Sign up for our newsletter and get the latest big data news and analysis.

Robots Keeping Shelves Stocked: How Machine Learning and AI is Helping the CD Industry Stay in the Game

In this special guest feature, Amjad Hussain, founder and CEO of Algo.ai,, observes that whereas 20 years ago, labels were able to flood stores with CDs in the knowledge surplus inventory would inevitably sell over time, doing so now could lead to huge losses. As such, companies are increasingly turning to new smart technology which harnesses the power of machine learning AI to accurately predict future demand using existing data. Algo.ai is a workflow automation platform which connects artificial intelligence, augmented reality, and automation to retailers, distributors, and manufacturers.

In the age of Spotify, it may appear CDs are predestined to suffer the same fate as ‘the radio star’. However, while sales are dropping year over year, the mid-year music revenue statistics from the RIAA (Recording Industry Association of America) report $245 million in CD sales in the first part of 2018 alone. Rather than drying up completely, the CD business is evolving into a niche, somewhat like vinyl, and this year store owners across the US, have been struggling to keep CD shelves stocked for popular releases.

Whereas 20 years ago, labels were able to flood stores with CDs in the knowledge surplus inventory would inevitably sell over time, doing so now could lead to huge losses. As such, companies are increasingly turning to new smart technology which harnesses the power of machine learning AI to accurately predict future demand using existing data.

Machine learning is perfectly suited to inventory optimization due to its ability to predict future buying behavior. So, how is ML helping the CD industry keep on whirling, despite the huge challenges it faces?

1: ML Can Help Labels Highlight Hidden Pockets of Faithful CD Fans

While leading retailers like Best Buy are cutting CDs from their stores, there are still hidden pockets of consumers who are still buying discs with gusto. For example, a recent Fast Company article notes high sales for CDs within the indie genre, and niche discovery platform Bandcamp reported 18% year-over-year growth in CD sales for 2017, up from 14% growth in 2016.

Life would be so much easier if niche music fans would all form communes and live in the same area, but as it stands, labels need to use big data and AI to find and target these groups.

Considering that most ‘big box’ stores have drastically cut the space available for stocking CDs, labels need to use data to choose which shops to stock with which records. Moving past the capabilities of descriptive, or even predictive analytics, prescriptive analytics combines optimization techniques popular in Operations Research with Machine Learning. Using a range of data inputs such as historic sales data, local demographics, and online reviews, we can make accurate sales predictions for each release, and from there prescribe the optimal stock level for each location.

2. ML Can Help Stock the Right Discs in the Right Stores

But when it comes to selling CDs, it’s not as simple as sending out batches to geographic hot spots. CD manufacturers release special edition discs, which carry a much higher margin than normal CDs. These discs are considerably more expensive, partly due to the fact that they can be re-sold at high values as collector items further down the line.

As such, music agencies need to use data to go deeper, to work out not only which artists are most likely to sell in certain locations, but which areas to send limited edition discs.

ML and predictive analytics can gather and analyze comparable data for a certain artist to find out how many discs were sold in particular locations, but also to include other factors, such as artists with a similar style who sold well in this area, artists from the same producer, or genre, to forecast how many will sell. But algorithms can also use this data to highlight which types of a certain disc sold, be it the normal or special edition, and with what frequency.

3: ML Can Help Brands React to Consumer Trends Quicker

In the internet age, consumer trends can change fast. A drunken rant on social media or poor review online can quickly sway consumers away from patronizing a particular artist or brand. As such, labels need to be able to react fast to any changes in the environment.

Smart inventory tools are combining ML-based anomaly detection, which looks at unusual changes in time series data, and natural language processing (NLP) to allow teams to pinpoint opportunities or issues immediately, understand the huge amount of information in front of them, and act quickly. Using natural language instead of dashboards or programming, business people can request information and give commands to algorithms simply using their own voices, making even the most “old-fashioned” feel comfortable using the new technology.

And in this industry, time is of the essence. Big box stores like Target will not hesitate to send back products immediately if they think they won’t sell. But with the superpowers of ML and NLP on their side, labels can move just as fast as consumers and limit risks.

So, can smart inventory management tools help CD manufacturers stay relevant for much longer, or will CD simply become the new VHS? After all, no one collects VHS, regardless of how much of a fan they are. But for now, thanks to a bastion of loyal CD fans, and a technological mix of AI technology, shelves are still being stocked with CDs. But just the right amount.

 

Sign up for the free insideBIGDATA newsletter.

Leave a Comment

*

Resource Links: