Is SKU still needed or XDM object field mapping using other product attributes can handle it?

[Rafael_Sahagun]

Hello,

Since SKU and classifications is being considered by this specific business too time consuming to maintain based on the amount of products and attributes, there is a discussion in about moving away from using a SKU in Adobe Analytics (and would think later when CJA is implemented) since there are product attributes already in the in the XDM object field and more can be added to different dimensions (evars in AA).

On the other hand I also thought about using only SKU and not having to send all the values in the XDM object (thinking of load speed and simplicity) and later use classifications and even add more attributes and make it retroactive, if not too much level of effort.

Is SKU still needed? we'd lose retroactive value loading option I'd thing but other vars still can be classified if needed even if not from the typical SKU (or tracking id) style value.

Any implications for AEP/CDP in going either route?

I think the best option is to look for optimized ways to keep SKU up to date and have the ability to add other attributes down the road if needed and use xdm mapping to populate product events a business wants and also is partly implemented.

Thanks in advance.

R

[Asheesh_Pandey · Accepted solution]

@Rafael_Sahagun  Yes, keeping a SKU or product ID is still important mainly as a stable key for retroactive classification, data joins, and product-level analysis in both Analytics and Experience Platform. I'd recommend following -

• Keep SKU or product ID as a stable key

• Minimize real-time attribute payload by including only key those needed immediately

• Use classifications (AA) or join tables (CJA) for additional product metadata 

• Automate SKU metadata ingestion into AEP via batch sources or catalog APIs

• In future, when using CJA or RTCDP, link datasets via SKU

This keeps your ecosystem scalable, joinable, and retroactively extensible, while keeping performance in check.