Localization is a key objective of most high-volume retailers today, and even if that’s not necessarily their strategy, having the right product that relates to the needs of your most valuable shoppers always in stock will support sustainable sales performance at the shelf.
Managing the product mix can be best accomplished with an assortment optimization solution and accompanying approach that puts customer needs and behavior at the center of your decision making. This process goes hand in hand with your merchandising efforts to ensure that planning and strategies can be effectively communicated and executed in store.
Discover tips and hints dedicated to helping you evaluate your assortment optimization and category management software.
Historical approaches primarily focused on the category, not necessarily the shopper
In the past, assortment decisions were commonly derived via an unstructured business process, using ad-hoc analysis from syndicated sources delivered by data providers such as IRI and Nielsen – and the final product mix was visualized by the resulting planogram. This, of course, was the process gap that the category management pioneers like Dr. Brian Harris and The Partnering Group designed in order to bring a more formal structured methodology to during the 1990s.
Evaluating how to manage the assortment tail was an idea reflected in questions I would receive from customers during the dot-com boom as they began to feel competitive pressures from new online competitors. Chris Anderson did a wonderful job explaining this predicament in his 2006 book The Long Tail. The common solution at the time was to view assortment related data in a Pareto chart – allowing the visualization of additional volume that each additional SKU would contribute if it were added to the mix. This insight was still only directional, as what was missing was a true measure of incrementality, since these analyses didn’t account for the potential effect of transferrable demand as an item entered or exited a category.
The assortment optimization solutions that followed, attempted to manage the assortment tail and did the necessary math to calculate sales for the new product mix, however they forced an endless trial-and-error game in order to achieve best results. In addition, most of these models were completely black box in nature and you had no way to understand why you got the results you did or configure your strategy and objectives to reflect the suitable tactics for a category.
Category objectives are not homogenous
Not all categories and products will/should have the same tactics. So-called “routine” categories are typically used to build transaction volume and profitability. “Destination” categories are often treated as must-win transactions or to generate customer excitement. “Convenience” categories may be used to enhance the retailer’s image. Lastly, “seasonal” categories are used to create excitement and generate profit. Legacy assortment optimization solutions often treated all product segments as if they played the same role, and because of this, recommendations would often be missing out on generating optimal results.
What should you consider when evaluating an assortment optimization solution?
1. Support for best-practice approach
There are many frameworks for category management that are designed to make the process more efficient for category practitioners however, in my experience, those that are most successful focus more on the end customer and their shopping preferences and behavior rather than being product centric. I subscribe to the Category Management Association’s CatMan 2.0 process, which provides a sound framework for most food, drug and mass retailers to build a strong foundation for their category management efforts, as it attempts to account for both behavioral and attitudinal shopper insight. Look for a solution that enables a best practice approach.
2. Consumer decision trees
Whether you have used formal research methods to develop consumer decision trees – or if it is years of collective experience – today’s assortment solutions can and should leverage this valuable insight, and provide a structured repository for this information.
3. Category assessment
A historical perspective on performance, and comparison to the market, is important to understanding missed opportunities and under/over performing product segments. Support for category assessment may lead you to redefine the category based on this understanding – consider, for example, the trend of organic items.
4. Objective driven
Look for a solution that will allow you to guide the results and drive the metric(s) important to you, i.e. loyalty, profit, volume. Next, look for one to provide you a recommended assortment mix that maximizes your objectives. Gone are the days of having to play the trial-and-error game, results need to be quickly and efficiently delivered. There should be some flexibility to set different objectives for each product group or geography to reflect different tactics for each product segment or store cluster.
5. Localized results
Delivering localized assortment recommendations is a must-have. The solution should be flexible and scalable to do this at a store- or cluster-level.
6. Full inventory model
It is important that the analyses account for thresholds for case pack, presentation quantities and days of supply.
7. Ability to control churn
This ability is needed for minor reviews to limit the number of items impacted.
8. Mandatory or prohibited items
Preferably, this is informed through an automated data-feed – providing the ability to make core items mandatory, to limit regional item availability, or to restrict prohibited items. Your solution needs to support this at the store and cluster level.
9. Modeling transferrable demand
A key functional requirement is to be able to model household data to calculate the true incrementality of each item, in relation to all other items in the assortment.
10. Demand forecast for new-to-market items and items not carried
All items under consideration of the assortment review should have a forecasted sales volume when actual sales is not available. Ideally, this is a store-level forecast.
11. Artificial intelligence
AI is the latest innovation in assortment optimization. The use of machine learning can be utilized to analyze big data sources, and to identify trends and patterns – before your competitors do.
12. Space aware
Understanding your available merchandisable shelf space is critical. Also critical is the true item capacity – calculating for different merchandising and stacking methods for each individual product.
13. Understanding the “what” and “why” of the results
Reporting and analytical views for both pre- and post assortment optimization should be readily available. This includes Category Scorecards, distribution reports, and yes… Pareto charts still have a place. Additionally, in the case results are unexpected, you should be able to have visibility into the logic of the system to validate why each assortment decision was made.
Having an effective and efficient way to handoff the assortment decisions to the merchandising team is important. This needs to include the adds/deletes/retains for each cluster or store.
15. Internal and external collaboration
Depending on your category management process, there may be collaboration – including approval steps – with internal stakeholders. It also may include collaboration with external stakeholders, both upstream and downstream. Ensure your solution has collaboration built-in.
16. Measure and track execution
Tracking actual results against projections provides for improved execution and better insight for future optimization.
Challenge traditional methods for a winning assortment strategy:
- Learn more about best assortment optimization strategies in our recent webinar
- Discover the 5 factors a category manager should consider when selecting a category planning solution
- See how you can deliver effective assortment optimization