong figure there costs a reader far more than a wrong app recommendation.
The minimum bar in this review methodology
A tool doesn’t get reviewed until it has been used on real client work for at least 2 weeks. Not a trial account poked at for an afternoon. Actual work, with actual deadlines, where the tool failing would cost me something.
That’s why there aren’t more reviews here. The bar is slow on purpose.
What gets scored
Six criteria, applied the same way each time.
- Real cost to a solo operator. The price you pay alone, on the plan that actually has the feature you came for, including the annual-versus-monthly gap
- Time to first useful result. Measured from signup, in minutes
- What it does when it fails. Offline behaviour, sync conflicts, and whether an error loses your work
- Data portability. Whether you can export everything, in what format, and how much you lose on the way out
- Solo fit. Whether the pricing and design assume a team you don’t have
- Works from anywhere. Whether it breaks on an Indian phone number, a non-US card, or an address outside the States
- Support reality. Response time on an actual ticket I opened, not the number on the marketing page
Where pricing figures come from
The vendor’s own pricing page, read on the date the post is written, with that date stated in the post. Regulated figures follow the same rule against primary sources such as the IRS guidance on nonresident taxation. Never another blog’s summary, and never a figure carried over from an older article of mine.
Software pricing moves fast and quietly. If you find a price here that no longer matches the vendor’s page, that’s a correction I want, and the the correction process covers how those get handled.
Every review names what it rejected
A recommendation means nothing without the alternatives it beat. So each roundup names the tools that lost and says why, in specific terms.
You can see this working in the accounting software comparison, where the popular pick loses on a concrete limitation, and again in the security tools breakdown, where several paid products lose to free ones. When I let an AI agent run on my own machine for a week, the write-up leads with what went wrong, because that was the useful part.
Affiliate links and rankings
No tool on this site pays a commission. Nothing here is an affiliate link, so nothing in a ranking has a revenue reason to be there. The FTC endorsement guides set out what has to be disclosed if that changes, and it would be, before publication rather than after.
If a commissioned link ever appears, the post carrying it will say so at the top, in the plain language the FTC asks for rather than buried in a footer. The full position sits in the advertising disclosure.
Screenshots
Where a category has a credible public baseline, the review is measured against it rather than against my own preferences. Security tools get checked against the CISA list of free cybersecurity services and tools, which is why so many paid products lose that comparison.
Screenshots on this site come from real sessions on my own account, with the interface in whatever state it was actually in. No vendor press kits presented as my own testing, and no mockups.
Where a tool sits behind a paid tier I don’t hold, I say so and use the vendor’s published material with attribution instead of pretending otherwise.
When a verdict changes
Tools get worse. Pricing changes, features move behind higher tiers, and companies get acquired. When a recommendation stops holding, the post gets updated and the old verdict stays visible with a note explaining what changed.
Reviews get re-checked at least once a year. The ones covering fast-moving AI tooling get looked at more often than that, because a six-month-old verdict in that category is close to useless.
Why this review methodology is stricter than it sounds. The tool review process has no revenue input at any stage, because no tool on this site pays a commission. That removes the pressure that bends most rankings before the testing even starts.
How tools are tested, in one sentence: used on real work first, ranked second, and dropped when they stop earning the place. This review methodology is published so you can hold the tool review process to it, and the advertising disclosure shows there is nothing behind it pulling the other way.