Answer an RFP support model question with a workable rota

Scope

Start with the people and work the service must support

Westhaven has ticket-office staff, membership administrators, event supervisors and visitors. They do not all need the same assistance or have the same authority. A visitor may need help retrieving a booking; only an authorized employee may request an administrative change. Identify the supported products, environments, sites and user groups before assigning channels. Include onboarding and transition support only where the offer funds them, and distinguish them from routine live-service assistance.

Build a short catalogue of contact types using the buyer’s work. A request for instructions, an interrupted sale, a suspected security event, a product enhancement and a complaint should not disappear into one undifferentiated queue. They can share an intake route, but their next decision and responsible team differ. ISO/IEC 20000-1 defines requirements for a service management system, while Part 2 offers application guidance. Neither supplies a ready-made Westhaven rota or proves that the bidder is certified.

Record exclusions as routes, not dead ends. If the bidder does not control the museum’s local network, explain who collects the first observations, who contacts the buyer’s network team and who tells the museum what happens next. An excluded technical component does not automatically remove every coordination duty. The tender and approved contract determine which duties remain with the bidder.

Coverage

A channel can be open while nobody is working the case

For each channel, separate receiving a message, acknowledging receipt, performing triage, starting specialist work and communicating progress. An automated ticket receipt proves only that the system recorded the contact. If live assistance ends at 18:00 but critical incidents can reach an on-call responder, say so. Explain which incident classes qualify, how the customer invokes that route and what happens to ordinary requests submitted at the same time.

Use a named time zone and a calendar that accounts for daylight-saving changes, public holidays and buyer-specific operating days. “Business hours” is incomplete when an English museum, a continental support team and an overseas supplier follow different calendars. Show whether a case already in progress continues after the desk closes and who decides. Keep the timing calculation aligned with the service-level schedule rather than introducing a second definition in the narrative.

Accessible contact is part of usable coverage. A person locked out of the product may also be locked out of a support portal that uses the same login. Provide an approved alternative and explain its verification limits. GOV.UK’s assisted digital guidance distinguishes helping someone use a digital service from assuming every user can complete it unaided. W3C’s Consistent Help criterion concerns the relative placement of repeated help mechanisms; it does not require a human to be available continuously.

Capacity

Show that the rota survives ordinary absence and competing work

A headcount is not a coverage plan. Identify the skills required at each point: customer communication, product diagnosis, data correction, supplier coordination and incident leadership may sit with different people. Check whether the proposed workers have the permissions and current knowledge to act. A name on an escalation list is weak evidence if the person cannot access the relevant tools or has no allocated support time.

Use contact demand by interval and channel, handling time, follow-up work and the mix of specialist cases to test staffing. GOV.UK’s support guidance provides a useful planning basis for estimating demand and reviewing performance; its examples are not targets to copy into an offer. Allow for leave, sickness, training, breaks and handover. Shared teams also need a rule for simultaneous incidents across customers. Do not allocate their full capacity separately to every contract.

For a simple illustration, a continuously staffed position requires 168 staffed hours per week. If one full-time equivalent contributes an assumed 30 hours of usable desk coverage after other duties and absences, the arithmetic gives 5.6 equivalents for that position alone. This is a planning lower bound, not a recommendation to hire six people: skills, shift distribution, local employment rules, peak concurrency and resilience may require a different plan. An on-call arrangement is a different service and must not be counted as an occupied desk.

Keep personal rota details and emergency contact information out of the public bid narrative unless an approved tender requirement calls for controlled disclosure. Give the evaluator roles, capacity basis, backup arrangements and evidence of readiness. The service owner should approve the allocation and the commercial owner should confirm that the price funds it.

Case ownership

The customer should not have to manage your resolver chain

Assign a customer-facing owner at triage and keep that responsibility explicit until closure or an accepted transfer. Technical work can move from the desk to a product specialist, then to the payment supplier, while the original owner continues updates and tracks the next decision. Sending an email to a supplier is a requested handoff. It becomes an accepted handoff only when the receiving party confirms the case, responsible function and next action.

Prepare a transfer packet containing the affected user outcome, scope, timestamps and zone, observed symptoms, checks performed, safe evidence, current workaround, priority basis and requested decision. Do not copy secrets or unrestricted customer records into another queue. Record the receiving case reference and any limitations on what can be shared. A supplier contract that allows only a designated administrator to raise incidents creates a dependency that the model must staff.

At Westhaven, a payment supplier may restore its interface while an admissions reconciliation remains unfinished. The support owner must distinguish supplier restoration, recovered customer operation and remaining corrective work. Where the bidder cannot control the supplier’s timetable, state the coordination promise it can make and the limitation it cannot remove. Do not turn the supplier’s published support description into a guarantee for the entire offered service.

Escalation

Define the action an escalation is supposed to unlock

Base priority on the buyer’s affected work, breadth of impact, urgency, available alternative and relevant risk. A single administrator unable to process a time-critical event may have more immediate impact than many users seeing a cosmetic defect. Record what is known and what is still being established. Let a qualified role review priority when facts change, but retain the original observations and any effect on the agreed clock. A label change must not quietly erase elapsed time.

Functional escalation brings another skill or permission. Management escalation resolves a blocked allocation, unaccepted handoff or commercial decision. A major incident needs a coordinated response and an owner for customer communication. Google’s incident-response account separates incident command, operations and communications. Use that as a role-design reference rather than a claim that every support case needs three new people.

Add explicit triggers: no accepted owner, repeated failed recovery, worsening impact, missed update, unavailable specialist or a supplier refusing the case. For each, name the authorized decision and fallback if the first escalation is unanswered. Suspected compromise moves to the approved cybersecurity incident process; NIST SP 800-61 Revision 3 addresses that risk-management context. A general support article should not prescribe an improvised forensic investigation or declare a legal notification deadline.

Closure and reporting

A smaller queue does not prove customers are getting help

Specify what the owner checks before resolving and closing a case. Confirm the relevant customer operation, describe any workaround and disclose remaining limitations. Keep a permanent correction or recurring problem linked to its own work record when restoration is temporary. If the customer does not respond, use the agreed reminder and closure process with a reopening route; do not equate silence with acceptance of every conclusion.

Complaints deserve a distinct owner and review, even if the associated software fault is fixed. ISO 10002 offers complaints-handling guidance rather than a substitute for the contract or a particular dispute mechanism. Keep the complaint’s outcome visible without manipulating incident priority to improve a dashboard. Similarly, a cybersecurity investigation can continue after a service is available again.

Report waiting time and unresolved age alongside first response, restored outcomes, repeated contacts, reopened cases and missed updates. Segment by request type, channel, coverage window and dependency so a fast routine queue does not conceal stranded urgent work. Google’s postmortem guidance provides a basis for learning from incidents and tracking improvements; the bidder must still assign each action and test whether it changed the service. Avoid rewarding ticket closure alone when several tickets concern the same unresolved museum event.

Bid answer

Walk the Friday case through the model before you release it

At 17:55 on Friday, several Westhaven tills stop completing admissions. The desk closes five minutes later. Ask the proposed team to identify the receiving role, collect enough impact information, invoke the correct route, obtain resolver acceptance and arrange the next customer update. Add a supplier that has not answered and an on-call specialist who is already handling another customer. The exercise ends only when the unresolved work has a valid owner and a funded next action.

Run a second case through the fallback channel while the support portal is unavailable. Test whether a user can find it, communicate accessibly and report the issue without sharing credentials. GOV.UK’s contact pattern is useful for making route and availability information clear. In a bid rehearsal, use approved test contacts and synthetic records. Do not create a false live emergency or disturb another customer’s support queue.

The answer can then describe the supported users and services, available routes, actual hours, case ownership, resolver and supplier handoffs, escalation authority, customer updates and closure. Attach or reference the coverage schedule in the permitted location. Label proposed coverage that still depends on recruitment, supplier agreement or buyer approval, and give its readiness condition. Do not describe it as current.

Reconcile the schedule with the service-level definition, price assumptions and transition plan. The support narrative owns the operating route; it does not create new response promises, service credits or recovery objectives by implication. Keep a version and a review trigger for changes in buyer hours, user volumes, channels, staffing, supplier entitlement or service scope. A polished answer is ready only when the offered people and arrangements can follow it.