Foundation / 06

Decide what belongs in a hosted system

Map applications and responsibilities before moving work, with practical checks for access, exports, recovery and a return path.

Cloud computing moves some part of an application's operation into infrastructure managed elsewhere. For a small business, the practical decision is which responsibilities move with it and which remain in the office. Begin with the work the application supports, then examine access, data handling, connectivity and the ability to leave the arrangement.

Compact storage equipment arranged on a desk
Compact storage drives, blank papers and a cable on a desk.

Map the current work before choosing a destination

List the people and systems that interact with the application. Include imported spreadsheets, shared folders, printers, scheduled exports and any specialist devices. Ask what happens at the edges of the main workflow: the monthly report, the occasional data correction or the year-end export can reveal dependencies that a routine demonstration misses.

Write down the current owner of the information and the person who administers access. Identify work stored outside the main application, such as attachments downloaded to a laptop. A migration plan should account for those working copies rather than treating the central database as the entire process.

Divide responsibility explicitly

A hosted service does not transfer every security task to its operator. The division depends on the service and its configuration, as the NCSC shared-responsibility guidance explains. The customer may still need to manage users, settings, devices and information even when underlying infrastructure is managed elsewhere.

Build a responsibility sheet for the proposed arrangement. Avoid a single row labeled “security”; split it into tasks that can be assigned and checked. Ask for the same clarity around backup, service administration and the steps required when an employee leaves.

TaskQuestion for the responsibility sheet
Account accessWho approves users and removes access?
ConfigurationWho reviews sharing and administrative settings?
RecoveryWho can retrieve an earlier version or restore lost data?
Service changesWho reviews notices and tests affected workflows?
LeavingWho exports information and checks that it is usable?

Weigh access against dependencies

Remote access may make a hosted application convenient for a distributed team, while also making internet availability and account recovery more central to daily work. Ask what remains available when the connection fails. If an offline mode exists, determine which tasks it supports and how later changes are reconciled.

Review how information is organized and exported. An export may contain records without preserving the screens, relationships or document attachments used in the application. Request a sample and inspect it with the person who will need the information. Also ask who controls the business account if the usual administrator becomes unavailable.

Pilot a complete business task

Choose a pilot with a beginning, a handoff and a finished result. For example, move a sample document through creation, review, approval and retrieval. Use test information where practical and agree who can see it. Record the places where people need new permissions, different file names or a revised procedure.

  1. Define the work and the evidence that will show it completed correctly.
  2. Check sign-in, permissions, linked devices and required imports.
  3. Test a realistic handoff between two roles.
  4. Attempt an export and a recovery task, not just ordinary editing.
  5. Record exceptions and resolve them before expanding the move.

Prepare the return path and final handover

Decide what would cause the move to pause or reverse. Keep the previous working arrangement usable for the agreed transition period, with clear rules about where authoritative information is being edited. Otherwise, two systems can acquire conflicting changes while people believe they are using the same records.

At handover, update the application inventory, access procedures and recovery instructions. The backup guide helps frame recovery questions; the network guide helps identify connection dependencies. If outside support is involved, include the new responsibilities in the written support scope.