Beyond the Next Step: Why Legacy Warehouse Logic Is Costing You Miles

In fulfillment warehouses a common strategy to gain picking efficiency is to cluster or batch pick multiple orders simultaneously.

The challenge with this strategy is that the “efficiency” of this picking relies on how your software sequences that batch; ultimately determining if your picker walks two miles or half a mile to complete the same set of walk. When an e-commerce picker sets off with a cart designed to hold 10 customer orders simultaneously, they aren't just looking for 10 items; they are managing a matrix of 10 to 100 distinct items scattered across thousands of square-feet.

Most teams generally rely on one of three core strategies to group and route these multi-order trips: Rule-Based WMS Logic, a Greedy "Next-Best" Approach, or a Predictive Path Optimization, like Lully’s Pick Optimizer.

Here is how each strategy handles a 10-order pick run, and why the decision engine, or lack thereof, behind your carts makes or breaks your labor margins.

1. Rule-Based WMS Logic: The Siloed Assembly Line

Traditional Warehouse Management Systems (WMS) rely on rigid business rules to structure work, such as "group 10 orders by client", "cluster all single-item orders," or "all orders in Aisle 1". Your team spends time not only implementing these “strategies”, but also maintaining them (if they remember).

Order Selection Strategy: The WMS slices the order pool based on your predetermined rules. The rules have no context of where items are, nor how full a piece of equipment actually is. It’s as simple as “find first 10 orders that match a criteria”

The Pick Strategy: The system forces the cart through a predetermined, linear sequence based on location sequence, regardless of where the specific items for those 10 orders are located.

Routing Strategy: Follows statically assigned location pick sequence

The Reality on the Floor: The picker takes their 10-order cart down nearly every aisle in the facility. Your cleverly placed rack cut throughs are ignored. The weight of equipment as units are picked wasn’t factored in. You’re pickers are hoping to get lucky with dense picks.

The Result: High predictability, but a massive waste. Pickers spend up to 60-70% of their shift pushing a heavy 10-order cart past empty shelves just to honor static system rules.

2. The Greedy Approach: The Local Firefighter

To avoid walking down empty aisles, many modern fulfillment apps use a greedy algorithm for order selection. Many software vendors will talk about this as “the holy grail”, they may even describe it as “orchestration”. Here’s how it actually works. The software looks at all open orders, has a criteria for selecting anchor order (oldest/smallest/largest, etc) and then selects the next 9 orders that are “most similar” to the anchor order. This sounds GREAT in theory, but leaves a ton to be desired in the real world.

The Pick Strategy: The algorithm quickly chooses an anchor order, often the oldest or smallest order, then quickly evaluates the remaining order pool across to select the next “best” order based on the previously selected orders. Best is often defined as having the same item as the previous order, but often ignores the fact that the order may expand the selected item pool.

Routing Strategy: The greedy algorithm isn’t creating trips with awareness of other trips, it's simply taking the best near term decision it can find. As a result, while the system can minimize distance of the trip by running a routing algo separately, doing so does not address the holistic problem and is likely to continue the congestion problems you already have.

The Reality on the Floor: You’ve effectively put glitter on top of the already not-so-great static rules.

Item A is common across all of the orders, which is great. But what if item “D” is on the other side of the building? 10 Orders all containing item “D”, or even “A, B, C, D” are entirely missed.

The Result: The greedy approach solves the problem of managing endless configs and profiles in your system, but it operates with severe tunnel vision. By constantly chasing short-term wins, it repeatedly creates its own traps.

3. Holistic Decision Engine: The Master Navigator

Predictive Path Optimization doesn't ask, "Which of these 10 orders has the most items in common?" It doesn’t consider “What is the shortest path for this inefficient trip?” It asks, "Given all orders in the system, how should they be organized to globally minimize travel distance, minimize congestion, and ship everything on time?"

The Pick Strategy: Before the picker moves their cart, the engine evaluates hundreds of thousands of groupings and route permutations across all customer orders simultaneously, accounting for aisle directions, physical constraints, and total item distribution.

The Reality on the Floor: The system recognizes that bypassing a nearby item in Aisle 3 early on allows the picker to grab it seamlessly on a natural return loop from Aisle 8. It maps out a single, serpentine sweep through the store floor that fulfills all 10 orders in one fluid motion.

The Result: Zero unnecessary backtracking, minimal physical strain on the picker, and there’s 20 more trips planned right behind it that are just as good. Steady, consistent, and efficient.

Strategy Comparison Matrix

The Bottom Line

When managing fulfillment operations, small grouping and routing errors multiply exponentially. Rule-based systems enforce unnecessary steps, and greedy algorithms fall into short-sighted distance traps. Only Predictive Path Optimization views all of the work to be performed as a single unified puzzle, delivering the shortest path, higher picks per hour, and lowest operational overhead.

Determining if a system is using a rules based approach is straightforward. But knowing if its using a simple greedy algorithm and disguising it as “intelligence” is a bit more nuanced. The easiest way to tell? Ask how often the order grouping analysis runs: When you hear something like “every few seconds”, you’re looking at a greedy algorithm implementation.

Tags: warehouse management path optimization fulfillment efficiency batch picking order routing