Many freemium services start out planning to attract users for free via PR, word-of-mouth and viral growth. Regardless of if actual growth is slow or quick, the question to grow via data-driven, trackable and paid marketing, a.k.a. paid acquisition, is likely to arise.
Looking at the entire conversion flow from a user registering for the free service, free users engaging with the service and finally converting to paid the flow looks like a leaky bucket. A significant percentage of users drop out during each stage of then funnel.
This will prompt a version of the question: Should we pay for pouring more people into the bucket when it has so many holes? Shouldn't we fix the conversion issues first and then quickly scale aggressively by ramping up advertising?
Unfortunately, the real world won't let you turn on paid acquisition and immediately start running at high speed with a high degree of control. You need to learn to walk before you can run. Unless you have the basics in place, you won't be able scale with control (which you want to have).
In order to have control when you scale, you have to invest ahead of time in people, systems, processes and market knowledge even while the foundation is a somewhat leaky bucket.
As you're learning about acquisition and plugging engagement and conversion holes, a good plan is to invest more in fixed costs (online marketing staff, engineering time, tracking systems) while investing relatively less in media spend.
Among the things that take time to learn and implement as an organization are:
* Where can we buy registrations cheaply or at least cost effectively?
* Which ads are effective (per channel and country)?
* How do we quickly translate ads into multiple languages and produce multiple language versions?
* How do we track user behavior from click to registration to multiple transactions over time? (Will take engineering time to implement and quality assure.)
* What's the revenue over time, per channel, from user cohorts coming via paid acquisition? And the only way to learn about behavior over time is to have old cohorts. (They'll usually perform worse than organic users, but you want to know how much worse and adjust your target.)
More posts on freemium strategy: