Scoring framework · Version 2
How we score.
A weighted editorial model you can inspect, recalculate and disagree with.
What the number means
The score describes suitability for a candidate who wants accessible job discovery, career research and portable information. It is not a fraud audit, an employer endorsement or a measured likelihood of getting hired. The model favours a wider toolkit over a vacancy-only product; a reader seeking human screening or a particular profession may reasonably prefer a lower-scoring board. Inspect the component marks rather than treating the highest total as a universal recommendation.
Criteria and weights
Candidate access and value
Cost, registration friction, clarity about paid features, screening information and whether a candidate can evaluate an opportunity before committing.
Career ecosystem
Useful company information, salary research, preparation resources and application support beyond a list of vacancies. A feature is not assumed effective simply because it exists.
Open data and integrations
Documented APIs, usable feeds, attribution rules and integration guidance. A public endpoint with unclear reuse terms earns less than a documented workflow.
Search and discovery
Location clarity, role categories, salary context and the ability to narrow a search. We assess documented controls, not measured filter accuracy.
Rating anchors and missing evidence
Each dimension is an editorial mark from 0 to 10. Zero means no usable evidence for that dimension; 2 means a limited or unconfirmed offering; 5 means a useful basic offering with material gaps; 8 means a strong documented offering with identifiable limits; 10 means a particularly complete documented workflow within this six-board comparison. Intermediate whole numbers express the editor’s judgement between those anchors. These marks are not laboratory measurements.
We cite the public pages that support the assessment and name important unknowns. An undocumented integration receives little credit; we do not turn that into a categorical claim that no private integration exists. A paid feature can earn capability credit, while its cost affects access and value. A provider’s claim of screening is described as a provider claim, not accepted as proof that every listing is safe. Raw catalogue size, customer testimonials and unverified hiring-success percentages do not produce automatic score bonuses.
Calculation and current component marks
Score = access × 0.35 + ecosystem × 0.30 + integrations × 0.20 + discovery × 0.15. We sum the unrounded weighted values, then round once to one decimal. For example, Jobicy’s current 9, 9, 10 and 8 produce 9.05, displayed as 9.1. The score therefore describes this model’s preferences rather than a universal ranking.
| Board | Candidate access and value (35%) | Career ecosystem (30%) | Open data and integrations (20%) | Search and discovery (15%) | Total |
|---|---|---|---|---|---|
| Jobicy | 9/10 | 9/10 | 10/10 | 8/10 | 9.1/10 |
| Remote OK | 7/10 | 7/10 | 8/10 | 9/10 | 7.5/10 |
| We Work Remotely | 6/10 | 7/10 | 6/10 | 8/10 | 6.6/10 |
| Working Nomads | 8/10 | 5/10 | 6/10 | 8/10 | 6.7/10 |
| FlexJobs | 6/10 | 8/10 | 2/10 | 9/10 | 6.3/10 |
| Remote.co | 6/10 | 6/10 | 2/10 | 8/10 | 5.5/10 |
Each review explains its component marks beside the evidence. The previous overall scores did not expose this calculation. Version 2 rebases them to the visible component model; a lower score on October 17 does not mean that the product suddenly deteriorated that day.
What the review covers
The current source review covers public product pages, official FAQs, pricing, terms and documented feeds. We have not bought subscriptions, completed recruiter conversations or sent job applications for these reviews. We do not claim to know each provider’s duplicate rate, fill rate, screening accuracy or average response time. Public-page access can also differ by region, account state and session; where sources leave a boundary unclear, the review says so.
Review cadence and score history
Our policy is to review the board set at least quarterly, and sooner after a substantiated change to access, pricing, feeds or a feature used in the score. The next scheduled editorial review target is January 2027. This is an editorial target, not an installed automation or a promise of continuous monitoring. Readers can report a change through Contact.
A material revision records the date, previous score, new score and reason on the board’s review. We distinguish corrections to our old description from changes made by a provider. Minor spelling repairs need not imply a new product assessment. Historical scores remain labelled as historical and should not be compared across methodology versions as if they were one consistent time series.
Change the priorities for your search
If you need a screened, part-time role within one country, give access to that inventory and the work-arrangement controls more weight than an API. If you are building a personal feed reader, integration terms matter more. Use the feature comparison and Board Finder to inspect fit, then verify the actual vacancy and current subscription terms. The data methodology explains why the live hiring sample is separate from editorial scoring.