What product data should not be stored in a PIM system

What this episode covers

The fifth episode answers the question that saves the most trouble during an implementation: which data not to put into a PIM system. Karolina introduces the split between hot and cold data, shows it on a laptop as an example, and explains why this is a technical distinction rather than an academic one.

Key takeaways

• Hot data changes often and is needed in real time: stock levels, promotional prices, availability, orders. Its home is an ERP or WMS.
• Cold data is stable: marketing copy, technical specifications, images, video, manuals. PIM systems were built for it.
• A PIM system is not built for real-time synchronisation. Exchange with e-commerce usually happens once or twice a day.
• Transactional and personal data stay out too: order history, payments, invoices, email addresses, phone numbers.
• One exception: a recommended retail price held as a fixed value on the product is cold data and can serve as a base for further calculation.

Hot and cold

Hot data changes often and is critical to current operations. Because of that dynamic nature it needs frequent updating and is usually required in real time: stock levels, current prices including promotional ones, availability in stores, order information. Those belong to ERP and WMS systems, designed for frequent updates and continuous synchronisation.

Cold data changes rarely and is stable over time. It is the basic specification and description of the product: marketing copy, technical specifications, images and product video, manuals. PIM systems were built precisely for it, because it can be centrally managed and enriched before going anywhere else.

Take a laptop. Into the PIM system go the name, category, marketing description, processor and disk details, the main image, dimensions, weight and the manual. Out stay the number of laptops in the warehouse, the current promotional price and the order history.

Why this is a technical argument

The key sentence of the episode is this: a PIM system can hold a very large amount of product information, but it is not technically designed to synchronise data in real time.

And with stock levels and promotional prices, real time is exactly what is expected. Synchronisation between a PIM system and e-commerce usually happens once or twice a day, which does not guarantee that the customer sees true availability. So that information should come from systems dedicated to order handling and sales.

The distinction also matters when planning integrations between systems: stock and price are a different data stream from a product specification, with a different cadence and a different source.

What else stays out

Beyond the hot and cold split there is data that simply does not belong in a PIM system, no matter how often it changes. Transactional data, meaning customer order history, payments and invoices. And personal data: email addresses, phone numbers.

The exception worth knowing

Prices can enter a PIM system when we mean the recommended retail price, held as a fixed value on the product. Such a price helps people segment products in the system, or serves as a base for further calculation in e-commerce.

In that framing the price stops being hot data and becomes cold, because there is no risk of delay in passing it on in real time. Then there is nothing against it.

How to apply this

During an implementation it is worth walking through your own product data attribute by attribute and asking one question about each: does this change often and relate to order handling and sales, or is it a stable part of the specification. If the latter, it belongs in the PIM system.

If you are wrestling with product data, get in touch. More on the PIM systems page.