How we score

Every StackReview score comes out of the same process. We publish it here so you can judge our verdicts — and disagree with them — on the merits.

How we test

We pay for our own seats

Every tool is bought with our own money on the plan a typical team would actually choose. No press accounts, no vendor-supplied demo environments, no guided tours from a sales engineer.

We run real work through it for two weeks minimum

We spend at least two weeks doing genuine day-to-day work in each tool — real documents, real projects, real teammates — instead of ticking off a feature checklist. Bigger platforms get longer.

We stress the edges

Large datasets, slow connections, offline use, permissions and sharing, mobile apps, keyboard navigation, exports, and migration paths in and out. Most tools look great in a demo; we look for where they strain.

We test the way a real team adopts

Onboarding from zero, inviting non-technical colleagues, wiring up two or three integrations, and living with the defaults before we tune anything. Adoption friction is part of the product.

The five-point rubric

Each tool is rated from 1 to 5 on five criteria. The headline score out of 5 is our considered overall judgement informed by those criteria — not a blind average, because a fatal weakness in one area shouldn't be smoothed over by strength elsewhere.

Ease of use: How quickly a new user gets productive, how predictable the interface is, and how much configuration is needed before the tool earns its keep.
Feature depth: Whether the capabilities go beyond the marketing page — advanced views, automation, permissions, APIs — and whether they hold up under real workloads.
Value for money: What the realistic plan costs per seat per year, what's locked behind upgrades, and whether the free tier is genuinely usable.
Performance & reliability: Speed with real data volumes, sync behaviour, mobile and offline handling, and the vendor's public incident history.
Support & docs: Documentation quality, response times on real support tickets we file ourselves, and the state of the community around the product.

What the numbers mean

4.5–5.0 — Best in class. We'd recommend it to most teams in its category without hesitation.
4.0–4.4 — Very good, with trade-offs worth knowing before you commit.
3.0–3.9 — Solid for the right team, but there are clearly better options for some use cases.
Below 3.0 — Only worth it in specific circumstances we spell out in the review.

When and how we update reviews

SaaS moves fast, so a review is a snapshot. We revisit every published review at least once a year, and sooner when a vendor ships a major release, restructures pricing, or has a significant outage or security incident.

When an update changes a score, we say so in the review rather than quietly editing the number. Pricing and features described in any review are accurate as of that review's stated date — always confirm current pricing with the vendor before you buy.

Readers catch things we miss. If you spot an error, tell us and we'll verify and correct it.

How affiliate relationships are firewalled from editorial

StackReview earns money through affiliate links, and we think you should know exactly how that is kept away from the scoring:

Testing and scoring are completed before any affiliate link is added to a review.
Commission rates are never a factor in scores, rankings, or which tools we cover.
No vendor reviews, approves, or previews our copy, and none can pay for placement.
We publish critical reviews of partners, and we link to competitors regardless of whether they pay us.

Full details on our affiliate disclosure page.

Want to see it applied? Browse every review.