Rebranding Without Losing Recognition: Mastercard, Burberry and Dropbox

Rebranding has a poor reputation because most public arguments about it are about taste. The documented cases are more useful than the arguments. The transitions below keep a recognition asset and change the system around it, and each one left a number, a date or a research method that a committee can actually argue with.

Mastercard's 2016 identity, by Pentagram partners Michael Bierut and Luke Hayman, was the brand's first change in 20 years. It replaced the stripes in the overlap with a solid orange, moved the wordmark outside the symbol and set it in lowercase. The colour shift was technical rather than decorative: the new mark uses an additive colour mix so the overlap reads brighter on screens, where the original subtractive mix darkened.

At the time, 2.3 billion Mastercard cards were in circulation carrying the previous mark. That figure is the scale problem in one line, and it is also the reason the wordmark did not disappear in the same project. The famous removal came three years later, once the brand could prove what the new symbol was doing on its own.

The cases below are not three styles. They are three mechanics, and each one answers a different question a board will ask before approving the work.

The two Mastercard moves were three years apart, and that gap is the most transferable thing in this article. At CES on 10 January 2019 the company announced the wordmark would be dropped entirely, citing its own research that more than three-quarters of consumers could identify the brand from the circles alone. The symbol had already been in market since 2016, measured and reported, when the name came off.

Burberry is the counter-example, and its designer said so himself. The August 2018 identity, by Peter Saville with chief creative officer Riccardo Tisci, replaced a logotype used since 1901 and the Equestrian Knight device with a new sans-serif logotype and a monogram, announced through the brand's Instagram. Daniel Lee, chief creative officer from 2022, reintroduced the archive Equestrian Knight logotype in February 2023. Saville's 2025 assessment was that the churn itself was the mistake, and that at any point over the last two years you could have found at least three different Burberrys circulating in the world.

The difference between the two stories is not talent. One team treated the change as a sequence of measured steps; the other treated it as a series of statements, and the asset the audience had learned was replaced before anything had been proved about its replacement.

Research before the mark is what separates a rebrand from a redesign. Airbnb's Bélo, launched in July 2014 and designed by DesignStudio, followed roughly a year of work that included one intensive week in which four team members travelled to 13 cities, stayed with 18 hosts across four continents and interviewed more than 120 people. The stated goal was a marque anyone could draw, which is a recognition target stated before the drawings existed.

Instagram's team ran the cheapest version of the same test in an office. Employees were asked to draw the icon from memory in five seconds, and almost everyone drew the rainbow, the lens and the viewfinder, so those elements were carried into the new icon and the interface was flattened around them. The test cost almost nothing and answered the only question that mattered: what is the brand when the person cannot look it up.

Mastercard's numbers arrived after launch rather than before it, which is the other legitimate order. Pentagram records 81 per cent spontaneous recognition of the new symbol without the name, and the 2019 wordmark decision rested on the company's own finding that more than three-quarters of consumers could identify the brand from the circles alone. A committee that skips this stage is choosing to find out in public.

Dates matter in this subject, because half the myths about rebranding come from compressing two decisions into one. The table records what changed, what was kept and what the record actually shows.

Read the first two rows together and the most quoted misunderstanding in this field disappears: the 2016 redesign and the 2019 removal were separate decisions, made three years apart, with a measurement in between.

Dropbox's 2017 system, built by Collins with the in-house Dropbox Brand Studio and specialist studios, is the clearest case of a change that was never really about a mark. Collins describes the rebuilt brand as organised around syncing teams rather than syncing files, with Sharp Grotesk typography, an isometric diamond and a Flex Deck layout logic. That sentence is the agency's own account of its work, which is exactly how a case study should be read.

The same distinction appears at a smaller scale in Mailchimp's September 2018 refresh, also by Collins. The recognition system stayed, the expression changed, and the commercial story underneath was consolidation: 35 solutions reduced to five portfolio categories. Neither project needed a different name to become a different brand, and neither asked the audience to relearn what the company was for.

A committee can use this as a test. If the proposed change alters what the company is for, it is a positioning project wearing a design costume, and it needs the research and the sequencing that a positioning project needs. If it alters how the company looks while the promise stays put, it is a refresh, and it should be priced and scheduled as one.

A rebrand is argued about in decks and decided at this distance: on a shelf, in a concourse, at a rack. The three photographs below are not illustrations of the cases above. They are the conditions a new identity has to survive before any case study about it is worth reading.

Burberry's reversal deserves better than being filed as a failure. The archive Equestrian Knight logotype came back in February 2023 because it still carried recognition that a monogram could not buy, and Daniel Lee was hired to change the company's direction. The expensive part was not the 2018 identity; it was the churn around it.

A rollout plan is where recognition is actually kept or lost. Mastercard's case shows the constraint in one figure: 2.3 billion cards carried the previous mark when the 2016 identity launched, and the wordmark came off three years later. No launch calendar can flip an asset that lives in that many wallets at once, which is why phasing is the professional answer to impatience.

There is also a cheaper option that committees forget: change less, test more. Instagram kept three elements that memory testing surfaced and rebuilt everything else around them, at a cost far below the price of relearning. The audience never had to decide whether the brand was still itself.

1. Name the recognition asset that survives before anyone touches a mark.Instagram kept the lens, viewfinder and rainbow after five-second memory drawings; Mailchimp kept its recognition system and changed its expression.
2. Put a measurement between the redesign and the removal.Mastercard dropped its wordmark at CES on 10 January 2019, three years after the 2016 identity, citing more than three-quarters of consumers identifying the brand from the circles alone.
3. Pay for the research before the launch, not after it.Airbnb's Bélo took about a year and included 13 cities, 18 hosts, four continents and more than 120 interviews.
4. Read agency case studies as claims.Collins' Dropbox account is the agency's own narrative, and its causal attributions are not findings.
5. Phase the rollout, because recognition lives in the installed base.2.3 billion Mastercard cards carried the previous mark when the 2016 identity launched, which is why the name came off in a later project.
6. Do not let a reversal become churn.Burberry's 1901 logotype and the Equestrian Knight were replaced in 2018, the knight returned in February 2023, and the designer called the intervening churn irresponsible.