Methodology

Scoring without a fake universal winner

Software fit changes with volume, platform, team and the job being automated. Our score reflects that.

Five dimensions

1. Workflow fit: can it perform the required work, including handover?
2. Commerce fit: platform, order data, channels, languages and adjacent integrations.
3. Operating cost: plan, usage, setup, QA and expected overages at the reader’s volume.
4. Implementation load: data preparation, configuration, testing and ongoing ownership.
5. Control and evidence: permissions, reporting, auditability and quality of source evidence.

Weights are versioned by use case. A small Shopify store prioritising speed should not receive the same ranking as a multilingual brand with a dedicated support-operations team.

Evidence penalty

Unknowns lower confidence, not necessarily product quality. A missing public price, unclear limitation or untested integration is shown as uncertainty. It cannot be silently filled with a favourable assumption.

Commercial firewall

Affiliate status is stored separately from product fit. The ranking function does not read commission rate, partnership tier or campaign terms.