AI Meets Category Management: What’s Ahead for Brands?
CATEGORY MANAGEMENT is perhaps the most ubiquitous business activity in the retail consumer goods industry. It has 30-plus years of history as a human-centered, expert process.
With the arrival of artificial intelligence (AI), some human-centered aspects are facing transformation.
“AI integration in category management is expected to usher in a new era of efficiency, agility, and strategic decision-making,” observes PwC in a research report, Reimagining Category Management with Artificial Intelligence. “As organisations [sic] embrace these technological advancements, they position themselves to thrive in an increasingly competitive and dynamic global marketplace.”
To understand these implications, it is worth considering what artificial intelligence is – and is not.
An MIT Professional Education bulletin defines AI as “the ability for computers to imitate cognitive human functions such as learning and problem-solving.”
Focus on the words “artificial” and “imitate.” AI solves many questions by sheer brute force – absorbing more data in minutes than we could ever read in a lifetime. They don’t think as humans do, but they are trainable. They can get better and faster over time at answering questions we need to ask routinely.
In his groundbreaking book, Understanding Media, communications philosopher Marshall McLuhan famously called mass media “extensions of man.” AIs share many parallels with this. They too are extensions of human capabilities, modeled after us, but ultimately synthetic.
Beyond the Walled Garden
Category management has proven to be a remarkably enduring business process on its own. The classic 8-step model was originally developed by Dr. Brian F. Harris, a USC professor, ca. 1989. Many variations have been put forward over the decades.
The model is notably recursive – conceived as a cycle of periodic improvements to the category plan. But it can also be rigid – especially category definitions and roles. Each category plan is ultimately an average across the range of shopper preferences and tendencies.
Always a highly data-driven activity, category management has long relied on two realms of information that focus on recent sales history as the basis of decision-making. Retailers control much of this within what some call their “walled gardens” – demand/POS data, shopper loyalty data, baskets, trips, inventory. Retailers have long recognized the power this affords them.
Category captains – brand partners – have traditionally folded in syndicated and household panel data, market share data and survey research to make the cases for their category and promotional plans.
Newer sources of data from the digital realm carry potential to change the game, with new ways to visualize future demand. Retail media has engendered whole new data sets based on message delivery, personalized offers and conversions. Digital life reveals new insights from analysis of text, images and other unstructured data.
Can AI Bring Faster, Better Decisions and Actions?
AI provides tools to get to the issues faster. It can tirelessly track more data sources on an ongoing basis. It carries immense potential to identify breakthrough opportunities. Some promising examples:
- Access more data from more sources previously inaccessible to category planners
- Recognize patterns in data humans can’t easily see
- Spot changing market conditions to respond more rapidly and accurately
- Monitor relentlessly, learn and send alerts to trading partners
- Evolve category planning from hindsight to foresight
- Identify most valuable shopper segments who buy the brand or will buy the brand, and define strategic opportunities
- Trigger “next-best action” merchandising/competitive responses.
While these are exciting areas of potential faster and finer category decision making, it brings a paradox for trading partners. Compliance and inventory inaccuracy remain steep challenges. What will happen when AI-enabled plan adjustments come at a higher frequency? Can you deliver?
When it comes to performance at the shelf, brands and retailers still need to confront a tough reality that has been holding back category performance for decades. IHL Group, which diligently tracks inventory accuracy, released retailer survey data late last year in its Shelf Intelligence Report that brings this issue into focus. When retailers self-reported their own performance, they rated their planogram compliance accuracy at just 57%. On-shelf availability accuracy was scarcely better, at 58%, while promotion compliance accuracy was 60%.
This sad level of follow-through suggests that much of the effort and expense invested in category and promotion planning today is not delivering on expectation. Worse, lax implementation makes an accurate understanding of plan effectiveness nearly impossible.
In an AI-enabled era just ahead, can we even hope to implement accurately at a faster pace? Perhaps AI can offer a remedy for that too.
Consider Next Best Actions
If you concede that you can’t implement every task well, it becomes very important to decide which to focus on first, which to handle if you can, and which to abandon as non-productive. The principle of next best action (NBA) holds great promise here.
When applied in marketing, NBA is defined as an AI-driven strategy that uses real-time customer data and predictive models to determine the single most effective message, offer, or action for an individual.
In retail merchandising, NBA principles can be applied to determine which at-shelf tasks have the highest priority on a given visit, given the available resources of the store team. AI tools are very good at providing this sort of quick guidance, especially when a merchant snaps before and after images of each shelf set and sends them to the cloud for analysis.
Long and Short of AI
As we consider the near-term impact of the AI revolution on category management, it may be instructive to recall Amara’s Law, famously uttered by tech guru Roy Amara way back in 1978: “Humans tend to overestimate the impact of new tech in the short run and underestimate it in the long term.”
While the AI hype cycle is in full swing, it is already creating near-term value for category managers on several fronts:
- AI automates some interactions using agents
- AI enhances human judgement, creativity, and collaboration
- AI cuts time spent on reporting; leaves more for building relationships
- AI demands less time for chasing data, but helps humans ask better questions.
Of particular interest to category planners, AI is allowing planners now to identify most valuable segments who buy a brand in a particular category, or who will buy the brand based on the indications that they're pulling in from other resources.
“With AI we can more clearly identify the size of the new prize,” said Dr. Harris, who continues to push the frontier of category management practice at his latest venture, Intent AI. “You’ll want to go after the breakthrough opportunities which motivate trading partners.”
