Seventy-Nine Days: on the recorded mistakes of rebrandingSeventy-Nine Days: on the recorded mistakes of rebranding

On April 23, 1985, Coca-Cola changed its ninety-nine-year-old formula. The decision had not been taken on a whim: over two years the company had run blind taste tests with roughly 200,000 people, and the new formula had beaten both the old Coca-Cola and Pepsi. Even so, the phone lines jammed, protests began, and on July 11 the old formula returned under the name "Coca-Cola Classic." Between the two dates, 79 days had passed. The tests had measured taste; nobody had measured how people would feel about the change itself. Every mistake in rebranding is a variation on that sentence, and nearly all of them are documented.

In January 2009, Tropicana switched to new packaging designed by the Arnell Group: the orange with a straw stuck in it was gone, replaced by a glass of orange juice. Sales fell by about 20 percent in two months, the loss reached $30 million, and on February 23 the old packaging came back. In October 2010, Gap replaced its twenty-year-old serif logo with Helvetica and a small blue square; the reaction was so harsh that the company first tried to "crowdsource" the design, then returned to the old logo six days later. In 2001, the British postal service spent about two million pounds to rename Royal Mail "Consignia"; sixteen months later it took the name back. In 2002, PwC's consulting arm named itself "Monday"; within a few months it was acquired by IBM and the name disappeared.

What the four events have in common is not a lack of talent on the designers' part. In all four, the company changed an asset the customer recognized and regarded as their own, without measuring that recognition. Tropicana's orange was recognizable on the shelf from ten meters away; the glass was not. What changed was not beauty but findability.

In 1979, Daniel Kahneman and Amos Tversky showed with prospect theory that people are roughly twice as sensitive to losses as to gains; in 1980, Richard Thaler named the overvaluation of what one already owns the "endowment effect." The customer regards the logo they are used to as their own; when the logo changes, the "modernity" gained is small and the sense of "my brand" lost is large. In 2012, Google's user research team gave this reaction a name: change aversion. The drop in satisfaction they measured at every redesign of a product occurred regardless of the quality of the design and faded with time; Zajonc's mere exposure effect was running in reverse, then re-establishing itself.

In 2010, Michael Walsh, Karen Winterich, and Vikas Mittal measured who reacts most harshly: the customers most committed to a brand responded most negatively to a logo change, while those with a weak relationship to the brand received the change more favorably. Rebranding is the job that most disturbs a brand's most valuable customers; that is why it is measured with them above all.

In 1973, the economist Michael Spence wrote signaling theory: a costly action proves to the other party a quality that cannot be seen. Rebranding is such a signal; it says "we have changed." The signal is credible if there is real change behind it, and empty if there is not. Gap's logo had put new clothes on a company in which nothing had changed; the audience recognized the empty signal at once. In 2000, BP spent about £136 million, implementation included, on the green-and-yellow sun designed by Landor and the slogan "Beyond Petroleum"; the signal was strong, but after the spill in the Gulf of Mexico in 2010 the same sun turned into the emblem of a promise not kept. Where it cannot be proven, the signal works against the brand.

Jenni Romaniuk's Building Distinctive Brand Assets, published in 2018, provides the measuring instrument for rebranding: every visual asset is scored on two axes, fame (how many people link the asset to the brand) and uniqueness (how many people link it to this brand alone). A high-high asset is untouchable; a low-low asset can change freely. Mastercard acted on this measure in 2019: when its research showed that more than 80 percent of people recognized the two circles without the name, it removed the word from the logo. When Johnson & Johnson dropped its 136-year-old script logo in 2023, it asked the same question in reverse and was criticized; Burberry discarded its knight in 2018 for a plain sans-serif logo and brought the knight back in 2023; Pepsi, in 2023, reverted the million-dollar new logo of 2008 to the bold lettering of the 1990s.

In the "simplification" wave of the 2010s, dozens of brands converted their serif, characterful marks into look-alike geometric sans-serifs and voluntarily lowered their uniqueness scores. The reason for the reversals is not nostalgia but the corner lost on Romaniuk's grid.

The dose of change has also been studied. In 2013, Bettina Müller, Bruno Kocher, and Antoine Crettaz tested logo redesigns in the Journal of Business Research: a measured evolution made the brand look more contemporary and preserved attitude, while a radical break produced loss among committed customers. Instagram's icon change in 2016 taught the second half of the lesson: the flat-colored icon that was mocked in the first weeks was accepted within a few months through Zajonc's mechanism, because the app was opened every day. A brand with a high exposure budget can digest radical change; an industrial brand seen twice a year cannot. The dose is set by how often the brand is seen.

A new identity arrives not in a day but with a program. First inward: employees learn the new identity before the customer does, because they are the ones who answer the first question. Then the dual period: the old and new marks live side by side for a set time; Mastercard's move to the nameless circles was spread over years. Then the digital trail: domain redirects, search engine records, map and directory listings, social accounts; when one is forgotten, the brand splits in two in the search results. Then the law: the old registration is not abandoned, and the new registration is secured before the new name goes public. Tropicana's and Gap's quick reversals teach one more lesson: the rollback plan is part of the rebranding plan.

New Coke sank with 200,000 tastings because it had measured the wrong thing. The right measurement has three stages: before the change, unaided recall and the Romaniuk grid; with the change, predetermined thresholds (which drop in recognition is acceptable, which one triggers a rollback); after the change, the same measures repeated at three and twelve months. If the threshold is not written down in advance, every drop is explained away with "they'll get used to it," and Tropicana's two months turn into two years.

Our protocol begins not with drawing but with an inventory: the existing mark, color, form, typography, and voice are scored one by one on the axes of fame and uniqueness; a high-scoring asset is kept and becomes the core of the new system. The second step is the signal question: what has actually changed in the company, and what will the identity prove; if there is no answer, we do not recommend a rebrand; we recommend a refresh. The third step is the dose: how often the brand is seen, together with Müller's finding on measured evolution, determines the breadth of the change. The fourth step is the transition program: internal launch, dual period, digital trail, and registration. The last step is measurement; the thresholds are written before the launch.

New Coke's 79 days are counted as the most expensive lesson in marketing history. The core of the lesson is simple: people take ownership not of the product but of their habits around the product; rebranding is an intervention made with respect for that habit.

Coca-Cola Company, New Coke chronology (April 23 – July 11, 1985; about 200,000 taste tests).

Tropicana packaging change (January–February 2009); Gap logo change (October 4–11, 2010); Royal Mail / Consignia (2001–2002); PwC Consulting "Monday" (2002).

Kahneman, D. & Tversky, A. (1979). Prospect theory: An analysis of decision under risk. Econometrica, 47(2), 263–291; Thaler, R. (1980). Toward a positive theory of consumer choice. Journal of Economic Behavior & Organization, 1(1), 39–60.

Sedley, A. (2012). Change aversion: Why users hate what you launched (and what to do about it). Google UX research.

Walsh, M. F., Winterich, K. P. & Mittal, V. (2010). Do logo redesigns help or hurt your brand? The role of brand commitment. Journal of Product & Brand Management, 19(2), 76–84.

Spence, M. (1973). Job market signaling. Quarterly Journal of Economics, 87(3), 355–374.

BP "Helios" identity (Landor, 2000) and reported cost (about £136 million).

Romaniuk, J. (2018). Building Distinctive Brand Assets. Oxford University Press; Sharp, B. (2010). How Brands Grow.

Mastercard (2019) nameless logo announcement; Johnson & Johnson (2023); Burberry (2018, 2023); Pepsi (2008, 2023); Instagram (2016).

Müller, B., Kocher, B. & Crettaz, A. (2013). The effects of visual rebranding on brand attitude. Journal of Business Research, 66(1), 82–88.

Zajonc, R. B. (1968). Attitudinal effects of mere exposure. Journal of Personality and Social Psychology, 9(2, Pt. 2), 1–27.