The stage is set for Amazon but is building an ad business a big enough priority?

Before I go any further, a few more numbers for context. Data to understand how Amazon’s ad revenue breaks down is nigh impossible to come by, but it’s fair to say that today this dominated by display advertising and more latterly programmatic with the launch of X. There is therefore no significant revenue coming from search today, a landscape that Google still dominates with 65% share in the US and over 90% in most of Europe. In terms of revenue, however, Google takes almost 80% of all search ad revenue, so why then is everyone so excited about voice search and Amazon specifically ?

The technology, and therefore consumer behaviour, is pivoting again. There was a time back in the early 2000’s, when search behaviour enjoyed a different question based format (remember Ask Jeeves) and before Adwords took its dominant position, the early movers were businesses now largely forgotten, namely Overture and E-spotting. Google quickly squashed them leading one to consider whether it is best to be a first mover or launch a superior product soon afterwards as a second mover. Many argue that Google Home’s smart speaker is a superior product but importantly Amazon got there first with Alexa. History, especially that of the mobile, is peppered with examples of devices and technologies being pioneered by one player (e.g. Nokia, Microsoft, Palm etc) and then these businesses fast losing their dominance to a product with superior user experience (e.g. Apple, Samsung).

According to Google’s own forecasts, by 2020, voice search will account for 50% of all searches globally. Google has rapidly pursued Amazon with its Home and Mini products enjoys a decade lead on Amazon in terms of understanding how to build an ad business and engage with brands and agencies. Conversely, Amazon provides direct access to consumer product and as a result has established relationships with manufacturers and retailers.

Amazon is notoriously difficult to navigate from an advertising perspective. Advertising is a rounding error in terms of revenue for Amazon today whereas for Google, it is their bread and butter. There are essentially two divisions that broker advertising/partnerships, which operate very separately; Amazon Media Group (AMG) which does bigger ticket brand integration versus Amazon Media Services (AMS), which is an auction, based, search type model. It is the latter that is more akin to Google’s core yet today perceived wisdom is that AMG takes more of the revenue.

So the big question is; can the pivot to voice provide the foundation for Amazon to beat Google at their own game, much like Apple beat Microsoft and Nokia in the race from feature phone to smartphone? Will Google’s expertise and focus on being an ad business give it the advantage or does the end-to-end marketplace and commerce proposition that Amazon has, favour them?

The reality is that both remain frenemies for brands – they have monopolies on distribution, whether that be distribution of content or commerce – but Amazon is arguably the fiercer force - overnight their insight through transactional data, allows them to effortlessly start up in a new sector as a competitor and steal market share, not just marketing dollars. Today, Google has a monopoly on search intent data and with such dominance in the search engine market share, is able to manipulate the economics of the auction for marketplaces that feed brands. Their analytics business and Doubleclick ownership means they do see the link between search and sales, however Amazon knows precisely what you have bought and your value as a customer. Amazon Prime was a masterstroke from Bezos, providing Amazon with an on-demand pipe into millions of households for ”stuff.” In a world of voice driven search, I would argue that transactional data provides Amazon with distinct advantage.

In my research for this piece, I reviewed a number of nascent studies on the emerging discipline of voice search. The first study from Kenshoo showed that whilst more people use Google search and shopping services for product discovery, 56% of all product searches start at Amazon. Perhaps, somewhat unsurprisingly, another survey from Bing, Microsoft owned competitor to Google in PPC, found that conversely 70% of retail based searches happen on a search engine like Google, Bing or Yahoo. It’s early days when it comes to voice based buying; 70% of Alexa “skills” (applications integrated in to voice platform) have just one review and retention beyond two weeks is said to be extremely poor. A Business Insider Intelligence survey of 1,000 users in 2017 found that only 9% of Alexa users had used voice commands to buy a product. These are low statistics but no doubt, before it scaled, Google faced a similar challenge – behavioural change takes time but we often underestimate how fast it can scale, if it delvers consumer value.

Unlike text-based or even image-based search, voice necessitates that the number of options served up will be far fewer. The concept of pages and pages of listings vanishes in this new paradigm. Think of a use case; you ask Alexa to “buy washing powder” – Alexa could respond by either asking ”which brand would you like to buy? ” or instead bypassing wide choice directly to “would you like to buy Persil or Fairy?” Much like Google’s quality score, an algorithm makes the decision on which brands are prioritised, so can Amazon but in a more compressed context. Today, Amazon uses its Choice algorithm that connects to its vast database of user reviews, previous product purchases and most popular purchases to determine which product to promote. Much like Google or Facebook, no marketer will ever see the precise methodology so undoubtedly the other factor considered would be the margin that Amazon makes on the sales of each product.

So everything is the same but everything is different. True disruption would be delivered if Amazon genuinely returned the products with the best recommendations and best prices for the customer searching, And that could still happen. Irrespective of that, this new paradigm requires a brand to think very carefully about the implications. If decisions are made on the best “value” products (in the same way as Google promises to prioritise quality content) then this forces brands to ensure that their products and customer service is flawless – a very good thing for the consumer but will create challenges for brands. If however, Amazon chooses to allow its biggest partners to circumvent the purity of the algorithm, and pay to be a preferred option (again much like Google Adwords) then a brand may be competing to be one of less than a handful of options.

As a marketer, the nirvana would clearly be a transparent model where quality and value are rewarded. Amazon’s recommendation engine delivered that exact disruption in e-commerce – by providing the fastest customer experience, breadth of product and reviews from real people, it became the go-to for buyers. However, much like Google, we all know that behind the scenes, Amazon is ruthlessly obsessed with winning market share and toppling competition. Amazon becomes God, determining the rules of its own auction.

Amazon has been rumoured to be talking to a number of big brands including P&G recently about how this model plays out. It is no surprise that the FMCG’s will be the first port of call as they are priority customers for Amazon across the piece. Brands that do not sell their products on Amazon will find themselves further down the pecking order. With fairly limited resources in advertising, Amazon will prioritise the biggest opportunity to grow revenue having written the textbook on growing adjacent revenue. As the Amazon Prime and Amazon Video subscription bases grow, these also offer huge value to brands. Amazon will continue to invest in original content production (media including the Kindle is already 18% of their revenues) to win attention further up the funnel. Combine this with their product marketplace and it would be unwise to bet against them.

For me, this means that whilst we will see head count growth in their media divisions, Amazon will first leverage direct brand relationships and likely also the management consultancies who focus on the broader digital transformation, outcome based initiatives. Marketers and agencies will continue to be frustrated at how little data is shared but I predict that they will invest in proving effectiveness in an original way. They have the opportunity to leverage the visibility of the transactions seen on their own platform to crack attribution far better than Google or Facebook.

So what does this mean for marketers and agencies? One thing is clear to me. We must learn the lessons of the past. Those that experiment and learn early, much like the SEM’s that built their business around the Google stack, will gain a competitive advantage. Amazon will increasingly go down an auction and programmatic route so brands will either need to invest in these skillsets in-house or work with agencies to support them. The Amazon ad stack today is nascent and clumsy but with engineering focus that could be changed overnight. Google and Facebook had to go on the journey of listening to marketers and agencies before they built dashboards, invested in audience measurement and insight teams. Snapchat is on that same journey right now and Amazon will have to do the same. At a time when agencies are in an existential crisis, does Amazon choose to invest in converting the agency community or instead concentrate on building out their direct-to-brand contact? Agencies should have cause for concern because Amazon needs agencies far less than Google and Facebook do. All marketing professionals, whether client or agency side, must double down on data and martech experience to have any chance of mastering Amazon. For me, Amazon will represent the next iteration of (big) data driven marketing – it has mastered cloud computing and AI but with a direct focus on commerce. No doubt, in their mind, injecting messaging in to the customer journey, whether that is whilst watching content or searching to buy a product, is a fundamentally simple mathematical and technical endeavour. Amazon will have significant power to manipulate the paths that prospective buyers take.

So what does all this mean? I predict that Amazon will have a few false starts before it succeeds, but as the kings of experimentation, it relishes failure as a route to learning, so it is just a matter of time before it wins share. Google and Facebook both face their own challenges in terms of brand safety and filter bubbles, whilst Amazon can remain laser focused on shifting product. Whilst 2018 will see some rapid progress in this regard and some big bets from Amazon, I believe this transition (particularly the pivot to voice) will take 24-36 months. Doug Anmuth, an analyst from Morgan Stanley has a punchier forecast, predicting that Amazon will reach $4.5bn in ad revenue this year and just shy of $7bn in 2019.

So Amazon is already on track to become the third largest scaled digital ad platform, but could it trouble Google and Facebook further? As marketing continues its direction of travel towards being a more data driven, and systematic discipline focused on outcomes, Amazon is arguably best placed. Google remains hugely reliant on its search business for revenue and no one has yet come close to toppling that empire, but that was pre-voice. Given Zuckerberg’s announcement last week, Facebook appears to be seeing cracks in its ad revenue model and will be on the defensive for most of the year. If I were sitting in the boardroom in Seattle right now, I would fancy my odds. The real question is therefore “Alexa, how seriously are you taking competing with Google and Facebook ?”

Paul Frampton was chief executive of Havas Media Group UK and is a regular contributor to The Drum.