Ticket Data Model Freeze (Onboarding)
A checklist with 7 steps: Inventory every field the client's current support process captures before opening a new platform.
By InnovaAI ResearchPublished
What are the steps?
Ticket Data Model Freeze (Onboarding)
- 01
Inventory every field the client's current support process captures before opening a new platform
Pull 30 closed tickets from the incumbent system and list each attribute agents actually fill in: order number, account tier, product SKU, region, contract ID. Fields that exist in the schema but are never populated are candidates for removal, not migration.
- 02
Separate required fields from nice-to-have fields and cap the required set at eight
Every required field adds seconds to each ticket and increases the chance an agent picks the wrong value. Keep only fields that drive routing, SLA calculation, or reporting the client has asked for in writing.
- 03
Map each channel to a default priority and confirm the mapping with the client in writing
Email, chat, phone, and social rarely deserve the same default urgency. A shared inbox tool like Front or Help Scout will apply whatever default you configure, so the mapping is a policy decision, not a product setting.
- 04
Define the status vocabulary and freeze it before any automation is built
Open, pending, on-hold, resolved, and closed cover most agency support retainers. Adding a sixth status later forces a rebuild of every trigger, view, and report that references the original set.
- 05
Assign an owner to each field and document who may change it after go-live
Without a named owner, fields accumulate one-off values during the first month. Record the owner in the client-facing configuration doc so change requests route to a person rather than a group thread.
- 06
Run 20 representative tickets through the frozen model and log every field that goes unused
Use real historical cases spanning billing disputes, technical faults, and feature requests. If a required field is blank on more than two of the 20, drop it from the required set before launch.
- 07
Version the data model and store the signed-off copy in the client's delivery folder
Date the version and note which platform instance it applies to. When the client asks for a new field in month three, the diff against this baseline is the scope conversation.