Link tags: abtesting

Is client side A/B testing always a bad idea in your experience? · Issue #53 · csswizardry/ama

Harry enumerates the reasons why client-side A/B testing is terrible:

• It typically blocks rendering.
• Providers are almost always off-site.
• It happens on every page load.
• No user-benefitting reuse.
• They likely skip any governance process.

While your engineers are subject to linting, code-reviews, tests, auditors, and more, your marketing team have free rein of the front-end.

Note that the problem here is not A/B testing per se, it’s client-side A/B testing. For some reason, we seem to have collectively decided that A/B testing—like analytics—is something we should offload to the JavaScript parser in the user’s browser.

Plain Text vs. HTML Emails: Which Is Better? [New Data]

Spoiler: it’s plain text. Every time.

Nothing boosts opens and clicks as well as an old school, plain-text email.

I feel vindicated.

People say they prefer HTML emails ..but they actually prefer plain-text.

This seems like a plausable explanation:

Think about how you email colleagues and friends: Do you usually add images or use well-designed templates? Probably not, and neither does your audience. They’re used to using email to communicate in a personal way, so emails from companies that look more personal will resonate more.

Now get off my lawn, you pesky HTML-email lovin’ kids.

Statistical significance & other A/B test pitfalls

Cennydd delivers a slap of common sense to A/B testing. With science!