I mean, it was more than marketing. It was hard to build an app that would use WP8 features on WP8 but also run on WP7. If you kept building for WP7, the WP8 features were unavailable; you'd end up with maybe a shared core and a WP7 app plus a WP8 app, and then WP8.1 did the same kind of thing, and finally WM10 let you build "universal" apps which would only run on WM10, which never managed to get an current install count more than WP8, and might not have managed to surpass WP7 either.

WP7 itself was a throwing away of compatability with Windows Mobile apps; reportedly, that loss of goodwill is why Google refused to support the platform. WM6.5 apps would have been out of place on WP7, but if Microsoft was still yelling about developers x3, maybe they would have made that transition possible --- require a new frontend for cohesive UX, but allow for core libraries from older apps to run as-is, instead of forcing the new apis.