A Governed Service Portal with Operational BI for a Retailer
Requests arrived through many channels with no single record, and nobody could measure support. We stood up a helpdesk fast, then redesigned it in place, adding roles, SLAs, documents and dashboards for staff and customers.
- Service
- Business Transformation
- Industry
- Retail & E-commerce
- Region
- Germany
- Engagement
- Lab-type ODC (dedicated team)
Outcomes
- One record per request
- SLA-driven routing
- Self-service case visibility
Stack
Odoo Helpdesk · Python · Metabase · ETL · PostgreSQL
Written byOriental Nippon Team
A German retailer's support team serves two very different groups from the same desk: internal colleagues and external trade customers. It needed a working helpdesk quickly, and it could not wait for a full process design first.
Later it needed much more: roles and permissions for both groups on the same ticket, controlled document sharing, workflows matched to ticket type, and a view of support performance that had never existed.
The problem
- Many channels, no single record. Requests arrived by several routes.
- One ticket, two audiences. Internal and external users needed different views, with no permission model.
- Attachments by email. Sensitive files had no controlled sharing.
- Uniform handling. An urgent trade issue took the same path as a routine internal question.
- No visibility. Neither managers nor customers could see how support was performing.
| Before | After | |
|---|---|---|
| Ticket record | Scattered | One helpdesk, live in weeks |
| Access | Same view for everyone | Roles for staff, trade users, admins |
| Routing | Uniform | Rules and SLAs by ticket type |
| Reporting | None | Internal and customer dashboards |
What we built
A fast-start helpdesk
Standard Odoo Helpdesk, live before any redesign.
A role and permission model
Internal staff, external trade users and administrators on the same ticket, each seeing what they should.
Controlled document handling
Files attached and shared from the case record.
Ticket workflows
Routing, escalation and SLA rules by ticket type.
A BI layer
Pipelines into Metabase, with internal dashboards and self-service views for trade customers.
Tickets now follow rules set by their type, with escalation built in:
Technology and the role it played
| Technology | Role in this project |
|---|---|
| Odoo Helpdesk | Tickets, roles, SLAs and documents |
| Python ETL | Pipelines from ticket and operational data |
| PostgreSQL | Warehouse layer |
| Metabase | Internal and customer-facing dashboards |
How we delivered it
We deliberately did it in the "wrong" order: a plain helpdesk first, then a period of watching how the team really worked, then the redesign.
- Fast start. Standard helpdesk configured and in use.
- Observe. How support actually works, with the team.
- Redesign in place. Roles, permissions, documents and workflow rules, added incrementally.
- Instrument. Pipelines and a warehouse over ticket and operational data.
- Publish. Internal dashboards first, then customer-facing views once the numbers were trusted.
01Fast start
2–3 wks
Standard helpdesk configured and in use
02Observation and mapping
2–3 wks
How support actually works, with the team
03Roles, permissions and documents
3–4 wks
Access model, file handling, sharing controls
04Workflow automation
3–4 wks
Routing, escalation, SLA rules by ticket type
05BI layer
4–6 wks
Pipelines, warehouse layer, internal dashboards
06Client-facing dashboards
2–3 wks
Self-service views for trade customers
Risks and how we handled them
| Risk | How we handled it |
|---|---|
| Designing for a process nobody follows | A plain helpdesk first, then observation, then redesign |
| Trade users seeing internal notes | A permission model separating staff, trade users and admins |
| Customers seeing numbers that are wrong | Customer dashboards only after internal numbers were trusted |
Outcome
Support became faster and much more controlled, the business can see its own service performance, and trade customers check the state of their cases without asking.
Starting with a plain helpdesk and redesigning around observed behaviour kept us from designing a workflow for a process nobody actually followed.
What the client keeps
- The production helpdesk
- The permission matrix
- Workflow and SLA configuration
- Pipeline code, in their own repository
- Editable dashboards and the metric definitions behind them
Sources and further reading
- Helpdesk, Odoo documentation
- Metabase documentation, Metabase
- Authorization Cheat Sheet, OWASP
Facing something similar? This is the kind of work our Enterprise & Digital Transformation team delivers every day. Talk to us about your stack.
