In Adobe Analytics has anyone modified the default 'Products' implementation from product name (words) to product number (number)?

[KG_10]

We found an inconsistency in the product name that is skewing reporting.  We would like to switch and use the product number instead.  We know there will be a disruption in the data & reporting, but think the long-term benefits are greater.  At the time of implementation using the product name was recommended. We are implementing the product classification import and need a valid product key.  Any learnings from anyone that has made a similar change?

[yuhuisg]

By "product number", I assume you're referring to the SKU. If so, then I've heard of some implementations that track the SKU in place of the product name, then use Classifications to add metadata. This is the same process as you've described.

Then, the difficulty is usually around managing the Classifications. Classification Rule Builder might not work since SKUs don't normally have obvious patterns to match to the metadata. If you rely on manual Classification imports, then you have to ensure that you're always importing up-to-date Classifications for new/updated SKUs. The best solution would be to automate the Classifications, using a script to get the latest product information routinely from your product database and send the metadata to AA via an API call.

Aside from handling Classifications, I don't think there are any huge roadblocks to doing what you've considered.

[KG_10]

Thanks for the reply.  Yes, we would switch from product name to SKU.  I've seen SKU implemented successfully at a previous employer.  We would be using a script from our product database to import the latest attributes on a regular basis into AA.

We know there will be a change in data from the point of implementing the SKU.  We were wondering if any other AA users had switched their product default implementation and were there any unexpected impacted items (so we could plan ahead). 

[Stewart_Schilling · Accepted solution]

Yes, I've done this quite a few times for clients.  Using a product number (whether it is the SKU or some other unique product identifier) is always the best way to go. 

Since you will be doing SAINT (classification) uploads anyway, here's the trick that will make your analysis of pre/post data much easier.  In addition to uploading classification rows with the new product keys (SKU's), also upload rows with the old product keys (product names).  It will look a bit like this: 

By doing this, you will be able to use the "Product Name" classification to report across the old data and the new data in a single report.

Note also that I added a few other classifications in the example above (Product Category, Product Gender, and Product Brand).  If you will be pulling info from your Product Information Management system anyway, you might as well get more than just Product Name.  Spend a bit of time exploring this topic - it can turbo-charge your reporting. 

[KG_10]

Thanks, Stewart, for sharing.  That's an interesting trick to explore & consider.  I'll look into it.