Lucky Brand's Approach to Successful Merchandise Allocation

blog-clothing-rack

The idea behind inventory allocation seems pretty straight-forward: Ensure stores are kept in stock with respect to their ability to sell merchandise. Simple enough, right? Yet, as an allocator, one end-of-season scenario you’re likely familiar with is the aching realization of a missed opportunity — where better inventory allocations in one store could have led to more transactions, fewer markdowns, and improved margins.

Of course, you and I both know the subtle art to successful allocation is far from simple. In fact, a recent U.S. retailer survey estimated inventory misjudgments, including misallocating inventory, account for more than half (53%) of unplanned markdown costs. This issue has prompted retailers, like Lucky Brand, to search for ways to infuse intelligence into their allocation systems and processes.

Mike Relich, COO of Lucky Brand, recently shared his experience leveraging Celect artificial intelligence and machine learning (AI/ML) software to successfully improve retail allocation results. 

Here are the five biggest takeaways:

1. Balance Gut and Intuition with Data

For a long time, retailers relied on averages to manage their business (for example, all large doors are assigned the same amount of inventory, etc.). Advances in AI/ML over the last five years now make it possible to inform inventory allocation decisions by considering how consumers like to shop, product nuances and a product's relationship to its surrounding assortment. In the short video below, Relich touches on how Lucky Brand leverages Celect Allocation Optimization to make recommendations for each item across stores, driving data behind the gut-driven decisions merchants make.

nrf-lucky-allocation-where-to-put-it-1-1

2. Make Use of Attribute Data

When it comes to fashion, items change from season to season so there isn’t typically much historical data for a particular product. In this segment, Relich shares how Lucky Brand uses product attributes to influence their allocation decisions. With Celect Allocation Optimization, Lucky Brand is able to identify which attributes are the most statistically significant and then uses this input to allocate new product in a smarter way.

nrf-lucky-attribute-level-allocation-buy-2

3. Break Away from the Old Paradigm

Normally, buyers rely on store groups to drive product buys. One particular buy will be for all stores, another will be for the top fifty, and so on. With Lucky’s approach, Relich’s team is leveraging AI and ML to understand the location-specific demand of existing inventory to a.) inform allocations and b.) avoid over-recommending styles to certain sets of stores, breaking away from the old paradigm between merchandise buys and allocations.

nrf-lucky-over-recommended-stores-allocation-3

4. Adopt a “Shuffle the Deck” Approach

Most allocation systems provide a line of styles for you to choose from. From here you make calculations based on weeks of supply or the latest three-week trend for that style, allocate, then repeat. The problem with this approach, according to Relich, is the allocation decisions you make don’t take into account the context of other styles – you’re making each decision in isolation from one another rather than looking at the whole pool of inventory and their context to one another. The “shuffle the deck” approach utilized by Celect allows Lucky Brand to allocate multiple styles at once to make much more informed, accurate allocation decisions with a full view of how each style interacts with each other.

nrf-lucky-why-celect-allocation-4

5. Integration and Adoption of Advanced Analytics

As with any technology integration, success relies on how easy it is to get the data out of your existing systems. However, when it comes to adoption, Relich describes his team’s reaction using Celect AI/ML and how it’s transformed the way they do business:

nrf-lucky-integration-adoption-allocation-6

“What blew my mind was the adoption of the tool – because with system implementations, you always set the expectations to be as follows: it’s going to take twice as long, do half of what you expect, and cost twice as much,” says Mike Relich.  “This was not the case at all with Celect. After sitting down with the allocation team, they loved it. The adoption has been so amazing – you can come in and select the items you want to allocate, the UI is great, it’s really compressed our training and it’s very very intuitive.”

You can access the full presentation by Mike Relich here.  

Topics: allocation, artificial intelligence, inventory optimization, brick-and-mortar retail, physical stores, machine learning, merchandise assortment

Ready for more?

Request a Demo