top of page
Search

Travel Coordinates in D365: Is the Juice Worth the Squeeze?

Beau Schwieso
1 day ago
6 min read

*Important note: Microsoft doesn't call this feature "Travel Coordinates" but I do. It's technically called "Warehouse spatial location" but that sounds too nerdy for me.



Your warehouse team can have a very productive morning according to their step counters and a very frustrating morning according to your shipping schedule.


Back down the same aisle. Across the building for one more item. Past the staging area, then away from it again.


Everybody is working. A surprising amount of that work might be walking.


That is why the idea of using coordinates to improve picking caught my attention. But before we start putting savings in a PowerPoint, let's get specific about what Microsoft actually delivered.


First, what is this feature called?

If you have heard it described as “travel coordinates,” (aka you're me and probably me alone. ok, maybe Dag also but you'd have to of been at the Houston DynamicsCon Roadshow and watching the same session as me) the matching capability in Microsoft's documentation is Warehouse spatial location in Dynamics 365 Supply Chain Management, the warehouse side of what many of us still call F&O.


Microsoft's setup documentation specifies version 10.0.49 or later and the Warehouse spatial location feature in Feature management. As of this November 9, 2026 review, the documentation still labels it preview, which Microsoft says is not intended for production use. That is a documented version prerequisite, not a verified general availability date.


For now, think sandbox evaluation and a business case you can validate.


What the coordinates actually do

This is not a reason to put “warehouse GPS” in your project charter.


The documented feature sequences work using location data. I would not infer live worker tracking, congestion detection, obstacle awareness, or turn-by-turn navigation from that description.


There is also a very specific boundary: only pick lines before the first put line are sorted. Later pick/put cycles remain unchanged. Work needs at least two pick lines, valid destination coordinates, and at least one valid pick location; failed validation leaves the original sequence intact.


That boundary matters when you estimate how much of your daily workload can benefit. Count eligible work, not every warehouse transaction.


It also does not fix poor slotting. Microsoft treats slotting as a separate capability addressing where items belong before orders are released. [3] If your highest-volume items live in an inconvenient corner, improving the order in which people visit that corner still leaves you with an inconvenient corner.


And I would be careful with the word “distance.”


A coordinate model can suggest that two locations are close. A person standing between them may see a rack, a cage, or an aisle they can only enter from the other end. City Block provides a grid-based distance estimate; I would not assume that it understands every permitted crossing in your building.


Likewise, the shortest walk may be a poor loading sequence. In food distribution, I would specifically test heavy cases, fragile product, temperature zones, and pallet-building requirements. Saving a walk is not much of a victory if someone then rebuilds the pallet.


Those are operational acceptance tests, not claims that the feature automatically enforces those constraints.


One technical detail worth testing: height

Microsoft's model adds the absolute heights of both locations for the vertical component. Two locations at Z = 2 therefore contribute 4 vertical units, not zero. [2]


That deserves attention if your operation involves lifts, elevated picking, or several rack levels. My recommendation is to validate the model against how your equipment actually operates before treating calculated distance as elapsed time.


More broadly, agree on a coordinate convention. Mixing aisle numbers on one axis with feet on another can produce a ranking that looks precise without representing useful physical distances. Document the origin, scale, and height convention, then have an operations lead walk a sample.


A field containing a number is not the same thing as trustworthy data.


A small implementation plan

Step

What I would do

Evidence needed to move forward

1. Establish the baseline

Observe comparable picks across several normal shifts. Separate travel, picking, waiting, and rework.

Current travel minutes per work record, completion time, errors, and order mix.

2. Select the scope

Choose work with meaningful travel and multiple pick stops. Identify sequencing constraints with floor supervisors.

A clearly defined eligible workload and agreed acceptance criteria.

3. Prepare the data

Map all pilot pick and destination locations, using one coordinate convention. Assign a maintenance owner.

Coordinates checked against the actual floor, including awkward routes.

4. Configure in a sandbox

Enable the feature, configure warehouse strategies, and assign location coordinates. Add Sort picking work lines after work creation in the wave template. [1]

Work sequences inspected before any operational trial.

5. Compare the alternatives

Replay representative work with the current sequence and each relevant configuration. Walk or safely simulate the routes.

Measured travel and completion time, plus wave-processing impact.

6. Make the investment decision

Review results with operations, IT, and finance. Confirm production availability and support before planning rollout.

A funded benefit, an owner, and a tested rollback approach.

I would include deliberately inconvenient cases: missing coordinates, one-pick work, multiple put cycles, widely scattered picks, and locations separated by barriers. If you use cluster picking or custom mobile flows, test those actual flows too. A clean sales-order demo is not evidence for every execution pattern.


For high volumes, Microsoft documents optional parallel batch processing for the sorting step. Measure first. Adding processing capacity before understanding the bottleneck is an expensive way to feel busy.


For rollback, retain the original wave configuration and test the return to it in the sandbox. Inspect existing work separately; do not assume changing setup reverses work that has already been sequenced.


Where the ROI lives

Finally, the juice is worth the squeeze part. If you've reached this far, great. If you haven't, not so great. If you skipped all the way here, even greater-er.


The first mistake is applying a travel improvement to the entire labor budget.


Suppose travel represents 40% of the time spent on the eligible picking process. If the pilot reduces that travel time by 15%, the direct improvement is roughly 6% of that process time before other effects.


Not 15% of the warehouse payroll.


The second mistake is assuming every recovered hour becomes cash.


Here is an illustrative business case. These are assumptions, not Microsoft benchmarks or a promised result.

Input or calculation

Example

Eligible pick work records per day

400

Measured travel time saved per eligible work record

1.5 minutes

Operating days per year

250

Annual capacity recovered: 400 × 1.5 ÷ 60 × 250

2,500 hours

Loaded labor value per hour

$30

Gross annual capacity value

$75,000

Share converted into a measurable financial benefit

50%

Annual realized benefit

$37,500

Annual maintenance and support cost

$5,000

Net recurring annual benefit

$32,500

One-time implementation cost

$25,000

Simple payback: $25,000 ÷ $32,500 × 12

About 9.2 months

First-year net return: ($37,500 − $5,000 − $25,000) ÷ $25,000

30%

The 50% assumption needs an explanation. Perhaps the improvement reduces overtime, avoids planned temporary labor, or postpones a hire that volume growth would otherwise require.


If staffing and paid hours stay the same, the immediate result is capacity. That may still be valuable. The team might clear the backlog, absorb growth, or finish before a carrier cutoff. Just give that capacity an operational use and an accountable owner.


If additional shipments create value, use incremental contribution margin and confirm there is demand to fill the capacity. Do not count the same recovered hour once as labor savings and again as extra throughput.


The implementation cost should include mapping, configuration, testing, training, and any necessary remediation. If an upgrade is required solely to obtain this capability, include that incremental cost too. Ongoing costs include coordinate maintenance and regression testing when layouts or processes change.


And run the pessimistic case.


If the measured improvement is only 30 seconds per work record, the same example recovers about 833 hours annually. At the same labor rate and 50% realization, the annual benefit is $12,500. After $5,000 of upkeep, simple payback stretches to 40 months.


Same feature. Very different investment decision.


What would convince me is fairly ordinary: a supervisor walking the proposed route, a picker explaining why it works, and finance agreeing on where the recovered time goes.

If those three agree, there may be real juice here.


Until then, the step counter is the only one guaranteed to hit its target.

DynamicsDad


















 
 
 

Comments


bottom of page