When we talk about running macro A/B tests, it's an example of a broader idea in growth: taking big swings.

Examples

When you make a landing page that's 80% shorter than your current landing page, that's a big swing.

Changing the color of your CTA button is NOT a big swing.

When you set up three Instagram ads — a screen recording video, an ad with only text in the image, and an image of someone tweeting about you — those are three big swings.

When you set up three Instagram ads — three screen recording videos with different body text — those are NOT big swings.

If you typically email users asking for a Yelp review, and then you A/B test texting them a meme like this asking for a review... that's a big swing.

(Maybe not one you should do. But a big swing nonetheless.)

In other words: "Big swings" means testing things that are wildly different from each other. Then doubling down on what works.

Why are big swings important?

When you take big swings, you win big and you lose big; the results are very different between swings. But you get to cheat because you can only keep the ones that win - and prune away the losers. So you improve much faster over time.

Find big winners. Quickly get rid of losers. It's evolution at work.