12 CPG Rebrands So Bad They Tanked Sales

A rebrand is supposed to be an upgrade. These are the CPG rebrand failures that turned into downgrades so severe the companies reversed them in weeks, sometimes at a cost of tens of millions of dollars. What links the worst of them is not bad taste, it is a misread of how a shopper actually finds a product on a crowded shelf, and how fast the register tells you when they cannot.

A logo or a package is not decoration. It is the search key a shopper uses to locate you in under two seconds, and the moment you change it you are betting that recognition survives the swap. When the bet is wrong, it does not fail slowly. It fails at the speed of the next shopping trip, which means it fails in your weekly scan data long before it reaches a quarterly review.

The 12 biggest CPG rebrand failures

1. Tropicana (2009)

The patron saint of the rebrand disaster. PepsiCo spent a reported $35 million replacing the orange-with-a-straw with an abstract glass of juice, and unit sales fell about 20 percent in under two months, roughly $30 million, because shoppers could no longer find the orange juice sitting right in front of them. The old carton was back within about two months. Distinctive assets are load-bearing walls, and this is what it looks like to take a sledgehammer to one.

2. Gap (2010)

The new logo lived six days. Six. The company asked the internet to submit alternative designs mid-backlash, which is the branding equivalent of asking the crowd at your intervention for drink recommendations. Reverted within the week.

3. Netflix Qwikster (2011)

An honorary non-CPG inductee that belongs here on principle. Netflix split off its DVD business, named it like a 2004 energy drink, and killed it in three weeks, before it ever launched. The rare rebrand that died in utero and still moved the stock.

4. Coca-Cola white holiday cans (2011)

Polar-bear cans that looked so much like Diet Coke that shoppers grabbed the wrong one and some swore the formula had changed, which it had not. Pulled within a month. People taste packaging before they taste the product.

5. Pepsi logo (2008)

Sales survived. Dignity did not. A leaked 27-page agency document explained the new logo through the golden ratio, the Earth's magnetic field, and the Mona Lisa. Inducted for the document alone.

6. Kraft (2009)

Traded one of the most recognized logos in food for a burst-and-swoosh that one critic called a logo for a company that does everything and nothing. Reverted within months. The company later solved the identity question by splitting itself in two.

7. RadioShack, 'The Shack' (2009)

Honorary inductee. A mid-life crisis rendered in signage. Sales kept falling, but now casually, on a first-name basis.

8. SyFy (2009)

Renamed so the brand could be trademarked, then announced as being for people who are less geeky. The audience, famously geeky, noticed immediately. Survived, but earned its seat here.

9. WW (2018)

Weight Watchers went wellness, and members could no longer say out loud what the company actually did. Enrollment noticed before the brand team did.

10. Ayds (1980s)

The reverse case: the diet candy that desperately needed a rename and waited too long, as sales reportedly fell by half. Sometimes the worst rebrand is the one you refuse to do.

11. New Coke (1985)

Technically a reformulation, spiritually the first rebrand fail, reversed in 79 days. Included out of respect, because it has tenure.

12. Consignia (2001)

Honorary international inductee. Britain's Royal Mail paid roughly two million pounds to rename itself into something that sounds like a shell company in a tax thriller. Reverted in 16 months.

The damage at a glance

What the shelf data says before the boardroom does

Line these up and they stop looking like a dozen unrelated blunders. They look like one avoidable mistake: a brand changed its search key and found out too late that recognition did not come along. The tell is always in the same place, sales velocity, measured as units per store per week, which is the number that strips out how many doors you are in and shows how hard each one is working.

Take the Tropicana pattern. A carton that tests fine in a lab still has to win at the shelf, where the shopper spends two seconds and no conscious thought. A brand running 12 units per store per week does not glide down to 10 over a quarter. It drops to 9 or 10 inside the first two reporting periods, and the drop is visible in week two, not in the quarterly business review three months later. The teams that reverted fast were the ones watching sales velocity weekly. The teams that watched shipments saw nothing until the distributor stopped reordering.

The same signal catches the quieter failure modes. A package that blends into the competitor's, like the white Coke cans, shows up as on-shelf availability and share of shelf that look fine while sales sag, because the product is physically present and simply not being found. That gap between present and purchased is the single most useful diagnostic a brand has, and it is invisible to anyone reading only a shipment report.

How to de-risk your own redesign

You cannot A-B test a national rebrand the way you test a landing page, but you can refuse to fly blind. Roll the new look into a subset of stores first, hold the old look as a control, and read per-store velocity weekly against that control for six to eight weeks before you commit the whole chain. If the new key is losing shoppers, the store-week data will say so while the change is still cheap to reverse.

A redesign that is quietly costing you shows up first as a per-store velocity dip against the doors that kept the old look. Scout tracks that store by store, so you can reverse in days rather than find out a quarter and thirty million dollars later. See the launches that made the same mistake in the 15 worst CPG product fails of all time, then pilot your redesign against a control.

The rebrand nobody remembers, and why that is the point

The rebrands that work share one quiet trait: you cannot name them. When a brand refreshes its look without breaking the shopper's recognition, sales do not move, no case study gets written, and the brand team quietly keeps its jobs. Success in packaging is invisible by design, which is exactly why the failures are so instructive. Every disaster on this list is a controlled experiment in what happens when you cut the thread between a brand and the shopper's two-second glance.

The practical takeaway is not that redesigns are dangerous. It is that recognition is an asset with a measurable value, and you only learn that value by watching what happens to velocity when you spend it. A brand that treats its logo, its color block, and its shelf silhouette as disposable styling is spending an asset it never put on the balance sheet. A brand that treats them as equity, and pilots any change against a control, gets to modernize without setting money on fire.

There is a timing lesson buried in the reversals, too. The brands that recovered, Tropicana, Gap, and Coca-Cola among them, shared one thing: they were watching closely enough to reverse in weeks, not quarters. The cost of a bad rebrand is roughly linear in how long you let it run, so the speed of your data is the size of your insurance policy. A team reading weekly store-level numbers caps the damage at a rounding error. A team waiting for the quarterly review pays for the mistake in full.

One more discipline separates the survivors from the cautionary tales: they decide in advance what number would make them reverse. A rebrand with no pre-committed kill threshold becomes a sunk-cost argument the moment sales wobble, because someone has already staked their reputation on the new look. A rebrand with a written rule, roll back if per-store velocity falls more than a set percentage against control for three straight weeks, takes the ego out of the room and lets the data make the call while the reversal is still cheap.

It is worth saying plainly that most of these companies were not careless. They hired good agencies, ran real research, and still shipped a disaster, because pre-launch research measures what people say about a design in a conference room, and the shelf measures what their hands do in half a second. Those are not the same test, and only one of them shows up in your sales data. The lesson of every entry here is to trust the second test, and to build your process so you see its verdict in days, not months.