Technical Advice I Wish I Received Before Go-Live

When you first go live on ServiceNow, the focus is all about hitting the date and getting the basics working. What nobody tells you is that without proper measures in place, the real technical lessons only show up afterwards.

Suddenly you can find yourself dealing with sluggish forms, integrations that pull too much data, or a data model that doesn’t survive a clone. None of these problems are unfixable, but I’ve learned that the way you design things up front makes the difference between quick tuning and constant firefighting.

Here's my advice.

[1] Protect the User Experience

You can get ahead of performance issues by thinking about design early. It isn’t just about indexing the fields people actually search on or keeping nodes balanced. The way the UI behaves day-to-day matters just as much.

Large reference fields such as users and groups should always be filtered with qualifiers or lookups. When you design them well, the system feels faster and more focused.  When you don’t, users feel the pain long before the logs show a problem.

[2] Know your Integration Limits (and Opportunities)

Out of the box, you have platform-level integration capabilities and IntegrationHub Starter. That covers more than most people realise if you make use of the free spokes properly.

The real trap is how data is moved. Too many integrations try to pull everything at once when the better approach is to paginate or batch. Combine that discipline with a clear view of what is included in the licence and what is not, and you avoid both overbuilding and overpaying.

[3] Design your Data Model

Whenever there is a request for new fields or tables, spend longer than you think investigating the “why” and the “how.”

ServiceNow gives you options: extend Task, use CMDB, rely on data lookups, or build custom tables. Each option has consequences.

One that often gets overlooked is how the model survives cloning. Without clone profiles, lower environments can lose critical data, leaving you with test systems that behave nothing like production.

Summary

Most technical issues after go-live are fixable. Performance, integrations and data models can be tuned once the cracks appeared. The key lesson is that building them right in the first place saves you from fixing the same problems again and again.