Why Split Testing Matters: Outsized Returns

Split testing, or A/B testing, can generate outsized returns because small tweaks often unlock surprisingly large improvements.

You’ve Already Done So Much Work

You’ve already spent the budget, picked the stack, the hosting, the plugins, built the site, ran the ads. You coordinated designers, devs, clients, stakeholders. And the site converts at 3% (or whatever your industry standard is).

That’s not bad, but if you did all that for 3%, and you can move it by another 1, 2, 3, 4 or 10% just with a relatively tiny amount of effort, that lift is going to outperform new ads, redesigns, and plugin changes every time. It’s going to improve your ROAS, make all traffic more valuable, and make the marketers look like geniuses.

This is where agencies can create disproportionate value for themselves and their clients.

Change Measurement

Make a change? Measure it!

If it’s better, you would be silly not to report on it. Bask in the glory of your successful update.

If it’s worse, you can end the test and there is no permanent loss. Report your learnings and iterate with a new approach until you get it right. It is much better to find your own mistakes before the client does.

Unlimited Upside. Very Limited Downside.

Life Cycle Management

Over time, websites generally get worse at converting traffic. Maybe there’s a strong start for a year or two, but eventually sites start to fall off.

Trends, errors, changes, and omissions start to show up and confidence declines. If changes aren’t measured or considered, there is a chance they are making your site worse at turning visitors into customers.

With Split Testing and CRO, instead of growing old and grumpy, you grow strong and wise – your good ideas make your website more and more successful, and those changes are measured to ensure success.

Changes and updates that don’t work for your overall goals are undone, or something else is tried until your goals are met.

Compounding over months and years, it is a cheat code to success.