Table of Contents
Target opened 133 stores across Canada in a matter of weeks in 2013. By April 2015 every one of them was gone, 17,600 people were out of work, and the parent company had written off more than $5 billion before tax. The plan had been built by clever people with serious data behind it. It failed because the whole thing rested on one untested assumption: that running a Canadian big-box chain is the same job as running an American one.
We tend to study winners. Turnaround heroes get the keynotes and the book deals, which is exactly the bias explored in what business case studies tend to leave out. Losers get a paragraph and a shrug. That is a shame, because the expensive mistakes repeat themselves with almost comic regularity.
Here are seven of the most costly ones, with the numbers attached.
1. Entering a market because the spreadsheet looked good
Target paid $1.8 billion for 220 Zellers leases in 2011, then tried to open more than 120 stores in under a year. The supply chain system was never properly configured. Shelves sat empty for months, prices scanned wrong, and shoppers who had crossed the border to visit Target stores in the US walked into something that felt like a discount warehouse with better lighting.
By the time management had the data to prove the model worked, they had already burned the capital and the goodwill.
The tell
Nobody running the Canadian operation had ever opened a store in Canada. Depth of local knowledge is expensive and slow to buy. Skipping it is far more expensive.
2. Protecting the cash cow instead of replacing it
Kodak’s own engineer, Steven Sasson, built the first digital camera in 1975. Kodak had the patents, the chemistry talent and 145,000 employees at its 1988 peak. It filed for bankruptcy in January 2012 anyway, because every internal argument ended the same way: digital would eat film margins.
Blockbuster turned down Netflix for $50 million in 2000. The reasoning was sound at the time and utterly irrelevant within a decade. The company went bankrupt in 2010.
Cannibalising your own product hurts for four quarters. Failing to do it hurts permanently.
3. Paying a premium for a story about synergy
The AOL–Time Warner merger was announced in January 2000 at roughly $165 billion. Two years later the company took a writedown of almost $99 billion. The promised magic, where internet distribution would somehow supercharge magazine subscriptions, never materialised in any measurable form.
Daimler paid $36 billion for Chrysler in 1998 and spent nine years undoing it. In both cases the acquirer paid a control premium for a narrative its own analysts could not model.
When the deal thesis needs a slide titled “cross-selling opportunities”, treat the price tag as a donation.
4. Cutting the line item that keeps you alive
Cost reduction is the easiest decision a board can approve, which is why it gets approved in the wrong places.
- Marketing. Kellogg’s kept and roughly doubled its advertising through the Depression while Post cut back. Kellogg’s came out of the 1930s as the category leader, a position it held for decades.
- Maintenance. Deferred upkeep at BP’s Texas City refinery was cited in the 2005 explosion that killed 15 people. Five years later, Deepwater Horizon cost the company north of $65 billion in penalties, cleanup and settlements.
- Support staff. Airlines that strip call centres to the bone save millions and lose tens of millions in churn from customers who cannot get a human on the phone during a storm.
The useful test is simple: does this cut remove a cost, or does it move the cost somewhere less visible?
5. Borrowing to the edge of the cliff
In 2005 KKR, Bain and Vornado bought Toys R Us for $6.6 billion and loaded about $5 billion of debt onto the retailer itself. Roughly $400 million a year went out the door in interest before a single toy was sold. The chain had a viable business and an unviable capital structure.
Chapter 11 came in September 2017. Liquidation followed in 2018, closing 700-plus stores and ending about 33,000 jobs. Thames Water in the UK has run a similar experiment with a debt load above £14 billion, and the bill is landing on customers and taxpayers.
Every one of those balance sheets was legible years in advance to anyone who bothered to read it. That is precisely the trade vulture investors build their returns on.
6. Tolerating a culture that charges you in instalments
Wells Fargo’s fake-accounts scandal broke in 2016. Sales targets had been pushing staff to open accounts customers never asked for since at least 2002, and the early warnings were dismissed as the complaints of a few underperformers. The final bill included a $3 billion settlement with the US Department of Justice and a Federal Reserve cap on asset growth that lasted years.
Uber’s 2017 crisis followed the same script. A blog post from a former engineer catalysed a year of investigations, a CEO exit, a lawsuit from its own largest shareholder and a rewrite of the company’s values.
Culture problems rarely appear as one big loss. They show up as a hundred small, rational-looking decisions that each shift a little more risk onto someone who cannot say no.
7. Refusing to kill a loser on schedule
Quibi raised $1.75 billion, launched in April 2020 and announced its shutdown that October. To its credit, the exit was fast. Plenty of companies are not so lucky. Google kept Google+ alive for eight years after it was clear the product had lost, and New Coke took 79 days to reverse only because the backlash was loud enough to be undeniable.
Sunk costs are the most expensive thing in most board packs. The money is already gone. The only real question is whether the next $10 million is a good use of capital, and teams that built the original plan are the worst people to answer it.
Write the kill criteria before you launch
Pick the number now. “If we have not hit 40,000 paying users by month 18 at a payback under 14 months, we stop.” A pre-agreed threshold turns an emotional argument into an arithmetic one.
Decision hygiene that costs almost nothing to run
None of the seven failures above came from stupidity. They came from decisions made quickly, defended publicly, and never revisited. Three habits raise the odds considerably.
Name the assumption that has to be true. Target Canada needed to prove that Canadian consumers would switch in volume at those prices. That was a testable claim, and it was never tested before the leases were signed.
Run a pre-mortem with the people who will do the work. Ask the team to write the post-mortem as if the project has already failed, then ask what caused it. Frontline staff usually know where the plan breaks; they are simply not asked.
Separate the decision from the person. Reversing course is cheap in month three and career-ending in month thirty. Build a review date into the decision itself, with a named owner and a fixed agenda item, so that changing your mind looks like governance rather than weakness.
The companies that survive their own bad bets are rarely the ones that made fewer of them. They are the ones that found out early, wrote off the loss without a cover-up, and spent the remaining capital on something that worked.
