Personas don’t dictate giant changes in established behavior. They instead serve to represent the people we are creating things for. Personas are not perfect. Neither are people. That’s why they’re effective tools.
That’s the biggest thing: they’re tools. They can be used incorrectly. So can user research. So can established paradigms, actually. (I can simply ape Amazon’s checkout process, but that won’t do me any good if I don’t know how site users respond to the nuances of my client’s product.)
Good personas are based on real user interviews. They help personify user research. They help us keep in mind the people we’re writing and designing and developing for. They aren’t gods.
The example website in your second blog post isn’t bad because of the personas. The personas didn’t force the designer or writer to create seven different messages - the designer or writer failed to understand how to synthesize the information the personas offered, which is to say they failed to create any kind of message hierarchy, tone hierarchy, audience hierarchy, etc.
So I think maybe the real message is, “Understand the role personas play,” not, “Personas are awful and will ruin your work.”
Personas help us understand that, indeed, different people come to the website looking for different things. They help inform writing and design. You mention Coke in your article, but fail to mention how Coke also has employed different messages for different audiences. There are polar bears, sure. But there are also celebrities. There are jingles. There are sports sponsorships. Each of these messages appeal to a different audience. And it’s the mix of these different messages that makes the brand what it is.
When a user lands on a website, the goal isn’t for them to see one message. The goal is to see the one message that matters most to them.