Case Background

What Actually Happened at Target Canada — and Why It Matters for This Post

The Decision Moment

You are the VP of Operations, four weeks post-launch. Corporate dashboards still show green. Store managers are calling in saying: “We have product; we can’t sell it.” Your CEO wants a 90-day recovery plan grounded entirely in descriptive analysis — no advanced modeling, no forecasting algorithms. Just: what does the data actually show, right now, at the ground level?

In 2013, a major U.S. retailer opened more than 100 Canadian stores in under two years. Within months, shoppers walked into empty shelves for everyday staples — even as distribution centers and store backrooms reportedly held plenty of product. It wasn’t a supply problem. It was a visibility problem. Leadership was looking at aggregate dashboards that masked what was happening at the floor level.

The root causes were operational: dirty product data (wrong units of measure, incorrect item dimensions, missing classification codes), system integration failures between the new ERP and replenishment systems, and an aggressive rollout timeline that skipped the basic data validation checks that would have caught problems before stores opened. The venture closed less than two years after launch.

Your discussion post asks you to step into that moment — not to explain what went wrong in hindsight, but to describe what a VP of Operations should have been looking at from day one, and what those views would have revealed.

100+
stores launched in under two years — skipping foundational data validation
~2 yrs
before full closure — one of the most expensive retail failures in Canadian history
5
descriptive artifacts your post needs to identify, define, and connect to operational fixes
500
words for the initial post — professional memo style, conceptual only, no numbers
Key Concept

Descriptive vs. Predictive Analysis — Know the Difference Before You Write

The assignment is explicit: conceptual, no calculations, no predictive modeling. That constraint is doing real intellectual work. It’s forcing you to think about what you can see in data right now — before any forecasting, before any machine learning, before any algorithm tells you what will happen next.

Descriptive analysis answers: What is happening? Predictive analysis answers: What will happen? The Target Canada case is a failure of descriptive visibility — leadership didn’t know what was happening on the floor. Not because the data didn’t exist. Because nobody was looking at the right cuts of it.

TypeQuestion It AnswersExample in This CaseAllowed in This Post?
Descriptive What is happening right now? Which stores have more than X% of shelves empty for a given category? ✓ Yes — this is the entire assignment
Diagnostic Why is it happening? Which item records have missing or incorrect unit-of-measure fields? ✓ Yes — part of what descriptive views surface
Predictive What will happen next? Which stores are likely to run out of stock in the next 30 days? ✗ Out of scope — don’t go here
Prescriptive What should we do to optimize? Which replenishment algorithm should we use going forward? ✗ Out of scope — don’t go here
📌

What “Conceptual, No Numbers” Actually Means

You’re designing the dashboard, not running it. Describe the variable — what it is, what it would show — without making up or citing specific percentages or counts. Say “shelf-level availability broken out by category and store” not “73% of shelves in Category A at Store 14 were empty.” The intellectual value is in knowing which variables to look at and why, not in producing a fictional dataset.

Core Assignment Requirement

The Five Descriptive Artifacts — What to Include and How to Frame Them

Each artifact needs three things: what it is, what problem it helps uncover, and how it would support an operational fix. That structure maps directly to the rubric. Here are five artifact categories that fit this case — and how to think through each one. You don’t have to use these exactly. Pick five that you can explain clearly and connect logically to the Target Canada situation.

01

Shelf Availability by Store and Category

In-Store Visibility
What It Is
A count or proportion of shelf positions that have product present, broken down by individual store and by product category. Not an aggregate — a per-store, per-category view that shows where product is physically absent from the sales floor.
Problem It Uncovers
Corporate averages might show acceptable overall in-stock levels while specific stores — or specific categories within stores — have critical gaps. A paper goods aisle that’s consistently empty is invisible in a systemwide average. This view makes individual failure locations visible.
Operational Fix
Store managers can be tasked with targeted replenishment on the specific categories showing the worst gaps. It also tells the DC operations team which product streams are failing to make it from backroom to shelf — so you know where to direct manual intervention first.
02

Item Data Completeness Rate

Data Quality
What It Is
A count of SKUs in the item master that are missing required fields — unit of measure, item dimensions, tariff codes, category classification, or language-specific labeling. Broken down by category, supplier, or DC.
Problem It Uncovers
This is the root cause artifact. Product can sit in a DC and never be replenished to the shelf because the system can’t process it correctly — wrong dimensions mean it can’t be slotted; missing codes mean the replenishment trigger never fires. The item data completeness rate shows you exactly how many items are operationally broken before a customer ever encounters the empty shelf.
Operational Fix
Prioritize a data remediation sprint on the categories with the worst completeness rates. Items can’t sell if the system doesn’t know what they are. This view tells you which categories to fix first — and gives you a measurable indicator of progress as the 90-day recovery proceeds.
03

Backroom-to-Floor Transfer Rate by Store

Flow Bottleneck
What It Is
A view of how much product is received into store backrooms versus how much of that product is actually making it onto the sales floor within a standard time window. Shows where product is getting stuck at the last mile — the backroom-to-shelf handoff.
Problem It Uncovers
This is why “we have product; we can’t sell it” is the defining symptom of the Target Canada case. Product existed. It just wasn’t moving from backroom to shelf. A low transfer rate in a specific store tells you the bottleneck isn’t the supply chain — it’s execution at the store level: insufficient labor, process failures, or system issues blocking the replenishment trigger.
Operational Fix
Deploy additional labor to high-backlog stores as a rapid intervention. Investigate whether the replenishment system trigger is functioning correctly for the affected stores. This view also reveals whether the problem is uniform across all stores or concentrated in a subset — which shapes where you direct resources first.
04

DC Inventory Age and Dwell Time

Distribution Center
What It Is
How long specific items or pallets have been sitting in a distribution center without being shipped to stores. Broken down by category, DC location, and item — shows where product is effectively trapped in the network.
Problem It Uncovers
Long dwell times point to systemic replenishment failures upstream of the store. If items are sitting in DCs for weeks, the system isn’t generating the outbound orders to move them. This could be a data quality issue (the item can’t be processed), a system integration gap (DC system isn’t talking to replenishment logic), or a slotting problem (item dimensions won’t fit into assigned DC slots).
Operational Fix
Manual review and expedited shipping of high-dwell items to stores with confirmed shelf gaps. DC managers can prioritize clearance of aged pallets while the underlying system issue is diagnosed. It also flags which categories need the most urgent data remediation — because long dwell times often trace directly back to bad item data.
05

Price Scan Error Rate by Category

Customer Touchpoint
What It Is
A count of transactions where the scanned price does not match the shelf label price, broken down by category and by store. Captures the data quality problem at the point where it meets the customer — the register.
Problem It Uncovers
Price mismatch errors are a symptom of item master data failures — when item data is wrong or incomplete, the system populates the wrong price in the POS. High scan error rates in specific categories confirm which parts of the item catalog have the worst data quality. They’re also a direct customer experience failure — shoppers lose trust immediately when what they see on the shelf doesn’t match what the register charges.
Operational Fix
Prioritize price data cleanup for categories with the highest error rates. Implement a manual shelf-label audit for worst-performing stores as an interim measure. The scan error rate also serves as an early-warning indicator — categories where it spikes first are your leading signal for deeper data quality problems elsewhere in the catalog.
💡

Choosing Your Own Five Artifacts

The assignment doesn’t prescribe which five you use. Pick five you can explain clearly and connect directly to the Target Canada scenario. Other valid options include: replenishment order exception rate (orders the system flagged but didn’t process), item setup completion rate by category (what percent of items have gone through the full setup checklist), store-level on-shelf availability vs. backroom inventory ratio, or supplier item data quality audit scores. Whatever you choose, make sure you can explain what it is, what problem it surfaces, and what fix it enables.

Core Case Question

Why “Green Averages” Misled Leadership — and What Your Post Should Say About It

This is one of the four case questions your post needs to address. It’s also the conceptual heart of the whole assignment. Get this right and the rest of your post flows naturally from it.

An average is a compression. It takes a distribution of values — some low, some high — and collapses them into a single number. That number might look acceptable, or even good, even when the distribution underneath contains stores or categories that are in complete operational failure.

Think about it this way. If 80 stores are running at 90% in-stock and 20 stores are running at 15% in-stock, a systemwide average might land somewhere that looks acceptable on a green-threshold dashboard. But those 20 stores are catastrophically failing their customers. The average told leadership “green” because it was built to report composite performance, not to surface outliers.

The average is the enemy of the exception. And in operations, the exception is what breaks your business.

— Core principle of operational diagnostics

What the green averages failed to separate in the Target Canada case specifically:

What Averages Hid

  • Store-level performance variation — the worst stores buried inside a system mean
  • Category-level failures — aggregate in-stock masking empty aisles in specific departments
  • The backroom inventory vs. floor inventory distinction — total product count looked fine; floor availability was not
  • Data quality exceptions — a dirty item record doesn’t lower an average if the system still counts the inventory as “available”
  • DC dwell time — product sitting in the network doesn’t show up as a shelf gap in aggregate metrics

What the Right Views Would Have Shown

  • The distribution of in-stock rates across individual stores, not the mean
  • Category-level shelf availability broken out from DC inventory
  • Exception counts — how many items flagged errors, not whether errors existed
  • Backroom-to-floor movement rates, showing where product was stuck
  • Item data completeness rates, showing which records were operationally broken
⚠️

The Framing the Assignment Is Looking For

Don’t just say “averages hide outliers.” That’s too generic. Connect it specifically to what happened: the aggregate in-stock metric looked acceptable because DC inventory and backroom inventory were counted alongside floor inventory — without separating them. The system saw product in the network and reported it as available. It wasn’t. Floor availability, item data completeness, and transfer rates were the views that would have surfaced the actual problem. Name the specific view that the average obscured, and explain why separating those views matters operationally.

Case Question

Which Segmentation Would Matter Most — and How to Argue It

The assignment asks which segmentation matters most: store, category, distribution center, or region. The correct answer isn’t one of those in isolation — it’s a cross between store and category. But your post needs to make the argument, not just name the winner.

🏪

By Store

Surfaces which locations are failing worst. Lets you prioritize where to deploy intervention resources first. A store-level view is where operational action happens — store managers, labor allocation, and in-store processes are store-specific. Without this cut, you’re managing a system average, not a real problem location.

📦

By Category

Shows which product types have the worst data quality, shelf availability, or transfer failures. Category-level failures often trace to category-specific item data problems — wrong units for a category, systematic misclassification, supplier data issues concentrated in certain departments. Knowing which categories are failing tells you which data remediation sprints to prioritize.

🏭

By Distribution Center

Identifies where product is accumulating in the network without moving. A DC-level view is essential for diagnosing upstream flow failures — if one DC is processing orders correctly and another isn’t, you have a system integration or process issue specific to that DC. Without this view you can’t separate a supply chain problem from a store execution problem.

🗺️

By Region / Language

Specifically relevant in Canada because Quebec’s French-language labeling requirements added a distinct compliance layer. Stores in French-language markets may have faced specific item data failures tied to bilingual label requirements. A regional segmentation surfaces whether performance failures were geographically clustered — which would point to region-specific regulatory or operational issues.

For your post, the strongest argument is that store × category is the most actionable segmentation. Store tells you where to act; category tells you what to fix. Together they give a VP of Operations a prioritized list of specific problems at specific locations — which is what a 90-day recovery plan needs. Pure store-level data doesn’t tell you what to fix. Pure category data doesn’t tell you where to go fix it. The cross-cut of both gives you that.

Case Question

How Descriptive Analysis Drives a 90-Day Recovery — Without Predictive Modeling

The assignment asks how descriptive analysis alone can guide a 90-day fix. This is where students often drift into vague language about “monitoring dashboards.” Be more specific than that. Descriptive analysis guides recovery by creating a prioritized action sequence — and that prioritization is exactly what a VP of Operations needs.

Think of the 90 days in three phases, each driven by a different descriptive view:

90-Day Recovery Framework — Descriptive Analysis at Each Phase
Days 1–30
Triage
Views driving action: Shelf availability by store and category; backroom-to-floor transfer rates. Use these to identify the ten worst-performing stores and the three most affected categories. Deploy field teams to those specific locations with authority to manually move product from backrooms to shelves while the underlying data problems are addressed. The descriptive views tell you where to send people — right now, today — without needing a forecast or model.
Days 31–60
Data Fix
Views driving action: Item data completeness rate; price scan error rate by category. The triage phase bought time. Now you’re going after the root cause. Rank all categories by item data completeness rate. The ones with the worst rates get dedicated data remediation teams — correct unit of measure, fix dimensions, resolve classification errors. Track daily completeness rates as the progress measure. No modeling required — just a count of broken records, declining over time as fixes are applied.
Days 61–90
Stabilize
Views driving action: DC inventory age and dwell time; replenishment order exception rates. With the worst data problems corrected, verify that product is now moving through the system correctly. DC dwell time should be declining. Replenishment orders should be processing without system exceptions. Store-level shelf availability should be improving in the categories that received data remediation. These descriptive views serve as recovery KPIs — not predictions about the future, but current readings that confirm the fixes are working or flag where more intervention is needed.
✅

The Key Argument: Descriptive Views Create the Priority Queue

You don’t need a predictive model to know what to do next. You need to know where the problem is worst right now, and what the underlying cause of that failure is. Descriptive analysis gives you both. It ranks problems by severity (store availability gaps ranked worst to best), points to root causes (item data completeness failures in specific categories), and provides a measurable recovery signal (completeness rates and shelf availability both improving over time). The 90-day plan is essentially: look at the distribution, not the average; fix the worst outliers first; track the descriptive view as your recovery KPI.

Writing Your Post

How to Structure the 500-Word Discussion Post

The assignment says it should read like a short professional memo. That means it needs to be structured, direct, and free of academic filler. No long preambles. No “In this memo, I will discuss…” Just get into it.

Memo Structure — “Make Reality Visible”
Opening
(~50 words)
Frame the problem, not the history. One or two sentences establishing the decision moment: what the dashboards say, what the store managers are reporting, and what the 90-day mandate is. Don’t spend half the post summarizing the case. Your reader already knows it. Get to what you would do.
Five Artifacts
(~250 words)
This is the bulk of the post. Name each artifact clearly. One sentence on what it is. One sentence on what problem it surfaces. One sentence on the operational fix it supports. Keep each one tight — about 50 words per artifact. Don’t overexplain. The memo style requires brevity and precision, not essay-length justifications. Consider using a numbered list or bolded artifact names to make the structure scannable — a memo reader should be able to locate each artifact at a glance.
Case Questions
(~150 words)
Address all four questions in the second half of your memo: why green averages misled leadership (the aggregation problem); which two or three of your five artifacts would have surfaced the crisis earliest (connect to earliest-signal views like shelf availability and item data completeness); what segmentation matters most (store × category cross-cut, and explain why); how descriptive analysis drives the 90-day plan without predictive modeling (describe the prioritization logic). Don’t treat these as four separate mini-essays — weave them together as a coherent operational argument. About two paragraphs covers all four.
Closing
(~50 words)
One crisp closing sentence or two on what the core lesson is: leadership can’t manage what it can’t see at the right level of granularity. The fix starts with disaggregation, not forecasting. Land on a specific, assertive note — not a general observation about data being important.
⚠️

What Loses Points on This Post

  • Treating the five artifacts as generic dashboard metrics — each one needs to be specifically tied to the Target Canada scenario, not just named as a type of data that’s generally useful in retail
  • Including calculations or percentages — the assignment explicitly says conceptual and descriptive only. No made-up numbers
  • Missing any of the four case questions — green averages, earliest-signal views, segmentation, and 90-day plan are all required in the post
  • Writing it like a reflective essay — this is a professional memo. Structure matters. Get to the point fast.
  • Generic artifact definitions — “sales data broken down by store” is too vague. Specify what exactly you’re looking at and what operational question it answers
Reply Post

How to Write the 50–100 Word Reply That Actually Adds Something

The assignment says: do more than agree or disagree. Build the discussion by questioning, extending, or refining their ideas. That’s a specific intellectual move — and it has three forms.

MoveWhat It Looks LikeExample Opener
Question Challenge an assumption in their artifact choice or segmentation without dismissing it — ask them to defend the logic further “You chose store-level availability as your earliest-signal artifact — but would that view have been available in the first week of launch, before store operations had generated enough scan data to be meaningful?”
Extend Take one of their artifacts and add a layer they didn’t consider — a different segmentation, a secondary use case, or a dependency they didn’t mention “Your item data completeness rate is strong — I’d add that it becomes even more powerful when broken out by supplier rather than just category, because it shifts accountability upstream to where the bad data originated.”
Refine Suggest a sharper or more specific version of their artifact — make it more operationally useful “Your ‘inventory visibility’ artifact is on the right track, but I’d tighten it to the backroom-to-floor transfer rate specifically — that’s the view that maps directly to the case’s central symptom of having product that can’t sell.”

Keep it under 100 words. Name their idea specifically — don’t write a generic comment that could apply to any post. One focused move is better than three shallow ones. End with a question if it fits — it keeps the conversation going, which is the point of a discussion board.

💬

What “Building the Discussion” Actually Means

Read your classmate’s post and find the one thing they said that you can push on or add to. Not their whole post — just one specific claim, artifact, or segmentation choice. A good reply sounds like a real conversation, not a response form. “I noticed you prioritized DC segmentation over store-level segmentation — here’s why I’d argue the store × category cross-cut gets you to the action faster” is exactly the kind of reply the assignment is asking for.

Need Help Writing This Discussion Post?

Our operations management writers understand case analysis, descriptive methods, and the professional memo format this assignment demands. Get a well-argued, rubric-aligned post drafted by someone who knows the material.

Get Professional Help Now →
Common Questions

FAQs: Target Canada Case Discussion Post

What exactly is a descriptive artifact in this context?
A descriptive artifact is a specific variable or view of operational data that makes the current state of a system visible — without prediction or modeling. In the Target Canada case, descriptive artifacts are the specific things a VP of Operations would look at to understand what’s actually happening on the floor: which shelves are empty, which item records are broken, where product is stuck. The key word is specific — not “inventory data” but “backroom-to-floor transfer rate by store.” Each artifact needs to have a clear operational meaning and connect to a specific failure the case describes.
Why did the “green averages” mislead Target Canada leadership?
Averages aggregate performance across all stores, categories, and locations — collapsing a distribution of values into a single number. If the majority of stores are performing adequately, the aggregate average can look green even when a substantial subset of stores is in complete operational failure. In the Target Canada case specifically, aggregate inventory metrics counted product in DCs and backrooms alongside floor inventory — so the system reported adequate coverage while shelves were empty. The average didn’t separate floor availability from total network inventory. It also didn’t surface item data exceptions — broken records don’t lower an average if the system still counts the item as existing in the catalog.
Which descriptive views would have surfaced the crisis earliest?
The two earliest-signal views would have been item data completeness rate and shelf-level availability by store and category. Item data completeness is a pre-launch check — it reveals how many SKUs are operationally broken before a single customer encounters the problem. It could have flagged the crisis before stores even opened. Shelf availability by store and category is a day-one operational metric — it would have immediately shown which stores and aisles were failing rather than letting the problem hide inside a systemwide average. Both of these views are straightforward descriptive counts, not complex models.
Do I have to address all four case questions in my post?
Yes. The assignment lists four case questions as required components of the initial post: (1) why green averages misled leadership, (2) which descriptive views would have surfaced the crisis earliest, (3) what segmentation matters most and why, and (4) how descriptive analysis drives a 90-day recovery without predictive modeling. All four need to appear in your 500-word post. The trick is integrating them into a coherent memo rather than answering them as four separate bullet-point responses. The strongest posts weave these into the artifact discussion rather than treating them as a separate section.
Can I use outside sources to support this discussion post?
The assignment doesn’t require external citations, but using one strengthens your post. The Target Canada failure is well-documented in business press. A credible external anchor: the Harvard Business Review analysis of Target Canada’s failure discusses the data quality and inventory visibility failures in detail. You can reference it without long quotation — a one-sentence cite to anchor your artifact choices in the documented root causes is enough. Check whether your assignment format requires APA citations for discussion posts before adding a reference list.
What’s the difference between a descriptive and a predictive approach to the 90-day plan?
A predictive approach would build a forecast model: “based on current inventory trends, these stores will run out of stock in X days.” A descriptive approach asks: “where are inventory levels worst right now?” The predictive version tells you what will happen. The descriptive version tells you what is happening — and that’s all you need to prioritize a recovery. The 90-day plan powered by descriptive analysis is essentially: rank current shelf availability by store and category (worst to best), deploy resources to the worst performers first, track progress using the same descriptive view as a daily recovery indicator. No algorithm required — just the right disaggregation of current-state data.
Takeaway

What This Assignment Is Really Testing

This isn’t a data science assignment. It’s an operations management assignment about what information a leader needs to see — and at what level of granularity — to manage a real operational crisis. The Target Canada failure happened not because data didn’t exist, but because the data was being looked at in the wrong way: aggregated to the point of meaninglessness, presented as averages that hid the distribution of actual performance across stores and categories.

Your post is asking you to demonstrate that you understand the difference between a metric that looks good at the system level and visibility that supports actual decisions at the floor level. Pick artifacts that connect directly to the case’s documented failure points. Explain the green averages problem with specificity — not just “averages hide outliers” but exactly which variables were being aggregated in a way that masked the real problem. Make your segmentation argument clearly. And frame the 90-day plan as a prioritized action sequence, not a vague call for “better data.”

If you need support writing, reviewing, or formatting this post — including the reply component — the team at Smart Academic Writing provides discussion post writing services, case study writing, and editing and proofreading across operations management and business courses.