Product Category Analysis: How to Determine Performance Drivers

Category analysis should ultimately enable category teams to make better decisions and drive brand growth. In order to do so, category teams must understand what’s driving business performance.

Performance drivers dictate where category teams can take actionable steps that will result in the greatest potential impact.

With the uncertainty of COVID-19, this degree of precision is essential.

Category managers at CPGs are feeling the pressure of siloed data and retail partners who are increasingly investing in their own analytics resources. Likewise, retailers are juggling rapidly changing consumer behavior and massive amounts of data. Plus, retailers still have a lot of ground to cover when it comes to building up their in-house analytics teams.

Let’s dive into how CPGs and retailers can determine performance drivers with category analysis.

  1. Why performance drivers are essential for category analysis
  2. Centralizing data for category analysis
  3. Understanding category performance with KPI trees
  4. Automating category analysis

Why performance drivers are essential for category analysis

When changes occur in category and brand performance, tons of different metric combinations could be the root cause. For example, let’s take any of these possibilities into consideration:

  • Changes in shopper baskets
  • Growing channels like curbside pickup and e-commerce
  • Supply chain disruption
  • New consumer trends
  • Fluctuating consumer confidence
  • Shifting customer loyalty due to early pandemic out-of-stocks

…and so on. This list is nowhere near exhaustive, and each of these factors could be simultaneously impacting performance. What matters the most is understanding the degree to which each factor contributes to and detracts from brand and category performance.

Only with this precise information can category teams create efficient and effective strategy.

Otherwise, category teams risk losing both growth opportunities and market share.

To put this problem into context, many shelf-stable categories like rice and pasta have excelled during the pandemic. With healthy top-line metrics, it’s easy to take that performance at face value. However, a performance driver analysis could reveal a significant decline in total distribution points, masked by other high-performers like volume price.

Without this critical information, supply chain issues could go unaddressed, and CPG category teams could lose market share in the long run as consumers steadily commit to alternative brands on the shelf.

That’s why understanding performance drivers is so essential. It creates the full picture of what’s happening to a category or brand so that every growth opportunity is on the table.

Centralizing data for category analysis

CPGs and retailers have different challenges with data, but both struggle to centralize data for a comprehensive view into business performance. Retail category managers need access to store and shopper data, such as panel, scanner, and loyalty data. Likewise, CPGs need access to internal and syndicated sources, such as retailer,  finance, and marketing data.

In both cases, data is often siloed— meaning, category teams can’t get a 360° view to truly understand performance drivers.

At least, not without manually merging these sources together. This process is time-consuming and prone to human error. In addition, siloed data creates an obstacle to cross-functional collaboration between category teams and other business functions. When every team has their own view of data, the complete picture is skewed.

The ideal outcome of centralized data is a holistic view of top-line and margin growth drivers. From there, teams can seamlessly collaborate to determine results from a single source of truth.

As such, category teams should strongly consider an analytics solution that enables them to easily connect their various data sources to a single hub. This kind of enabling technology should extend beyond data to the actual category analysis, which brings us to our next discussion point.

Understanding category analysis with KPI trees

Category analysis usually centers around a key performance indicator, or KPI. Sales in a category or brand is a common KPI to demonstrate overall performance across a specific period of time.

Understanding the drivers behind sales means understanding the complex relationships between metrics in large data sets. When it comes to analysis, there are innumerable metric combinations that could explain the root cause behind change. Out of all the possibilities that could impact category sales, which combination is right?

It’s impossible for humans to perform an exhaustive analysis that accounts for all of these factors. CPGs and retailers need insights that they can act on, so even the best category teams can’t assess every possible data combination, especially with tight deadlines and pressure from COVID-19.

This is where enabling technology like AI and augmented analytics can become a competitive advantage.

AI creates a common metric hierarchy and analyzes category performance across datasets. Machine learning algorithms can parse through data to analyze all potential drivers of a KPI in minutes or even seconds. From there, the ML process can determine which drivers are contributing positively and negatively to the overall performance.

Comparatively, an analyst would take days or weeks to identify drivers manually (and even then, their analyses would be limited). Plus, analysts approach data with their own biases and assumptions, whereas AI does not.

With COVID-19, safe assumptions are actually quite risky. Consumer behavior has changed rapidly, and it’s unclear which trends will stick; as a result, category teams need real-time insights into performance drivers to stay on top of opportunities and avoid pitfalls.

Granular, precise consumer insights can move the needle towards success.

That’s where a KPI tree is so valuable. A KPI tree provides a decomposition of the relationships between metrics so that category teams can easily understand the degree to which each metric affects the main KPI. This view enables category teams to weigh metrics based on their relative level of impact and make decisions accordingly.

Category analysis KPI tree shows which metric most impact sales

This KPI tree illustrates that declining total distribution points is the key detractor of sales with a -$21.5M impact.

If category teams can achieve this view into their data, they can direct their time and energy into the strategic activities that will drive growth.

While a KPI tree serves as an incredibly valuable visualization tool, it’s not complete without insights that explain the “why” behind category and brand performance.

Automating category analysis

The category analysis workflow we’ve outlined thus far includes:

  1. Centralizing various data sets into a single source of truth
  2. Analyzing every data combination to determine key performance drivers
  3. Generating category and brand insights (which we’ll speak to in this section)

This process can be automated with augmented analytics and AI. As described in the previous section, AI can actually perform analysis to exhaustively test every data combination and generate a KPI tree. This is only one component of what AI can do for category teams.

Along with a KPI tree, AI can generate natural language insights that highlight important information. These insights can draw attention to critical metrics, frame KPIs within the context of larger trends, and explain not just what’s happening in a category, but why.

This “why” is critical because it puts performance drivers into a larger narrative. Total distribution points may be the most significant decliner impacting sales, but what does that mean in dollars? To what degree are other metrics, like out-of-stocks, playing a role in the TDP losses?

Insights can answer these questions, providing a comprehensive narrative of brand and category performance.

In order to manually generate these insights, category teams would need to interpret data and visualizations to draw conclusions, selecting which aspects of the story to include. The challenge with this method is that it’s time-consuming and prone to bias, just like merging and analyzing data.

Ideally, category teams should spend their time acting on the insights they receive and making data-driven decisions. Understanding performance drivers enables them to operate with agility, precision, and certainty— essential advantages for the COVID-19 landscape.

Ultimately, category teams should consider automation through augmented analytics to get the best view into performance drivers.

Additional Resources

  • CPG Analytics Guide: CPG analytics allows you to leverage syndicated and first-party data to understand category and brand drivers. Learn how to enable cross-functional collaboration and get faster insights amid COVID-19.
  • Retail Analytics Guide: Retail analytics can connect all of your data sources, producing real-time insights that help you stay on top of changing consumer behaviors. Learn how to plan effectively even amid the pandemic.
Scroll to Top