Turnarounds are never reversed. They are abandoned when the architect leaves
Forty-four turnarounds on record and not one was undone. The failures do not look like reversals, they look like unfinished plans under a new chief executive. Intel, Burberry, and what the completed ones did differently.
Zero out of forty-four
Every decision type in this record has some reversals. The Turnaround has none. Thirty-nine of the 44 verified records held, four were partial, and not one company undid its recovery plan.
That is not a 91% success rate, and reading it as one is the mistake this page exists to prevent. A turnaround is not reversed because there is nothing to reverse it to. Nobody goes back to the strategy that nearly killed the company. What happens instead is that the plan stops halfway, and the record captures that as partial.
What a stalled turnaround looks like
Intel decided in February 2021 to fund a from-scratch foundry out of the cash flow of a shrinking product business. Foundry Services did $952 million of revenue against a $482 million operating loss in 2023, and the product business fell to $53.1 billion in 2024, roughly a third down over three years. The record's verdict is precise: the strategy's architect was pushed out while the fabs kept rising, leaving the funding question unresolved for successors.
Burberry decided in 2006 to walk away from profitable low-end check volume and reclaim control of distribution. Revenue grew 102% and operating profit 82% over the tenure that followed, and the company's value went from about £2 billion to over £7 billion. The verdict notes what happened next: the underlying tension recurs, requiring a fresh recovery plan under every chief executive since.
Neither is a failure of diagnosis. Both are plans longer than the tenure of the person who started them.
The ones that finished
Domino's publicly self-criticized its own pizza and relaunched in 2009, and the relaunch was followed by years of same-store sales growth. Best Buy sold its floor space to vendors in 2012 and got the cost cuts, the online growth and the revenue recovery. Chipotle rebuilt throughput and digital from 2018 and digital went from a trickle to over half of sales. Abercrombie & Fitch ran an institutional buying-discipline rebuild from 2022 and beat its own plan two years early.
The completed ones are shorter than the stalled ones, and they are specific about the mechanism. Throughput. Buying discipline. Floor space. Not transformation, not repositioning: a named operating change that a successor would find it awkward to stop.
How to run this decision
Drawn from what separated the records that stuck from the ones that were undone. A step nobody in the record can be seen taking is not here.
- 01
Check the plan against the tenure, honestly
If the plan takes seven years and chief executives at your company last four, you are not designing a turnaround, you are designing the first half of one. That may still be right, and it changes what you should commit to and what you should leave reversible.
- 02
Name the mechanism, not the destination
Chipotle rebuilt throughput. Best Buy sold the floor to vendors. Those are operating changes with owners and numbers. Premiumization, transformation and repositioning are destinations, and a successor can redirect a destination without appearing to cancel anything.
- 03
Sequence it so the first result arrives inside eighteen months
Not because short-term results matter more, but because an unfinished plan with nothing to show is the one that gets quietly dropped at the handover. Domino's had visible same-store sales movement before the plan was complete.
- 04
Fund it from something other than the business you are fixing
Intel's foundry was to be funded by a product business that was itself shrinking, which made every bad quarter in the old business a question about the new one. Where that structure is unavoidable, say so out loud at the start and agree what happens when the funding source dips.
- 05
Write down what the next chief executive must not reopen
Burberry's recovery has been re-run under every chief executive since 2006, which means the 2006 decision was never converted into something structural. Put the parts that must survive a handover into contracts, ownership or capacity, not into a strategy document.
- 06
Declare it finished, in public, on a date
Turnarounds in this record do not end, they fade. An explicit end point converts the work into a baseline that the next plan has to start from, rather than a plan the next plan replaces.
What it looks like when this is going wrong
- The plan is longer than the average tenure of the role running it.Intel, 2021. Burberry, 2006.
- It is funded from the cash flow of the business it is meant to fix.Intel, 2021.
- The plan is named for a destination rather than an operating change.The four partial turnarounds on record.
- Each new chief executive announces a fresh recovery plan for the same problem.Burberry, every tenure since 2006.
- Nothing visible will have changed at the eighteen-month mark.Intel Foundry Services, 2021 to 2024.