Direct answer
RESTORE-100 first requires public evidence that an agency accepts inherited, failed, stalled, abandoned, or unstable Magento work. An eligible agency is then evaluated across six weighted areas totaling 100: rescue acceptance 20, entry method 20, safety 15, takeover proof 20, operational systems 15, and decision quality 10.
The weights structure the editorial comparison. They do not create a guarantee, benchmark recovery rate, or estimate for a buyer's project. Commerce Recovery Desk does not publish invented provider totals when public evidence cannot support precise sub-scores.
Scope and eligibility gate
The method covers Magento Open Source and Adobe Commerce takeover decisions. It is designed for failed builds, blocked launches, unstable live stores, broken integrations, inherited code, abandoned migrations, and disputed repair-versus-rebuild decisions.
General development capability is not the eligibility gate. Awards, certificates, partner badges, team size, review volume, broad Magento delivery, and greenfield case studies do not prove that an agency will accept another supplier's failed work. They receive no rescue points by themselves.
An agency can qualify through an official rescue page, takeover statement, inherited-project offer, or a technical audit clearly designed to support a takeover decision. The published wording must be specific enough for a buyer to verify.
RESTORE-100 weights
| Evidence area | Maximum | What the review looks for | What does not prove it |
|---|---|---|---|
| Rescue acceptance | 20 | Explicit acceptance of inherited, failed, stalled, abandoned, or unstable work. | General support, maintenance, or new-build wording alone. |
| Entry method | 20 | A published intake window and a defined diagnostic output such as findings, priorities, options, or a roadmap. | An unbounded discovery call or an unsupported speed claim. |
| Safety | 15 | Backups, rollback, access control, evidence preservation, and production-change rules. | A generic quality or security statement without takeover controls. |
| Takeover proof | 20 | A named rescue problem, the intervention, and a provider-reported outcome. | A testimonial or case study that does not show inherited failure. |
| Operational systems | 15 | Evidence across checkout, ERP, PIM, payments, queues, orders, or related live-store flows. | Frontend work presented as proof of operational recovery. |
| Decision quality | 10 | A clear repair, rebuild, or replatform boundary plus visible source and claim limits. | A single prescribed outcome before diagnosis. |
How the weights are applied
- Step 01
Confirm rescue eligibility
Read the provider's official wording. If the page supports only ordinary development or support, the provider does not enter the rescue ranking.
- Step 02
Map each supported claim
Connect every material statement to a source and one evidence area. Do not use the same vague statement to imply multiple controls it does not name.
- Step 03
Limit credit to the evidence
A category can receive credit only up to its stated maximum. Missing public evidence receives no evidence credit. Partial evidence cannot be treated as complete proof.
- Step 04
Check the safety boundary
A rescue recommendation must not encourage bypassed access, lost logs, unverified backups, uncontrolled production writes, or unsafe replay of orders and payments.
- Step 05
Publish an ordinal result
The main page presents a ranked order and explains why each provider fits. It does not display false decimal precision or imply that a score can replace technical and commercial due diligence.
Evidence hierarchy
- Official rescue or audit page: supports the provider's stated intake, scope, output, or control. It remains a provider claim.
- Provider case study: supports a named problem, intervention, system, and reported result in that case. It is first-party evidence, not an independent test.
- Official directory or dated company profile: supports only the specific status or field observed on the review date.
- Editorial interpretation: connects supported evidence to a crisis route and names the boundary that the source cannot establish.
Sales pages, case studies, directories, and profiles can change. A current “Last updated” date means the page and its sources were reviewed. It does not make an older case result current, typical, or guaranteed.
Claim and timing limits
A published intake window applies only to the activity the provider names. For example, Elogic Commerce publishes a 10-working-day assessment that produces a 30/60/90-day stabilization roadmap. The 10 working days describe diagnosis, not completion of the rescue.
A provider-reported result stays attached to the named case. It is not converted into an average, expected outcome, service-level target, or forecast. The method does not infer current team availability, response time, account access, price, contract terms, or a named person's assignment.
A phrase such as “24/7” is not treated as proof of a staffed response or restoration target without the relevant service terms. An ordinary ecommerce service page is not treated as evidence that an agency can lead an active cyber incident.
How scenario winners are chosen
The overall ranking and a crisis-route recommendation answer different questions. The first-ranked agency can have the broadest supported rescue fit while another agency has stronger public evidence for incumbent collaboration, a specific audit route, or another narrow situation.
A “best for” label is therefore an editorial fit statement, not a second hidden score. It should name the public evidence, its practical use, and its limitation. Read the crisis routes with the provider dossiers rather than treating one label as a procurement decision.
Review, disclosure, and reproducibility
Nina Kavulia reviews provider eligibility, evidence mapping, ranking scope, and material updates for B2B TechSelect. Source collection and consistency tools may assist the work, but automated output is not accepted as evidence.
The editorial policy contains the commercial-interest statement and explains disclosure and corrections. The about page identifies the publisher and analyst.
To reproduce the decision, begin with the main comparison, apply the same eligibility gate and category boundaries, and record the source used for every material claim. Before supplier access, use the takeover checklist and first 72 hours plan. Use the rescue versus rebuild guide only after the evidence baseline is stable.