Transkrip penuh untuk Shallow DeepDive #18: The Coke rebrand is a rulebook, not a logo

was that the brand didn't look the same everywhere. 00:01:56.743 --> 00:02:01.616 Two hundred markets, hundreds of bottlers, thousands of local agencies, every one of 00:02:01.696 --> 00:02:06.215 them making something in Coke's name this afternoon, each one interpreting the 00:02:06.295 --> 00:02:10.755 guideline slightly, each one under deadline, each one granting itself a small 00:02:10.835 --> 00:02:11.551 exception. 00:02:11.631 --> 00:02:12.346 That's drift. 00:02:12.426 --> 00:02:13.526 It's never dramatic. 00:02:13.606 --> 00:02:18.361 It's a thousand small deviations, none of which matter alone, that together move a 00:02:18.441 --> 00:02:20.070 brand somewhere nobody chose. 00:02:20.150 --> 00:02:22.783 I watched this happen at a much smaller scale. 00:02:22.863 --> 00:02:27.559 When my old ad agency back in New York grew from two of us to a team of more than 00:02:27.639 --> 00:02:30.271 twenty, the first thing I lost wasn't quality. 00:02:30.351 --> 00:02:31.214 It was sameness. 00:02:31.294 --> 00:02:34.693 Every designer's hand started showing in the client's work. 00:02:34.773 --> 00:02:39.351 The blue crept brighter on one account because a version had performed well and 00:02:39.431 --> 00:02:40.884 everyone copied that file. 00:02:40.964 --> 00:02:45.365 Headings drifted into title case on another because a new hire preferred it. 00:02:45.445 --> 00:02:47.075 No single decision was wrong. 00:02:47.155 --> 00:02:51.674 The accumulation was, and by the time I noticed, the deviations had become the 00:02:51.754 --> 00:02:52.676 working standard. 00:02:52.756 --> 00:02:56.214 Fixing it meant a disruptive reset instead of a quiet nudge. 00:02:56.294 --> 00:03:00.282 That's the problem Coke has, multiplied by every country it sells in. 00:03:00.362 --> 00:03:02.582 So the real rebrand isn't the identity. 00:03:02.662 --> 00:03:04.291 It's what sits underneath it. 00:03:04.371 --> 00:03:08.890 Alongside the new look, Coke built a new Brand Center and what it calls Design 00:03:08.970 --> 00:03:13.136 Intelligence tools, developed with Adobe under the name Project Fizzion. 00:03:13.216 --> 00:03:17.440 The reporting we've seen describes it as the brand guidelines turned into 00:03:17.520 --> 00:03:22.275 machine-readable rules that live inside Illustrator and Photoshop, so the palette, 00:03:22.355 --> 00:03:27.109 the ribbon placement, the proportions of the square, get checked while the work is 00:03:27.189 --> 00:03:31.532 being made, and flagged when it's off, before anyone sends a file to print. 00:03:31.612 --> 00:03:36.190 Reportedly it replaces sign-offs that used to take days with a check that takes 00:03:36.270 --> 00:03:36.985 seconds. 00:03:37.065 --> 00:03:39.992 We'd treat the speed claims with the usual caution. 00:03:40.072 --> 00:03:43.707 The shape of it we believe, because it's the obvious next step. 00:03:43.787 --> 00:03:44.768 That step is this. 00:03:44.848 --> 00:03:48.188 For a hundred years a brand guideline has been a document. 00:03:48.268 --> 00:03:48.983 A PDF. 00:03:49.063 --> 00:03:53.936 A beautifully produced book that flatters the agency, reassures the client, and gets 00:03:54.016 --> 00:03:58.889 opened once by the person who actually produces most of the brand's output, on 00:03:58.969 --> 00:04:00.481 first day, and never again. 00:04:00.561 --> 00:04:03.134 The guideline only worked if someone read it. 00:04:03.214 --> 00:04:07.969 Coke has moved the rules from the document into the tool, so the guideline doesn't 00:04:08.049 --> 00:04:09.797 have to be read to be followed. 00:04:09.877 --> 00:04:10.858 That's the change. 00:04:10.938 --> 00:04:15.693 The most recognised brand in the world admitted that its rulebook wasn't 00:04:15.773 --> 00:04:18.877 anything, and fixed the rulebook rather than the logo. 00:04:18.957 --> 00:04:23.830 We said something close to this back in episode ten, when we argued brand 00:04:23.910 --> 00:04:28.664 need an AI chapter, and in episode eight, when we argued a brand has to be legible 00:04:28.744 --> 00:04:30.433 to machines as well as people. 00:04:30.513 --> 00:04:33.853 Coke has gone one further and made the machine the reader. 00:04:33.933 --> 00:04:38.039 A rule that only works if a tired person remembers it isn't governance. 00:04:38.119 --> 00:04:38.835 It's hope. 00:04:38.915 --> 00:04:42.667 A rule that runs in the background while they work is governance. 00:04:42.747 --> 00:04:45.792 Now the caution, because there is one, and it's real. 00:04:45.872 --> 00:04:47.325 A machine can hold a rule. 00:04:47.405 --> 00:04:48.740 It can't hold judgement. 00:04:48.820 --> 00:04:51.158 It can tell you the red is the wrong red. 00:04:51.238 --> 00:04:56.052 It can't tell you the ribbon shouldn't be there at all this time, or that the local 00:04:56.132 --> 00:05:00.415 market's Ramadan pack needs something the global system never anticipated. 00:05:00.495 --> 00:05:05.191 Encode the rules too tightly and you don't get a coherent brand, you get a frozen 00:05:05.271 --> 00:05:10.025 one, with every happy accident that keeps an identity alive filtered out before it 00:05:10.105 --> 00:05:10.821 happens. 00:05:10.901 --> 00:05:14.948 The thing worth noticing in Coke's version is where the judgement sat. 00:05:15.028 --> 00:05:19.783 People went through a hundred and forty years of archive and decided which handful 00:05:19.863 --> 00:05:21.670 of elements were non-negotiable. 00:05:21.750 --> 00:05:23.911 Script, ribbon, square, red and white. 00:05:23.991 --> 00:05:24.706 Four things. 00:05:24.786 --> 00:05:29.541 That decision was made in a room, by designers, and it's the actual design work 00:05:29.621 --> 00:05:30.337 this project. 00:05:30.417 --> 00:05:33.285 The software just stops those four things leaking. 00:05:33.365 --> 00:05:37.884 Get the list wrong and the software enforces the wrong brand very efficiently. 00:05:37.964 --> 00:05:42.247 Which brings it down to a scale that matters to anyone listening who isn't 00:05:42.327 --> 00:05:43.043 Coca-Cola. 00:05:43.123 --> 00:05:46.167 You don't need Adobe to build you a compliance layer. 00:05:46.247 --> 00:05:48.998 The lesson is smaller and more useful than that. 00:05:49.078 --> 00:05:53.656 Decide what's genuinely non-negotiable in your brand, and keep that list short, 00:05:53.736 --> 00:05:57.724 because everything on it is something you're about to defend forever. 00:05:57.804 --> 00:06:02.559 Write those few things so plainly that a tired person at nine at night, or a model 00:06:02.639 --> 00:06:05.625 given the file, can apply them without a phone call. 00:06:05.705 --> 00:06:10.519 Then name one person whose job is to notice drift, and put a review on the 00:06:10.599 --> 00:06:14.823 before the project closes, not after the first time something goes wrong. 00:06:14.903 --> 00:06:15.884 That's governance. 00:06:15.964 --> 00:06:16.945 It fits on a page. 00:06:17.025 --> 00:06:21.191 Most brands we meet have the two-hundred-page book and none of the 00:06:21.271 --> 00:06:24.728 The last thing worth sitting with is the direction Coke ran. 00:06:24.808 --> 00:06:29.681 Most rebrands start from the assumption that the problem is the look, and end with 00:06:29.761 --> 00:06:33.396 new logo, a launch film and the same drift twelve months later. 00:06:33.476 --> 00:06:37.994 Coke looked at itself, decided the logo was fine, and concluded the discipline 00:06:38.074 --> 00:06:38.790 wasn't. 00:06:38.870 --> 00:06:43.743 It's the least glamorous conclusion a brand can reach about itself, and it's the 00:06:43.823 --> 00:06:44.804 most of them need. 00:06:44.884 --> 00:06:48.047 Thank you for indulging us in today's Shallow DeepDive. 00:06:48.127 --> 00:06:50.406 We hope you'll join us for the next one. 00:06:50.486 --> 00:06:51.703 Until then, take care.