A client's operation continues when something fails on our side. In a marketplace model, "our side" includes partners — so continuity is designed into how campaigns are placed, not bolted on afterwards.
Every campaign above 10 seats is placed with at least two qualified providers wherever the campaign design allows, with a 24-hour resumption target if one provider fails. This is the single most effective continuity measure available to us, and it is why the model exists in the form it does. Clients are told about it at proposal stage.
| Scenario | Response |
|---|---|
| Partner site outage — power, connectivity, premises | Volume shifts to the secondary provider; target resumption 24 hours |
| Partner business failure or termination | Transition per 18_Vendors/05_Performance_Management_And_Termination.md; data returned or destroyed with certification |
| Mass attrition or strike at a site | Secondary provider absorbs; hiring surge initiated |
| Platform or hosting outage | Failover per the hosting architecture; status communicated to affected tenants within the incident SLA |
| Data loss | Restore from backup; RPO and RTO per the client's contract |
| Cyber incident or ransomware | Isolate, contain, restore from clean backup, breach process per 02_Data_Protection_Policy.md |
| Key person unavailability | Documented deputies for every critical role; SOPs mean processes do not live in one person's head |
| Natural disaster or civil disruption at a location | Shift to other network locations; remote working where the campaign's data rules permit |
| Client-side outage | We report it, hold capacity per contract, and do not silently idle staffed seats without telling them |
Stated per client in their contract. Where none is stated, internal defaults apply:
| Service | RTO | RPO |
|---|---|---|
| Hosted platforms | [[RTO_PLATFORM]] | [[RPO_PLATFORM]] |
| Campaign operations (dual-placed) | 24 hours | Not applicable |
| Internal systems | [[RTO_INTERNAL]] | [[RPO_INTERNAL]] |
Automated, encrypted, stored separately from production, with restoration tested on a stated cycle. An untested backup has not been proven to exist.
| Step | Who | Timing |
|---|---|---|
| Internal notification | Delivery Manager → Function Head → Director | Immediately |
| Client notification | Account Manager | Within the incident SLA — before the client notices, wherever possible |
| Status updates | Account Manager | At the P1 cadence, whether or not there is progress |
| Resolution confirmation | Account Manager | On restoration |
| Written incident report | Function Head | Within 5 business days |
We tell the client before they find out. A client who discovers an outage from their own customers has lost more confidence than the outage itself cost.
Continuity arrangements are tested on a stated cycle — a failover exercise, a backup restoration, and a tabletop walkthrough of a partner-failure scenario. Test results and remediation actions are recorded. An untested plan is a document, not a capability.
Reviewed annually and after every material incident.
Registration is free. Browse every live project with its scope, seats and payout, then apply to the ones that fit.
Register free See live projects