Sometimes it seems that technology can deprive us of our humanity. Too much screen time, not enough true reality. Whether we’d like to admit it or not, most of us are guilty in some capacity of spending more time focused on social media likes from strangers than making important connections with the people who matter to us most. It’s not a stretch to say that scenario is similar to what happens with category practitioners – we get so consumed with what we can measure and analyze on our own that we lose sight of what’s really important.
It makes sense that category management has operated this way. Traditional curation has focused mainly on margins, top-line sales, volume, store size, and location. Don’t be alarmed, these are of course still important criteria. But like a retweet from a verified Twitter account, unless you can identify the factors that led to your success and use them to establish a replicable model for making it happen again, the sense of achievement will rapidly wear off.
The challenge in both scenarios is that we don’t have a way to always know what our audience wants or have a winning formula to address their seemingly arbitrary whims.
Predicting future demand – Fad vs. True Trend debate
It helps to consider this in the context of shifts in consumer demand. Sometimes tastes evolve slowly, but then again, we would have all seen fads and trends affecting both our daily lives and businesses that appear to have hit the market like a tsunami. The question is, did you know each of them were coming before they impacted your life or work, and do you know how long they would last?
Today, the likely answer is a resounding ‘no.’ Most retailers, with support from their CPG manufacturers attempt to optimize their categories and make planning decisions by compiling data from POS systems, loyalty programs and syndicated data sources. This information is used to identify gaps or duplicates in their inventory, high-value product all-stars and low-performing duds, and to establish new evaluation criteria that will improve accuracy on the next cycle – such as reducing slotting fees to open up the market to smaller, promising their suppliers that have their finger on the pulse of unique consumer groups. However despite this aspiration, many categories are still riddled with ineffective duplications (anywhere between 14-17% duplication in category) of product attributes, but why?
Besides the inaccuracy issue, these category review processes are time consuming and often quite disruptive. Even when data is fully available, and when the talent is there to review it, they still have the fatal flaw of only looking backwards. They aren’t able to properly consider the complexity of the human element, and they offer no reliable way to predict what comes next.
The question you have to ask is: “Am I willing to overlook the next can’t-miss product, or be stuck with massive low selling highly discounted inventory when a fad suddenly dies out simply because I had to guess at what my customers wanted next?”
AI heralds the opportunity of continuous assortment optimization
Fortunately, the old ways of category optimization are set to expire. With the rise of machine learning and natural language processing (NLP), category managers are able to leverage the full context of their consumer behavior – not just what they buy, but the full scope of their shopping and lifestyle behaviors.
AI is becoming a massive supporting technology for CatMan 2.0 – in which retailers have the ability to incorporate the full landscape of shopper insights. In other words, AI leverages rules-based algorithms to automatically analyze large amount of data to identify patterns, ultimately “learning” how they interact to better understand current and future trends. Machine learning has the power to review and learn about customer behavior at a far deeper level than any merchandising team could before it and provides prescriptive recommendations to merchandisers that are easily acted upon. The system tells the merchandiser what’s important, instead of the other way around, supporting an agile merchandising approach.
Natural language processing provides another interesting development to capitalize on new data analysis capabilities, both on spoken or written English. For example, NLP can analyze social media conversations and gauge the dichotomy between what consumers say and what they do and understand true intent – accurately and before your competitors do.
What does this actually mean for daily lives of category practitioners?
When it comes to identifying trends, significant demographic shifts, and your most valued customers nuances, AI can provide a comprehensive and penetrative understanding of shoppers’ buying patterns. It can keep shelves stocked with the right merchandise mix and ensure that the supply chain is aligned to eliminate expensive out-of-stock or overstock scenarios. It can perform these tasks in an automated, predictive, and real-time manner.
It can also determine when a category review is critically needed and when it is not. This makes the AI process far less cumbersome and disruptive than traditional, calendar-based reviews. In fact a number of retailers I have spoken too who look out how best to leverage this technology, they can envisage the review calendar potentially disappearing all together. Instead a continuous refinement of categories, driven by AI, becomes a welcome sight for many retailers. Rather than big bang reviews causing disruption of inventory in their supply chains, high store operations costs and sometimes shopper confusion in the aftermath of major resets at the shelf, a steady alignment to market and consumer needs driving a sustainable category growth trajectory is understandably a very appealing concept.
The irony of human based merchandising
It’s not lost on me that establishing a prescriptive, human-based merchandising strategy relies on artificial intelligence and machine learning. But AI isn’t designed to operate independent of people; it’s designed to augment our capabilities and make our work more effective. Fortunately, as machine learning and NLP proliferate the retail industry, barriers between people and business process break down – and we’ll find it much more rewarding than just chipping away trying to find small incremental gains across disparate parts of the business.
Take your next step
Learn more about how applying the “Rule of 17” can drive profitable revenue growth in your category planning processes.
Meet CINDE, the AI-powered personal decision coach. Learn how she’s giving an unfair advantage every day to category managers.