As many of you know already, main objectives for start-ups, especially in early phase, is keeping a stable growth in acquisition of new customers. Especially in e-commerce space, this results in a very targeted strategy for Online marketing channel managers to optimize their spending in line with companywide overall CAC limits.
CAC stands for Customer Acquisition Cost and is computed as overall marketing spent over number of new customers acquired in a time period of reference. If you are under budget constraints (the case for most marketing plans) in terms of total spending, let’s say in a month, then you face an allocation problem between different marketing channels. The situation might look like as follows in terms of potential performance-driven channel split in a given day:
From a quick reasoning, you can think that the Google SEM is the best because it would let you acquire customers for the lowest cost, but looking at CAC is only half the story. What we need to factor in the equation is also lifetime revenue (or profits, depending how your company is managing its CLV model) for those new customers, by the different acquisition channels. You can find below a very basic representation of this idea:
* this is an estimate over a 2 yrs period of how much revenue on average a new customer from that channel will contribute to the business (considering her/his first purchase).
** here defined as Total Revenue – Total marketing spend. For simplicity, it does not take into account other costs (such as COGS or Operations).
When you consider estimates/metrics other than just acquisition cost (such as the value that new customers will bring to your business), your marketing strategy can be optimized to result in better customers and less money spent in marketing. As per example above, you might prefer to focus on Display network as acquisition channel because it has the highest overall expected profit of the three sample ones.
In the next issue we will be covering more in details how to actually calculate Customer Lifetime Value following one of most used approach (with Historical CLV using Average Revenue Per User and Cohort Analysis). Then eventually we will briefly discuss more advanced techniques for Predictive CLV (the one I am using) that are based on regressions and Bayesian Inference.
Image source: Flickr