Hidden secrets of The Amazon – your heard it here first folks!
Conversion Optimisation

Hidden secrets of The Amazon – your heard it here first folks!

7th April 2014

Admit it, if you work in ecommerce or tech the chances are you are that you’re a bit obsessed with Amazon. Their scale, their tactics, their dominance, their enigmatic CEO Jeff Bezos: it’s all fascinating. Not the least because Amazon is reasonably secretive about what goes on behind the scenes. Luckily for attendees of ClickzNY, Bryan Eisenberg, long-time Amazon “watcher” was on hand to talk us through some of his observations about what makes the retail behemoth so great.

“The most important single thing is to focus obsessively on the customer. Our goal is to be earth’s more customer-centric company.” Jeff Bezos


Eisenberg’s interpretation of this quote is more nuanced than simply ‘the customer is always right’, he believes that this quote speaks to Amazon’s intense focus on personalization for the customer and that this is the secret to their current success. The good news is that most businesses can employ some of Amazon’s tactics themselves.

Amazon’s Four Pillars:

1)    Customer Centricity

2)    Continuous Optimization

3)    Culture of Innovation

4)    Corporate Agility

Customer Centricity

Amazon believe in data.

Data about what their customers are doing on their pages, every mouse click, every hover, every page viewed, every product viewed then rejected. They then apply this data to your next visit, improving the experience for you in ways you won’t even notice. Paraphrasing from The Everything Store, ‘Amazon is in the business of helping a customer buy a book, Amazon is not in the business of selling books’. Again, this is another rather nuanced interpretation but this means Amazon uses all data for every product chosen or not chosen by a user and applies it across other users, thereby improving the overall experience. This tactic is working : 2/3s (66.8%) of Amazon’s traffic is repeat traffic.

Amazon then apply the data they have from their users about the products they do and don’t buy to inform their inventory selection. This is a marked departure from historical inventory sourcing in the book industry alone. Book shops’ and wholesalers inventory evolved through relationships with publishers and bulk deals and this was what ended up on the shelves for the consumers. Amazon reversed this by understanding the consumer demand and then sourcing based on that.

Building from this, Amazon went from being an online bookshop to the everything shop it is today. As ecommerce has grown and the technology that enables it has improved, so has Amazon’s reach and now they are competing with Walmart and other traditionally bricks and mortar retail giants. Amazon, despite not having any physical shopfronts still has one advantage over Walmart – Amazon knows who their customers are when they arrive on the site, they have their own data and also the data they can collect through their Amazon Affiliate Program, which has ~ 500,000 members. To put it another way, Amazon can see user data from another 500,000 sites in addition to their own very powerful domain. As of right now, physical retail analytics for the likes of Walmart are nothing like as comprehensive and the upshot of it is that Walmart doesn’t currently know who you are when you walk into one of their shops. Amazon does not have this problem.

Continuous Optimization

Amazon works hard to maintain their position in terms of attracting traffic whilst still offering a good price for their customers. In 2013, 30% of all ecommerce sales in the USA were through Amazon and they will work hard to maintain and improve this. One tactic is that of price changing: Amazon changes prices 2 million times a day across all of their product listings. The freshness of the page is just one factor of interest to Google when it comes to organic results so Amazon’s commitment regularly updated pages not only helps them control pricing and therefore their profits, it is also benefiting their visibility on organic search.

Their culture of continuous optimization extends to more than just pricing.

4 Key Inputs for Amazon:

1)    Selection: Amazon have 7 x the number of products as their competitors

2)    Price: Amazon strives to have prices 5 – 13% cheaper than their top 5 competitors

3)    Availability: Amazon turns their inventory oven 20 times a year. Most orders fulfilled by Zone 0-1 shipping. (ie very local)

4)    Experience: all departments are monitored in terms of customer satisfaction.

Culture of Innovation

These are just some of Amazon’s innovations:

Amazon Prime

Amazon Prime drives valuable repeat traffic – Amazon prime members visit the site more often and spend 3-5X as much as a non prime member. Even the recent price rise to $99 for the year doesn’t seem to have affected prime’s popularity.

Social Commerce

Amazon included user reviews before anyone else and pushed back on publishers who were initially reluctant. Whilst most retailers have caught up and now include reviews, few have introduced two features that are particularly useful, the ‘most helpful?’ button and the most critical filter. Both of these features allow users to sort reviews by the most helpful (that users vote on) or the most critical, two filters that drive a lot of conversions. When you factor in that 20% of people purchase products based on helpful reviews, it is clear that many retailers are missing out.

Amazon Associates 2006

Launched affiliate program Amazon Associates 2006.

The affiliates show banners, stores and have API access on their sites. As previously mentioned, Amazon drop a pixel on their affiliate sites which allows them to benefit.


Amazon were doing 200+ tests a month in 2004. There isn’t any more recent data about the frequency of their testing but it is conservative to assume that this rate of testing has at least stayed the same. Clearly, testing is taken extremely seriously however there are some rules. No-one is allowed to test on an Amazon product page above the fold without the express permission of Jeff Bezos. This final level of permission gives us some indication as to how carefully optimized the existing Amazon site is, based on this culture of constantly testing and iterating.

Research & Development

In 2013, Barnes + Nobles’ revenue was $ 6Billion.

Amazon spent $6.5Billion in research and development alone in 2013. Amazon is investing heavily in R&D and this investment, whilst expensive helps keep them ahead of their competition. It also allows them to enter new verticals and quickly become successful.

The image below puts Amazon’s R&D investments in context, namely they are competing with the tech giants in this area, not other retailers.


Image Source

Corporate Agility

Amazon, as with other agile technology companies, is set up rather like the internet itself, as Frank Rose suggests:

“In terms of organization, effective digital-age companies (for example, Apple)

tend to resemble the Internet – decentralized and interconnected”

Amazon’s teams are set up for execution, as demonstrated by their rapid MP3 store redesign to commemorate the death of Michael Jackson, which took less than 2 hours to go live.


Smaller companies can benefit from Amazon’s testing and implement the features that Amazon maintain, as this is a good indicator that it has been successful. And whilst you may not have Amazon’s $6.5billion budget to spend on R&D, there are a number of smaller companies, listed at the foot of this article, offering comparable products that can be used to improve your customers’ experience.

Amazon has one flaw:

Online commerce is only10% of all retail spending, the rest is real live shopping in bricks and mortar shops. At the moment, Amazon is only competing for 10% of all the shopping dollars out there, if you have a physical location alongside a good website you are in the best possible position to fulfill your customer’s needs online and off. At present, physical shops are not innovating at the same rate as online businesses. There are lots of established retailers doing interesting, successful and even humourous online marketing campaigns but the retail experience and feel of the physical shop does not match the online personality.


A bigger issue is that retail analytics for physical shops are much more primitive than their online counterparts. This is now slowly starting to change however, and by marrying your online and offline data you will be equipped to understand your customers and what they are looking for. RetailNext for example, can generate heatmaps of where your customers are in the shop, where they are congregating, which in turn can inform you where there are bottlenecks or other delays. Nomi is a newer service that helps you track users in the shop and, if a known customer, they can send email and text updates with coupons or other news. This physical retargeting echoes what we already see online and as the two worlds merge more closely together, businesses who apply the online analytics to their physical locations will be the ones to benefit.


Lots of 3rd party tools that mimic some of Amazon’s features:

  1. Monetate: takes 3rd party data and provides info about visitor. Retailers can buy this. You can then email and launch personalize email. Monetate helps you launch personalized campaigns very quickly and scaleably. It also offers visual searches, namely an image of the product for which you are searching.
  2. Sightly– helps retailers create video ads and personalize and localize them. They are YouTube partners. This helps businesses quickly scale video ads.
  3. Optmyzr – this product generate scripts to help improve ad quality scores
  4. Lattice – plugs into all sorts of databases to assess business expansions (eg signing a new lease) this information can then be relayed back to retailers who can then use this information to contact the expanding company for a product they might need in the event of this expansion.
  5. Iterate – a Colorado-based company that sources disruptive and innovative products for their retail partners.
  6. CommerceSciences – Little tag launches as people navigate your site, gauges their behaviour and attributes. You can then use this data to group users by criteria such as being cost conscious or security conscious. CommerceSciences will then provide information based on these buckets and what their system believes the user needs to finally convert. This can reportedly lead to a 25% – 30% lift in conversions.
  7. Kaggle – opens up data for a contest. Data scientists compete to surface outliers or anomalies with your data and solve your business problems.
  8. Nudgespot – Nudgespot is a very useful tool for small businesses as it helps you tailor and personalize all of your communications.
  9. Enigma – a useful site that pulls in reams of public data from NGOs, governments, companies and organizations.
  10. BloomReach – is a comprehensive product that allows you to personalize landing pages and more for your users, tracking them across channels and devices.
  11. Bazaarvoice – a tool to show user reviews. Widely-used.


Whilst Amazon’s domination can seem daunting, the correct response to such scale is to use Amazon as a case study and piggyback on their research to benefit your business. Keep an eye on their shopping cart language and internal linking products and see if any of them can be replicated on your site. If you are selling a more niche product, you have the luxury of tailoring your content to those users, something Amazon, in selling virtually everything, cannot do. Amazon may dominate 30% of all ecommerce in the USA, but that just means there is 70% up for grabs.


About the author

Sarah Kershaw is a search analyst based in New York who engages in freelance writing in her spare time, writing about trends in digital marketing, the future of news and fine art. As a search analyst, Sarah is interested in UX and IA and tends to get very animated when talking about fonts and colours.


Written By
This post was written by an author who is not a regular contributor to State of Digital. See all the other regular State of Digital authors here. Opinions expressed in the article are those of the contributor and not necessarily those of State of Digital.
  • This field is for validation purposes and should be left unchanged.