One fundamental difference between marketing for freemium products and marketing for traditional software products is that the marketing function doesn’t end when a user downloads and installs a freemium app. A freemium user is not automatically a revenue stream — in fact, before making her first in-app purchase, a freemium user is nothing but a cost: she’s taking up space on your server and using up bandwidth with her requests. This characteristic of freemium products muddies the definition of marketing; since not every user is a paying user, marketing for freemium products isn’t just the process of acquiring users — it’s also the process of trying to convert all users into paying users.
This new convention splits the marketing function into two functional groups: upstream marketing and downstream marketing. Upstream marketing is what’s done to acquire users; downstream marketing is what’s done to monetize (and maximize the revenue from) existing users. Both of these marketing functions are tied together through data: as a user flows through the freemium app, she creates data artifacts that contribute to the app developer’s ability to segment and optimize the experience of all users.
This new conceptualization of marketing necessarily encroaches on traditional product management and design functions: a product is not static if it has to adapt itself to user behavior, so product development becomes a real-time outgrowth of marketing. I don’t think this development is strictly what Regis McKenna referred to when he said that Marketing is Everything, but it’s nonetheless the case: product design is informed by marketing in freemium apps. Data-driven real-time design is the raison d’être of the freemium model: with lots of data, the experience of a very small, paying minority can be optimized to such an extent that their payments will eclipse the revenues that would have been earned had the app been paid for by everyone.
This reality adds a technical component to every aspect of a freemium marketer’s job. User acquisition can’t be engaged in recklessly; it has to be justified by LCV, which in turn is calculated and projected in real time based on user data. Apps must be instrumented very thoroughly so that every action contributes to the user profile model. Upstream marketing entails bringing in the best users at the best (not necessarily lowest!) prices; downstream marketing entails presenting users with the optimal number of opportunities to buy the virtual goods they’re most likely to enjoy.
Imagine walking into a store and seeing the shelves rearrange themselves before your eyes so that the items you most wanted to buy were placed next to the entrance. Now imagine that the store knew what items you’d most like to buy on your first visit because you decided to shop there after seeing a particular billboard, and people with whom that particular billboard resonated generally exhibited similar shopping patterns. And now imagine that all of the estimations and projections the store made about you were correct. That’d be a paradigm shift in retail, right? Well, that’s the power that the freemium business model provides app developers, but the key to taking full advantage of the model exists in the connection between upstream and downstream marketing functions — being able to classify a user based on the limited knowledge of how they entered the process.
So what are the limitations of the freemium model? When would implementing the freemium model be inappropriate for an app? The answers to these questions are determined by the scale of the app: the freemium model requires the large volume of data on which the concept is predicated to be used for optimization. In other words, if a freemium app can only be supported by upstream marketing (user acquisition), then the freemium model won’t be exploited to its fullest and very likely the app won’t produce more revenue than it would have as a paid app. The success of a freemium app requires three things:
- A massive potential audience. Niche apps shouldn’t be freemium; not enough people will use them to provide enough insight into how to optimize the experience.
- A robust analytics architecture that can track users by source and estimate and project LCV in real time. Without this, freemium doesn’t work: acquisition campaigns won’t be informed by data.
- A large, diverse product catalogue. The ability to optimize the monetization experience is the keystone to the freemium model; if the app relies on only one or two paid products, a mixed set of user tastes will be impossible to cater to at an individual level.
Data (and, by proxy, optimization) is the connection between upstream and downstream marketing; without it, an app can’t be successful under the freemium model.