Is Amazon’s Voice Assistant Alexa a modern Trojan Horse?
If you are not familiar with Homer’s classic poem, The Odyssey, here is a quick reminder of how the Trojan horse episode unfolded:
After 10 years of unsuccessful siege trying to take over the city of Troy, the Greeks decided to change their approach. They built a huge wooden horse which they left in front of the city of Troy and pretended to sail away. The Trojans, thinking the Greeks had finally given up, pulled the horse into their city as a victory trophy and threw a big party. That night, a little group of Greek soldiers who had in fact hidden in the giant wood construction, crept out. They opened the gates for the rest of the Greek army, who had sailed back under cover of night and easily destroyed the city of Troy — thus ending the war.
What is the link with Amazon?
The Gift of Alexa
Amazon pushed most of its own tech products (kindle, tablets, TV fire sticks, and speakers) during Black Friday 2018 but the lion-share of the merchandising placements and the heaviest discounts went to its range of Alexa-powered Echo smart speakers.
Alexa is a voice-controlled virtual assistant created by Amazon in 2014. It gives users the ability to control products throughout their home, listen to music, order products online and more. You can ask Alexa pretty much anything (including useless stuff), order a pizza or get some laundry detergent delivered next-day, all of this from the comfort of your sofa and without even needing to type anything.
During Black Friday, you could grab the Alexa-powered Echo Dot for as little as £19.99, a 60% discount vs. the normal retail price (£49.99). Now 60% discount might not sound like such a big deal (especially for Black Friday), but it is when you break down the production costs. According to firm ABI research, the total cost of components used to produce the Echo Dot sits around £27 ($34.87 exactly). This cost, which does not include labour, shipping, power consumption, and so on, would mean that Amazon lost at least £7 for each Echo speaker sold on the Black Friday deals (Amazon said they sold “millions”).
This specific example tries to illustrate a bigger point: Amazon is and has been structurally losing money with Alexa (reportedly hundreds of millions of dollars in 2017 alone).
Why is Amazon being so generous?
Remember the Trojans
One lesson from Homer’s Trojan horse episode is that an unexpected gift should always be looked at twice. We highlighted earlier what customers were getting out of Alexa (unprecedented convenience from a great cheap piece of technology) but one should also ask what Amazon gets out of it? After all, if the tech giant is willing to slash prices and lose hundreds of millions of dollars on a new product line, there must be a good reason.
Here are three:
1) “Hey Alexa, make customers buy more”
Alexa’s main purpose is to make sure Amazon consumers, well, consume. This comes as no surprise: if you have an Alexa device sitting in your kitchen, you will inevitably end up buying more stuff from Amazon as well as further embedding yourself into the whole ecosystem by using it services (Amazon music, Audible…). For each Echo device sold, Amazon therefore banks on increasing the average spent per customer; a strategy referred to as creating ‘customer stickiness’, which Amazon has been excelling at for more than 13 years with its loyalty programme called Prime.
Amazon customers can, for £79 per year ($119 in the US), become Prime members and get free next day delivery, as well as access to services like Amazon music or Amazon video and other benefits. For customers, Prime combines convenience and great value (the ‘real’ value of the program has been estimated at $785/year). For Amazon, it’s the best way to push customers to spend more and more often. Prime members spend on average more than twice as much on Amazon as non-Prime members (this gap keeps growing) and 46% of Prime shoppers make a weekly purchase on Amazon vs. 13% for non-Prime.
In its infancy, Prime was often criticised for being a money drain but given how instrumental the programme has been to the company’s growth in the past 10 years, Amazon can only hope Alexa follows the same footprints.
2) “Hey Alexa, boost Amazon’s private label brands’ sales”
Amazon has become the default product-search engine of the web with more than 50% of consumers starting their product search on Amazon (vs. only 28% on Google). The search query data (i.e. what people type when searching for a product) combined with a goldmine of user information, put Amazon in a unique (unfair?) position to figure out exactly what customers want — and they can produce these items cheaper than anyone else thanks to economies of scales and cheap access to capital.
Enter Amazon private labels. You might be familiar with Amazon Basics, the brand which launched in 2009 (with a handful of low-cost, generic phone cables) and that now offers more than 1,500 products. Well this is only the tip of iceberg: Amazon.com now houses more than 120 of its own private labels (nine times more than in 2016) but only occasionally displays that they are owned by Amazon. These brands include for instance Mama Bear (diapers), Happy Belly (coffee), Solimo (razors), Rivet (furniture), several apparel and jewellery brands as well as a brand called Wag (pet food) which sounds surprisingly close to the UK pet food brand Wagg…
Amazon private labels are expected to reach $25 billion in sales by 2022 and will play an increasingly important role for Amazon. How does Alexa help this strategy?
“Total sales across Amazon private label portfolio are expected to reach $25 billion by 2022”
Amazon works like a search engine: 90% of its users start with a search query and only 30% ever go to page 2 of results. Therefore the key for a brand to succeed on the site (let alone exist), is to rank on top of page 1. With voice ordering, customers can’t scroll through product listings anymore. Instead they must rely blindly (literally) on Alexa’s recommendations, typically two product results for voice-activated shopping searches.
What happens when a shopper searches for an item on Alexa that is offered by Amazon private label? According to Bain, Amazon first recommends its own private label products 17% of the time, despite the fact that these products make up only 2% of products sold. In other words, Amazon tends to favour its own brands vs. other brands thus further reducing an already very crowded ‘virtual shelf’.
When one accounts for the fact that voice search is predicted to reach 50% of total product search by 2020, this is pretty bad news for small and medium brands who are seeing their chance to rank on top of page 1 seriously reduced (bigger brands will most likely pay to keep getting the best placements). Given how much SMEs have contributed to Amazon’s growth and success over the past decade this is tough (did anyone say Nucleus?). This is potentially bad news for customers too: who wants to use a search engine biased towards recommending products which offer Amazon more margin instead of recommending the most relevant ones?
3) “Hey Alexa, spy on customers”
The use of the word ‘spy’ here is voluntarily provocative as I don’t believe Amazon wants to spy on its customers. However, it is not news that the Four (Facebook, Apple, Google and Amazon) are collecting an unprecedented level of personal users’ data to better understand their customers’ behaviours and habits (i.e. sell them more stuff) and the rapid adoption of smart speakers and AI powered voice technology takes customer ‘stalking’ to another level.
First, let’s get the big question out of the way: is Alexa always listening? The answer is yes and no. Yes, it does, as it is the only way it can respond to the trigger word ‘Alexa’ when customers dictate a command. But no, this information isn’t recorded, stored or used in any way. Amazon has no incentive to do so: it’s too much data to analyse and too big of a risk to take with customer trust, a big value of Amazon’s.
However, all voice requests, labeled as “Alexa interactions”, are recorded, stored in the cloud and linked to a specific Amazon customer’s account. Amazon analyses all those interactions in order to ‘improve its services’. This new set of data goes well beyond online shopping (it also includes questions asked to Alexa, messages sent via the assistant,…) and will give Amazon deeper penetration into the private lives and desires of its consumers.
What does Amazon do with this data?
The new data will be combined to existing user’ data from Amazon (e.g personal addresses, browsing history, search queries, order history, reviews…) to build even more accurate customer profiles and further personalise the shopping experience (think recommendations). This will eventually enable Amazon to… you guessed it, sell more stuff.
But there’s something else.
In the past 5 years Amazon branched out into a new and very lucrative (higher margin) business: advertising. While Amazon doesn’t sell its customer data, Amazon Media Group (AMG) its media branch, gives brands the ability to market their products to Amazon customers both on and off Amazon (on news sites for instance) based on demographics, shopping history or browsing behaviour. Amazon’s ad business has been growing extremely fast (AMG revenues doubled year-on-year to $4.6 billion in 2018) but still only represents 4.2% of a $111 billion digital ad market dominated by Facebook and Google.
And this, in my opinion, is the main reason for Amazon’s aggressive move into the smart speaker market. By positioning Alexa as the new ‘operating system’ of millions of households worldwide, Amazon could access an unprecented amount of data and offer an extremely granular level of insight into its customers. This would put AMG in a great position to capture a bigger share of the digital ad market and catch up with its competitors.
Amazon announced last Tuesday that Black Friday was the “biggest holiday shopping weekend ever for Echo devices, with millions sold worldwide” and that “Echo dot was the #1 selling product on Amazon globally, from any manufacturer, in any category”. Yes, the market for smart speakers is significant and growing fast (estimations value it at £3.5bn in the UK and $27.8bn worldwide in 2022) but the stakes are higher: Alexa has the potential to become the new operating system of millions of customers around the world, thus reinforcing Amazon’s position both as an online retailer and as an advertiser. Is this bad?
It depends on what angle you look at it from.
For other tech giants, this is bad news. Alexa reinforces Amazon’s already very strong advertising business which will eventually capture market shares from the two biggest digital advertisers (Facebook and Google) as well as potentially compete with existing smart phone operating systems (Apple’s iOS and Google’s Android). Responses to Alexa are different: Google and Apple both sell their own smart speakers (respectively Google Home and Homepod) featuring their proprietary voice-controlled virtual assistant and have a similar approach to user’s data. Facebook on the other end, recently announced that its new Portal smart speakers will be equipped with… Alexa.
For brands selling on Amazon, Alexa presents both opportunities and threats. It has the potential to completely reshape search, merchandising and advertising on Amazon thus rewarding companies who understand the new landscape and take the right actions. There is a risk though that some companies (mainly SMEs) will get left behind as Amazon uses Alexa to increasingly push its own brands and big brands who pay for preferential placements.
What about customers? It is probably too early to say but given that half of all UK households will own a smart speaker by 2022, it is only fair to ask the question. Here, the differences between Alexa, Google Home or Siri become somewhat irrelevant. All three companies argue that, in order to offer the best possible experience with technology, they need to capture, store and analyse users’ data. Their hidden agenda, as discussed earlier, is the monetization of data through advertising (although to a lesser extent for Apple ) which seems to leave customers faced with the Faustian bargain of convenience vs. privacy.
Should we have to sacrifice one over the other?
Snips, a French start-up which raised €22 million to date, is working on a block chain-based AI assistant which stores and process user’s data locally instead of sending it to the cloud. Snip directly challenges Amazon and Google’s approach by offering a decentralized machine learning solution which guarantees privacy. As consumers become increasingly aware of the privacy concerns with voice assistants, there might be more opportunities for innovative companies with such solutions.