One innovation the freemium model brings to bear is analytics as a fundamental component of the product development lifecycle: since distribution (and thus customer adoption costs) are 0, behavioral customer data is available with enough volume to develop new streams of revenue from it. The freemium model accords an analytics team the opportunity to conceptualize new sources of revenue; under traditional models, analytics at best can merely optimize existing revenue streams, and at worst it simply reduces expenses.
This phenomenon is an outgrowth of the inverse relationship between user base size and barrier to adoption (cost). At a price point of zero, the size of a product’s potential user base size is unlimited; as the price point increases, the number of potential customers decreases. This is obvious, but what is less axiomatic is how that relationship empowers analytics as a product development tool.
As a product’s user base increases, the data artifacts it creates become more valuable because their volume renders them actionable. Optimizing a 1 million-user revenue stream is easier than a 100-user revenue stream: the data used to make decisions is more reliable, trends are easier to spot, and results are easier to communicate to management because black box argumentation isn’t necessitated.
I define the phrase “black box argumentation” as a faith-based description of a specific result that relies on statistical or algorithmic techniques too complicated to explain in a presentation to management. These are “trust me, I’m a scientist” means of influencing a decision based on the results of an experiment or analysis, and they are very rarely persuasive – the more opaque a process, the less likely its results are to influence decision making. Analyses are most effective when they’re based on a simple operational conclusion that can be clearly communicated (Powerpoint, Excel) and conceptually absorbed.
So how does freemium facilitate the development of autonomous data products — without relying on black box argumentation — when seeking green lights from management? By providing large volumes of data on which to conduct straightforward analyses. This is one of the primary benefits of “big data” as an organizational strategy: it shifts complexity onto data architecture and processing from analysis, allowing for results to be communicated more clearly and convincingly. Thus, within the context of freemium, data products are easier to pitch and implement at the organizational level, imbuing the organization’s analytics team with a product development and management directive.
So the belief that analytics systems exist simply to reinforce product maxims that an organization already believes to be true doesn’t apply under the freemium business model – since big data is the foundation of freemium, analytics is a product driver. Analytics under freemium is not an intellectual or philosophical initiative; for freemium, analytics is the scientific framework used to generate revenue.
This perspective is at odds with some legacy definitions of analytics, which generally fall along the lines of “a reporting system that alerts the organization to problems”. Analytics in freemium should describe the current state of the product – its usage, its growth, revenue statistics, etc. But analytics should also be useful in delivering meta data about the product that can be used to not only enhance its current functionality but to develop additional functions that may not have existed in its original development scope.
The question, then, is how this is achieved. The most obvious means of using analytics to produce new revenue is to invest in it sufficiently through hiring the right people and building out the right infrastructure. But for an analytics group to possess the agency to produce new revenue streams, the entire organization must be dedicated to data-driven product development. This is probably the hardest aspect of implementing freemium: integrating “reporting” and “design” functions under the product umbrella group as opposed to segmenting them.
I think the company which best embodies the notion of “analytics as a revenue driver” is LinkedIn. Page four of its Q4 earning report reveals stable page views for the past four quarters, but pages five and six report growing revenues and changing revenue make-up. Why would revenues grow when page views remained stable? Because the percentage of revenue obtained through the “marketing solutions” and “talent solutions” product verticals increased. “Marketing Solutions” and “Talent Solutions” are (seemingly) comprised substantially of data products; the suggested jobs presented in the sidebar of the LinkedIn homepage. LinkedIn has been generously public about its use of data in driving revenue — some recommended reads are:
- Building Data Products Using Hadoop at LinkedIn
- Data-Infused Product Design & Insights at LinkedIn
- Data In and Data Out: Using Hadoop to Create Data Products at LinkedIn
- Developing Data Products
I’ve written before that big data is the foundation of the freemium model, but this is only true if analytics is approached with genuine intent. Analytics can’t be pursued “part time” or with minimal expenditure in mind – as a fundamental constituent of the product development process, analytics must be staffed, resourced, and prioritized with the same commitment as any other product-focused initiative.
Recent