A long time encyclopedia editor (who worked at Britannica, World Book, and Encarta) told me that door to door encyclopedia sales were dying well before Encarta. Due to the rise in two income couples, which radically narrowed the time window when you could get anyone to give you time at the door...
I think the phase that usurped Ecarta's position had more to do with good quality web search and to a lesser extent the crowdsourcing model of Wikipedia (run by a foundation rather than a for profit mega corporation) than to content or data standards.
Search within an encyclopedia (e.g., Wikipedia) doesn't rely on Web search generally (e.g., DDG).
I can see your point that content outside a curated context replaces an encyclopedia to an extent, and up until about 5-10 years ago, as clickbait and black-hat SEO finally won out, much noncommercial Web content was at least informative, if of generally lower quality than traditional printed sources. Convenience has a huge edge over quality, though. These days, it's Wikipedia's curated content, especially for complex breaking news stories, that is my first go-to. After Idleword's (Maciej Cezglowski's) piece on Hong Kong hit HN last week, I finally took a look at Wikipedia's article on the protests there. A full 73 page long article with references on a protest movement only a couple of months old. That's staggering, and exceeds all but the very best news sources. (Another case I'd noted was the Oroville Dam case, where Brad Plumer's article for Vox was the only trad med piece I could find even remotely approaching Wikipedia's article. The first instance of this I recall was the 2004 Boxing Day quake/tsunami in Indonesia / Indian Ocean.)
Wikipedia, on the other hand, greatly enhances Web search, though it also benefits generally from the high profile resulting from that.
Wikipedia's crowdsourcing, the reliance on underlying technologies (Wiki, HTML, mediawiki markup), and on a huge set of organisational systems, standards, practices, and solutions Wikipedia and the Wikimedia Foundation have arrived at, were transformational. Though those too strongly resemble the largely equivalent or analogous practices of earlier encyclopaedic efforts dating back to at least Diderot, as well as other reference works (OED, see Simon Winchester's The Professor and the Madman).
(Note: updated to add 2nd 'graph beginning "I can see your point...")
The arrival of good internet search along with the ever increasing pool of web content doomed a curated content approach such as was used by Encarta (which could never aspire to be as broad or deep as the web). I can tell you, based on watching sales figures and talking to many customers, that web search was a much bigger factor for the business than was Wikipedia. In any case it would have been much trickier to get the level of crowd input given to Wikipedia by a for profit enterprise. Regardless of tools or formats used.
Btw despite the brief lifespan of the product, Encarta gave Microsoft reasonable return on investment. And arguably more importantly, it helped to quickly entrench the Multimedia PC standard, especially in homes. Which helped reduce Cost Of Goods Sold for all of the company's products, which netted a very nice return in cost savings.
