Copy what everyone does badly, not what everyone does well

Rory Sutherland has a talk where he takes apart something almost nobody questions: benchmarking. Every company sends people out to study the market leader, write down what they do well, and bring it home to copy. It feels rigorous. It feels like due diligence. Sutherland's point, borrowed partly from Roger Martin, is that it's actually a trap. Do it long enough across an entire industry and every player ends up looking the same, competing on the same margins, fighting over the same shrinking difference. He calls the result an over-served market. Everyone doing the well-known things very well, nobody doing the neglected things at all.

The fix he proposes has a name that sounds like a typo the first time you hear it: reverse benchmarking. Instead of studying what the leader does brilliantly, you study what they do badly, and you fix that instead.

The restaurant that took it literally

The best version of this story is Will Guidara's, told in his book Unreasonable Hospitality. His restaurant, Eleven Madison Park, was ranked around 50th in the world. He took his team to eat at restaurants ranked above them, and told them not to take notes on what those places nailed. Take notes on what they were mediocre at.

Two things kept coming up. The coffee was fine, nothing more. And anyone who ordered beer got treated like a second-class guest compared to the wine drinkers. So Guidara didn't hire another chef or redesign the dining room. He made one of his staff, a genuine coffee obsessive, the restaurant's coffee sommelier. He made a craft-beer nut in the kitchen the beer sommelier. Two small, cheap, specific fixes, aimed at exactly the gaps nobody else in that market had bothered to close. The restaurant went from 50th to first.

Nothing about that fix required more money than the competition was already spending. It required looking somewhere else.

Why this actually works as a shortcut

The reason reverse benchmarking is a genuine innovation shortcut, and not just a cute reframe, is that it points you straight at the one part of the map nobody else is looking at. If ten competitors all benchmark against the same leader, they all end up polishing the same five things that leader already does well, because that's what shows up in every audit and every competitor deck. The neglected stuff never makes it into anyone's slide, because nobody's measuring it and nobody's proud of it. Which means it's completely uncontested. You don't need to out-innovate anyone there. You just need to show up.

Sutherland calls the obsession with easily measured metrics an "arithmocracy," a system that rewards whatever's countable and quietly ignores whatever isn't. Reverse benchmarking works precisely because it goes hunting in the part of the business nobody's counting. The gap is real, it's usually cheap to close, and it's invisible to every competitor still busy benchmarking against each other's strengths.

The version that matters for you

You don't need a five-star restaurant or an Ogilvy vice chairman to use this. Next time you catch yourself studying a competitor, a client's old agency, or a product you admire, flip the question. Don't ask what they do brilliantly. Ask what they clearly can't be bothered with, the thing every customer mentions once and then stops mentioning because they've given up expecting it to change.

That gap is where the fun, the difference, and the actual value are sitting, completely undefended, waiting for someone to notice they're allowed to just walk in and fix it.