Blog
Predictions 2021: PIM
DynamicWeb
Product Information Management (PIM) is the foundation of a modern eCommerce strategy. In an increasingly AI-driven world, the ability to structure and expose your product data to AI-facing applications will be a critical differentiator.
To stay competitive in the ever-expanding eCommerce landscape in 2021, savvy businesses are now looking to product experience as their number one differentiator.
Predictions 2021: PIM
As manufacturers, distributors and wholesalers adopt Product Information Management (PIM) software to provide better online product experiences across channels, they often realize just how much effort it can take to aggregate, categorize and enrich all the necessary data. As an outsider to the product data that fuels eCommerce-related industries, one may think that this information is ubiquitous and readily available online waiting to be captured. Unfortunately for wholesalers and distributors, accessing and aggregating that information for their own purposes across thousands of products can pose a real challenge.
If only you could snap your fingers and all that manual work would be done!
Also don’t miss our other Predictions 2021:
→ Predictions 2021: Content Management
→ Predictions 2021: Digital Marketing
Predictions 2021: PIM and AI
Luckily, that world may not be too far off.
AI and machine learning continue to embed themselves into the applications and processes we as humans use everyday. Have you ever enjoyed a movie that was recommended by Netflix? How about ordering a product with Alexa using only your voice?
One thing is for certain: structuring your data and exposing it to AI-enabled applications will become increasingly important for online consumers and internal employees to provide modern product experiences.
Computer vision and product enrichment
One of the most prevalent AI processes is Computer Vision – enabling thousands of images to be viewed, cross-referenced, identified, and categorized in seconds. We see this type of technology whenever we perform a reverse image search. Another computer vision application you may have used is the “search by camera” feature within the Amazon.com app.
Computer vision is not only used to identify what an image contains, but can also assign additional attributes with some additional logic. Here’s an excellent example of this using Google’s Cloud Vision API and a rubber glove as the product.
It won’t be long before advanced PIMs using AI will be able to utilize descriptors paired with image data to automatically categorize or enrich products, like personal protective equipment. Imagine what this looks like at scale, when you point the computer vision processor at a folder containing 50,000 product images.
Natural language processing, voice search, and ordering with digital assistants
Another popular application for AI is Natural Language Processing (NLP), which processes text to identify patterns and ultimately derive meaning that can be used for next steps or analysis.
Voice commands and voice search are already becoming a popular way to control devices and applications (like Amazon Alexa). PIM solutions will enable organizations to more readily participate in growing trends like voice search and voice command with structured, widely-distributed and constantly up-to-date product data.
Bringing it all together with machine learning
Like any emerging technology, AI isn’t error-proof. Human intervention and auditing is sometimes required to make adjustments.
Enter machine learning.
As a human operator confirms or denies results, machine learning can be used to train applications on how to better perform by reducing errors and improving accuracy. The promise held by the blending of computer vision, NLP and ML paints a fantastic future.
Imagine processing images to not only categorize thousands of products, but to further enhance this data with other relevant and relatable data from within your own organization or via cloud services. Or consider the benefits of product analysis or scoring, with deeper insights on how to optimize your product data than today’s “completeness” scoring models that are found in intelligent PIM solutions.
Time will tell how soon these predictions come to fruition – but we wager they’re not too far off.
Bonus: How do others foresee the future of PIM? 🔮
The above predictions are based on our assumptions. To broaden the perspectives, we've asked three of our partners across the world to share their predictions on the future of PIM. You can watch the livestream on-demand or listen to it as a podcast.