Most ServiceDesk Plus setups do not fail on the install. They fail in week three, when the categories are wrong, the mail fetching was never tested, and half the staff were never imported. The fix is boring and total: do the configuration in the right order.
ServiceDesk Plus will happily let you click settings in any order - and it will just as happily let you build on sand. This guide gives you the sequence that experienced admins use, and the reason behind each step, so your setup survives contact with real tickets. The order is: organisation and mail, technicians and roles, categories, requesters, templates. Each step is small; together they are the difference between a help desk and a form nobody trusts.
Create your organisation with its sites and departments first, because tickets, SLA policies and reports all hang off these names later - renaming a site after six months of tickets mangles your history. Set the time zone and business hours now, for the same reason: the SLA clock is only as honest as this calendar. Then configure the mail server, both directions. The incoming mailbox is fetched at intervals and every mail in it becomes a ticket - this is your safety net for staff who will never open a portal. The outgoing settings send the acknowledgements, assignments and resolution notes. Test both directions with a real mail from your own inbox before moving on. A help desk whose first test ticket silently vanishes teaches the whole office one lesson: this new system eats requests.
Technicians are your licensed seats, so add them deliberately. Give one or two people the administrator role - they can change settings, and that power should be rare. Everyone else gets a technician role matched to what they should see and do. Then organise technicians into support groups that mirror how problems arrive: Hardware, Network, Software and Accounts, or whatever your office's real split is. Groups matter because assignment rules, SLA escalations and reports all speak in groups. If your structure is "whoever is free picks it up", say so honestly and use one group - a fake structure produces fake routing. Note who is in each group and who backs them up on leave days; the ticket queue does not know about holidays until you tell it.
The category tree - Category, Sub Category, Item - looks like housekeeping and is actually your reporting future. Every chart you will ever show a manager groups by these words. Build it short and in requester language: Internet and Network, Email, Printer, Laptop and Desktop, Software, New Joiner and Exit. Under each, two or three sub categories at most. Resist the urge to encode your org chart or your troubleshooting tree into it; a requester choosing a category is guessing at a label, not diagnosing. A practical test: read each category name to someone outside IT. If they cannot tell you what belongs in it, rename it. You can add categories later with far less pain than merging bad ones, so start lean.
Requesters - the staff - should arrive by import from Active Directory or LDAP, not by typing. Imported requesters log in to the portal with the account they already use, their department and site come along for reporting, and new joiners and leavers stay in sync with the directory instead of with someone's memory. Send one imported colleague a test: log in, raise a ticket, watch the status change. Only after people exist do templates make sense. Build request templates for your five most common asks, each asking only for what a technician needs to begin: for a new joiner, name, department, start date and manager; for a printer problem, floor and printer name. Every extra field is a small tax on every future ticket, paid by the busiest people you have.
See the admin console, the mail settings and the ticket structure being built live in the video walkthrough.
Watch Full VideoManual lists rot: joiners cannot log in, leavers keep access, departments go stale. An AD or LDAP import keeps the requester list, logins and reporting fields matched to the directory your office already maintains, and requesters stay free - licences are counted per technician only.
Start with roughly six to ten top-level categories in plain staff language, with two or three sub categories each. You can add more any time. Merging or renaming categories after months of tickets is the painful direction, so lean beats complete at the start.
Yes - roles, groups and assignments are everyday admin work, not one-time decisions. What is genuinely hard to change later is the organisation structure and category history, which is why this guide does those slowly and the rest quickly.