Of course it didn’t.
by Ken Howell
AI wasn’t part of the conversation when Target launched in Canada in 2013. ChatGPT was still years away, and very few organizations were thinking about artificial intelligence in any practical way. Yet thirteen years later, I think the Target Canada story has become one of the best business lessons for organizations rushing to embrace AI.
Canada Day has become a bit of a tradition for me. I usually end up outside tackling another project around the house, but I also carve out time to listen to Terry O’Reilly’s Under the Influence. Last year, one of those Canada Day mornings inspired me to write The Grass Can Wait. This year, another episode took me in a completely different direction.
I’ve written before about why I enjoy listening to Terry O’Reilly. In my article Silent = Listen, I talked about his ability to tell stories that stay with you long after the episode ends. This one did exactly that.
As Terry reflected on the rise and fall of Target Canada, I found myself thinking back to 2013 and just how much excitement surrounded its arrival.
Here in Winnipeg, it wasn’t just another retailer opening another store. Target was going to transform one of the city’s best-known pieces of real estate. The old Winnipeg Stadium site beside Polo Park was being redeveloped into what many expected would become an incredible shopping destination, adding to what was already Winnipeg’s premier retail district.
Like many Canadians, I’d shopped at Target whenever I travelled to the United States. I enjoyed wandering the aisles, but I always made a stop at the clearance end caps. Somehow the bargains there were even better than the bargains throughout the rest of the store. Then came the announcement that Target was taking over many of the old Zellers locations. That made the launch feel even more familiar. Many of us had memories of shopping at Zellers, buying toys for our kids, or coming home with a Zeddy teddy bear. Everything about the launch suggested Target understood Canadian shoppers.
The stores even looked familiar. The colours weren’t dramatically different from Zellers, but everything looked brighter, cleaner, and more modern. The promise was compelling.
Then the doors opened.
When the Experience Doesn’t Match the Promise
The stores were beautiful, but the shelves weren’t. That image has stayed with me ever since because it seemed impossible that one of North America’s most successful retailers couldn’t stock its own stores. Customers hadn’t lost confidence in the Target brand. They lost confidence because the experience didn’t match the promise.
For years, I thought it was a supply chain problem. Later, I assumed it was an implementation problem. Then I read The Last Days of Target Canada in Canadian Business, one of the best business post-mortems I’ve ever come across. It completely changed my perspective.
It wasn’t really about trucks. It wasn’t really about software. It was about data.
Inventory sat in distribution centres while customers stared at empty shelves. Product records didn’t accurately reflect reality. Information wasn’t complete, wasn’t consistent, and wasn’t flowing through the business the way it needed to. Add an aggressive implementation schedule and the assumption that Canadian customers would behave exactly like American customers, and the technology never really had a chance.
After more than three decades working in technology consulting, I’ve learned that when major implementations struggle, the technology itself is rarely the root cause. More often, it’s the quality of the information underneath it and the business processes built around it.
That’s why the Target story feels so relevant today.
Why This Matters Today?
Every executive is hearing about AI. Every boardroom is discussing it. Every software vendor promises it will change the way we work. Many of those promises are real, but I sometimes wonder whether we’ll be talking about some AI initiatives five years from now the same way we talk about Target Canada today.
AI doesn’t create data problems. It reveals them. Often faster than any technology we’ve implemented before.
If your customer information is inconsistent, AI will expose it. If your product data is incomplete, AI will expose that too. If years of business decisions have left different systems telling different versions of the truth, AI won’t magically reconcile them. It will simply make those inconsistencies impossible to ignore.
Could today’s AI have helped Target Canada? I think it probably could have accelerated some of the enormous data cleansing effort. It might have identified inconsistencies much sooner. What it couldn’t have done was decide which information was correct, redesign broken business processes, or challenge leadership assumptions about entering a new market.
Those weren’t technology problems. They were business problems. Rooted in data, business processes, and decisions that technology could never make on its own.
Looking back, I don’t think Target Canada failed because it lacked technology. It failed because it couldn’t trust the information its technology depended on.
Thirteen years later, AI hasn’t changed that lesson. If anything, it’s made it even more important.
Every organization wants to move faster with AI. My advice is simple. Don’t start with AI. Start with the data that AI will depend on.
If you’re exploring AI within your organization, we recently published a companion thought leadership article at Paradigm, From AI Hype to Execution: Introducing Paradigm’s AI Strategy and Data Readiness Offerings, which expands on why successful AI initiatives begin with trusted data long before the first AI tool is deployed.
References: