There's enough differences between what is meant by "network effect" that it's difficult to understand why companies might be on one end or the other (or even which line of business we're discussing).
Overall I like the work here and I personally haven't done the legwork to come up with my own theory on a moat visualization. However, I do feel this framework is a stretch and would need more justification for a) why these two axes matter the most and b) better definitions for the axes.
Microsoft has commoditized their ecosystem, look at all of the equivalent PC makers, equivalent Silver/Gold/etc. parters, LARs, VARs, number of anti-virus vendors (and other software companies, etc.) all replaceable and Microsoft drives them business through an overlay sales and marketing team. They are the best in the world at this, Cisco has been the 2nd best.
Facebook is more vulnerable to loss of 3rd party content than the article states. They have many categories of content now and they've done an excellent job of building deep hooks making their platform the ideal place to publish so the moat is deep and wide but it isn't an invincible business. The # of DAU and the amount of activity on the platform would drop if Facebook lost any of it's third party content sources (1) publishers, (2) businesses, (3) community organizations, (4) topical groups, (5) marketplace sellers, (6) event organizers, (7) .., (..) .., (N) .. [I could go on for a while longer on all of the different ways Facebook is a platform within a platform fighting on multiple fronts, the 6 give a good enough of an idea to get started..]
If however, you're on the other end of the spectrum, and your network effect is internalized and suppliers undifferentiated, it doesn't make sense to ask what the killer app is.
Where the moat map fails for me is when I try to slot Homejoy in there. Like Uber, they have an externalized network effect. They assumed, at first, that their suppliers were fungible commodities; but found out that they weren't. People tend to want cleaners into their house that they trust. So by that notion, according to the moat map, they should have had a moat. But Homejoy ended up not doing well, for various reasons. So I'm guessing the moat map isn't the whole picture.
By contrast, you can't get the benefit of Microsoft's Windows ecosystem, or Apple's app ecosystem, without first giving Microsoft/Apple the full amount of revenue they expect from you.
In other words, a differentiated supplier / externalised network moat is only a moat if the vendor can get all its revenue upfront, or can reasonably ensure that the moat will keep customers locked in for a long time.
Google gives slightly more deference to established content providers, but not much
Re: YouTube: Apparently, this deference is directly proportional to economic power. Small creators have basically no power. Large creators have some. Corporations have enough power to make demands and cause YouTube to change things overnight. Some upstart should make note of this huge class of long-tail marginal creators become serfs.
Ultimately, the future culture will come out of this edge of the culture. YouTube and Google seem to have reached a point, where one can use their culture/groupthink against them, and they will steadfastly refuse to change.
I think there's a missing nuance here with respect to profit. I think developers are drawn by how efficiently they can make money on a platform, not strictly by how many users a platform has.
Microsoft is one of the key example used. In their heyday they were deeply integrated. They had the whole stack and they sold it enterprise-level and enterprise-wide. Companies where "Microsoft shops" for this reason. This philosophy was one the the reasons Microsoft was late to the Internet party.
I also disagree deeply that "Facebook could lose all of its third party content providers overnight and still be a compelling service." Equally on that spectrum Facebook is right next to Google. Who surely couldn't lose all their third party content overnight?