Back in 1999, I worked for a small San Francisco start-up. They were primarily a tech group building productivity applications using a new combination of technologies: JavaScript and HTML, which we called DHTML. (Most people know this technology combination as AJAX today.) To this day, I have never worked with software engineers as good as this crew. They are hands down my favorite dev team.
But back in 1999, DHTML was new. No one was using it.
I had the opportunity to usability test our first product and prototypes: a web-based presentation and a word processing application. Basically PowerPoint and Word for the web. I'm talking about creating, editing ppts and word docs in a browser. Netscape 4 browser. A whole new paradigm. My testers understood the concepts quickly and loved the ease of use. Many got how paradigm shifting this was and loved the idea. The company planned to build an entire suite of productivity applications using this technology.
But then, we got acquired. (Well, it wasn't so simple as that.) And next thing I knew, I was down in Mountain View presenting the technology to large internal groups. My role shifted and in the next year, I must have presented this technology (thanks to the amazing dev team) to over 30 groups inside AOL. Many groups immediately saw the user benefits - stuff that you and I take for granted today. Pop-ups, mouse-over dynamic windows, i-frames, divs.
Inevitable, the question would be asked about metrics. Back then, all metrics were page refresh based. Ads were displayed based on each page refresh. So if there was no page refresh (because we no longer had to refresh the page each time a user did something) that decreased the number of page views, which decreased the number of ads that could be served. Sure, we'd show that people stayed for minutes on a page, instead of leaving immediately, but even this metric was too new a concept.
Our new technology could not be "successful" in the terms of the old paradigm. Sure, we'd make the application more interactive, user friendly, basically have the user stay in the property using the application a LOT longer and increase usage, but.... it didn't show an increase in metrics that defined success at the time.
And that's because, the new paradigm was not "successful" by the metrics of the old paradigm. The mental mindset had to shift. The definition of success had to shift. And it took 4+ year before that happened.
The success metrics of the current paradigm stifled the adoption of the new paradigm.
And that's just one challenge when you're working with/creating paradigm shifts.