Should we change our logo?

The Cadence Question Is the Wrong Question

Companies that ask "should we change our logo" are usually asking it as a yearly or quarterly review item. The answer the question deserves is structural, not periodic. The right question is which class of brand this is, because each class has a different set of rules and each class fails for a different reason.

Three classes, three rules. The remainder of this page reads them against five anchor cases and one squint test.

Class One. The Never Test — Recognition-Equity-Rich Marks.

Coca-Cola's script has carried the brand for more than a century. The company has adjusted the script in dozens of small ways across decades, none of which the public has ever been asked to react to. Mercedes-Benz, Ferrari, Levi's, Caterpillar, Tiffany sit in the same category. The mark is recognized everywhere the company sells and the recognition is itself the asset being protected.

The never test is one question. If the existing mark were stripped from the asset tomorrow, what fraction of customers would still recognize the brand on the shelf, on the page, in a half-second glance? If the answer is high, the right cadence is functionally never. Refresh the mark only as a hand-restoration of an old painting: invisible from a distance, careful in detail, no announcement.

The failure mode in this class is the loud refresh. Tropicana sat inside class one and the company treated it as if it were in class two or class three. The recognition equity disappeared the moment the announcement landed.

Class Two. The 5-Year Test — Mid-Recognition Marks.

Airbnb's Belo was introduced in 2014 and has held since. Slack's mark refreshed in 2019 and held. Mastercard refreshed in 2016 and held. Burger King refreshed in 2021. The pattern across these cases is a 5 to 10 year cadence of intentional refresh, paired with a product or experience evolution that the new mark is meant to reflect.

The 5-year test for this class is two questions. First, has the product or category around the mark changed materially since the last refresh? Second, does the existing mark feel constrained when applied to the surfaces the brand has added in that period (mobile interfaces, voice surfaces, AR surfaces, AI surfaces)? If both are yes, a refresh is structurally indicated. If only one is yes, the refresh can wait another cycle without cost.

The failure mode in this class is refresh for the sake of activity. A mid-recognition mark refreshed without a structural reason draws attention to a change that does not improve the brand. The cost is recognition equity decay without a corresponding gain in product clarity.

Class Three. The 12-Month Test — New or Low-Recognition Marks.

Companies less than two years old, or older companies whose mark has never gained meaningful recognition, sit in class three. The recognition equity available to lose is small. The cost of a change is mostly the design and asset replacement budget rather than the destruction of a long-built asset.

The 12-month test is one question. Does the mark serve the company now better than the proposed replacement would? If the answer is honest yes, hold. If the answer is no, change. The decision is closer to a product decision than to a brand decision and the speed of resolution should match.

The failure mode in this class is the cost of changing too often. Companies that change their mark three or four times in their first five years build no recognition at all and become hard to find in search and harder to remember in conversation. The 12-month freedom is real but the cost of using it casually is high in a different way.

The Squint Test

Across all three classes, the squint test is a single instrument that catches the worst class of mistake before launch.

Look at the existing mark from across the room with eyes half-closed. Then look at the proposed new mark the same way. If the new mark is less identifiable than the old one at distance, the change is reducing recognition equity. Most of the reversals in the archive failed this test before launch and shipped anyway.

Gap's 2010 sans-serif wordmark failed the squint test against the existing serif. Tropicana's redesigned package failed the squint test against the orange-with-straw. Pepsi's 2008 mark passed the squint test by a narrow margin and produced a non-reversal but a long debate. Mastercard's 2016 refresh passed the squint test easily and shipped without controversy. The test is cheap, repeatable, and ignored more often than any other diagnostic.

The Five Cases, Quickly Read

Coca-Cola. Class one. Refreshed dozens of times without announcement across a century. Never reversed because never loud. The recognition equity is the asset.

Pepsi 2008. Class one with class two ambitions. The new mark was intentional and well-funded but produced a years-long debate about whether the visible equity loss was worth the visible novelty. Not reversed, not celebrated.

Airbnb Belo 2014. Class two. The new mark replaced the cursive wordmark and survived an initial controversy because the company continued building the product, photography, and host system around it. A class-two refresh that worked because the structural work followed.

Mastercard 2016. Class two, quiet by design. The two-circle mark was refreshed without disruption. The case is the reference for how class-two refreshes are supposed to be handled.

BP sunflower 2000. Class one positioning attempted as class two refresh. The new mark carried environmental positioning the company had not earned operationally. The 2010 Deepwater Horizon spill exposed the gap. The cost of the refresh in lost credibility exceeded the visible budget by an order of magnitude.

The Operating Pattern

Logo changes fail when the brand and the design team agree on the change before agreeing on the class the brand sits in. The class determines the cadence. The cadence determines the budget. The budget determines the rollout. The rollout determines the result. The reverse order, which most refresh projects follow, is what produces the reversals catalogued elsewhere in the archive.

Most companies asking the logo question are class two. Most companies behaving as if they are class three are actually class one. The misclassification is the source of the loss.