The 'conversion option' is just a bottom of funnel component of your acquisition strategy, and the costs to support those freemium users is still a component of

The 'conversion option' is just a bottom of funnel component of your acquisition strategy, and the costs to support those freemium users is still a component of your marketing spend. For established marketing channels/user segments, you should be able to identify your response curves and probability of conversion at each stage of your marketing/acquisition funnel, including from freemium to paid user. If freemium conversions is a component of your sales process, you should absolutely be attributing the costs associated with supporting freemium users towards marketing, net any freemium-specific revenue streams such as ads. It's a fairly substantial component of the fully burdened cost per acquisition, and by allocating the expense towards marketing rather than diffusing it into operations, you get a much more actionable view of your true CPA. Being able to do so at a granular level requires putting in effort to integrate your marketing analytics with your operational analytics. But even just looking at it in aggregate is beneficial to understanding acquisition costs. Network effects complicate that calculation a bit, depending on the nature of the network effect. Freemium Dropbox users sharing files could lead to net sign ups, as you mentioned. But the value from that is incremental user acquisition, and is still solidly a component of your marketing/acquisition costs. Contrasted with your social network example, where there is intrinsic leverage afforded by the size of the network that's independent/distinct from the potential to drive incremental user acquisition. VC is its own beast. As you said, it's essentially a form of economic dumping. They prioritize growing market share first, and identifying network effects and associated leverage potential after the fact. Only after all that do they really mature their business model, and does it make sense to try to understand what component of network effects is a marketing benefit vs. a product benefit. Uber is a perfect example of this - their original pitch[1] was very solidly as a taxi app, and one that used exclusivity as a selling point no less. Virtually all of their pitched areas of growth were around potential for optimizing personal transit, with only a single mention of branching out their infrastructure to other location based services applications. Now they've distanced themselves as much as possible from their identity as a taxi service and repositioned themselves as a logistics and transportation juggernaut. And the ubiquity, availability, and consistency of their network is a core component that enables that product identity. [1] https://medium.com/@gc/the-beginning-of-uber-7fb17e544851 https://medium.com/@gc/the-beginning-of-uber-7fb17e544851