Walking the show floor at CEDIA and through many other venues, I hear the question come often: “what’s the licensing cost to integrate Alexa Voice Service into our product?”. It takes a bit to understand why Amazon would give away a service that definitely costs money for it to run. The servers to power voice interaction, NLU, text to speech, etc. all consume real electricity and then there are all of the maintenance and other costs.
However, Amazon’s bet is that the ubiquity of Alexa will lead to more purchases. When someone does end up purchasing products through Alexa, because Amazon owns much of the distribution and marketing, the margin it gets is much higher than if some other channel made a similar product (20+% vs 2–8%).
The result is that even a slight increase in sales is a large contribution to profit. This week there was news out that this is in fact the case for Amazon — that Echo owners are purchasing more product. Whether this is correlation or causation, I’m not sure. The TechCrunch article states 10% increase in purchases, likely from Amazon Prime users. A 2013 study found prime members spent $1,300 annually (probably higher now, so let’s keep it at that level to be conservative).
A 10% increase would be another $130 / year. For the 1.6M that the same NPD study shows for Echo owners (but could be >5M), that would be an increase in Amazon spending $208M a year in additional sales as a result of the Echo! If we assume the 5M, the Echo is adding $1B revenue annually — or maybe $200M in pure profit.
This could get close to making the whole Amazon Alexa venture profitable today.