Normalization Conundrum in Large SN environment. 10K+ tables.

[Beany81487]

Hi everyone,  

I’m working through some challenges around **normalization and mapping** in a large ServiceNow environment, and I’d really appreciate some guidance on where to begin.  

Here’s the scenario:  

• Multiple CIs are linked to multiple contracts.  
• Those contracts are in turn tied to multiple catalog items.  
• We also have multiple vendors and product records, some with duplicate or inconsistent names. 
• Some are created through automation, and some have been entered manually. 
  I am not sure if the environment has ever had a mass clean up and restructuring of assets. I don't think they have. 

As you can imagine, this creates complexity in reporting, fulfillment, and overall data integrity. The biggest struggle I’m running into is knowing the **most effective starting point** for cleanup.  

Should the baseline focus be on:  

• **Vendors and products** → normalize and rationalize first, then re-map downstream?  
• **Contracts** → treat them as the anchor object and then rebuild relationships?  
• **Catalog items** → ensure alignment and fix the back-end D-mapping so items aren’t left out of flows?  

I’m aware that incorrect or incomplete D-mapping at the catalog layer can leave items orphaned or misaligned, which makes automation and reporting unreliable. That’s the area I feel least confident in, and I want to avoid reinforcing bad relationships in the data model.  

If anyone has tackled normalization at this scale, I’d love to hear how you approached it, what “baseline” you started with, and any best practices for working across contracts, CIs, vendors, and catalog items in a structured way.  

Thanks in advance for your insights!  

— Cristina

[Mark Manders]

Start with how the process should be and have agreement on that. You can start a cleanup, but if you move all the snow of the sidewalk and when you are done, a snow plough clears the snow from the street onto the sidewalk, you have wasted time. You need to know how your records are going to be created as of now. You need to know how the contracts are going to be related as of now. And so on.
As soon as that's in place, you can start cleaning up. Tweaking of the process is still possible (hindsight and what more), but it will remain clean after you are done with the first phase.

And the start point of the clean up is indeed normalization (rationalization is part of the process). Get rid of duplicates. Make sure you understand the reporting needs as well! Nothing is more frustrating than taking on such a huge workload and then find out that the final result does not allow for all necessary reporting.