Navigating Amazon’s River of Data

The online marketplace is not just a selling platform, it’s a powerful search engine used to deliver insights on SEO, product configurations, and more.


E-commerce activity in the United States reached a tipping point last year, according to the Commerce Department and Internet Retailer, with online sales taking a double-digit share (10.6 percent) of 2015’s more than $3 trillion in total retail sales for the first time, up from 9.8 percent in 2014. Total North American e-commerce sales reached $341.7 billion in 2015, and one company—Amazon—claimed a whopping 28 percent of that figure.

With a presence in the United Kingdom, Ireland, France, Germany, Italy, Spain, the Netherlands, Australia, Brazil, Japan, India, and Mexico in addition to the United States, Amazon generated worldwide revenues of $107 billion in 2015, from more than 304 million registered users and 55 million Prime members. And with more than 1 billion transactions processed per year, the company is not only a dominant e-commerce storefront, but also an ideal source of big data.

Amazon is a data-driven company. It collects data and uses it to understand consumer demand, popular product categories, price elasticity, demographics, sales trends, product development, keyword behavior, and customer feedback. And while it can be a challenge for third parties to access, data from Amazon can be useful to marketers in managing business on its marketplace and in other distribution channels.

Insights from Order Data

When you sell on Amazon’s third-party marketplace, customers buy directly from you. That creates a feed of orders containing the exact time of purchase and the location of the purchaser, as well as other important information you can use to improve your strategies. Here are just a few of the indicators marketers can track using Amazon data.

Advertising attribution. If you buy television media, you probably also use a media attribution platform to measure the effect of each ad spot. These systems typically rely on orders from the sites and phone numbers included in the ad spots. Consumers don’t always use the sales channel specified in the ad, however—they will check their favorite retail and online stores to see if the product is offered where they prefer to shop. And for an increasing number of consumers, the first choice is Amazon.

A media attribution platform is only as good as the data coming into it, so marketers need to integrate their Amazon order feeds into those platforms. Adding orders from Amazon will improve your software’s ability to analyze ad campaigns, optimize creative messaging, and right-size media buys.

PR attribution. Your public relations team is placing products in magazines, TV shows, and blogs to drive brand awareness and increase sales. With “earned” media, calls-to-action can’t always instruct consumers where to buy, so a visible web presence is critical. The impact of a successful PR placement varies widely among sales channels.

With one recent placement, the product landed exposure on Good Morning America. Sales on Amazon increased 600 percent in the following three days, and 100 percent on the brand’s own website. But the company wouldn’t have known what produced those extra sales without access to Amazon order data.

Attribution of a sales promotion. Most consumers interact with a product in multiple channels before purchasing. If you offer a discount in one channel, it may increase revenue in that channel and decrease revenue elsewhere. To accurately evaluate the effect of the promotion, you need to measure its results across all channels. This is particularly important in online channels, since it is so easy to compare prices from multiple websites at the time of purchase.

One of our clients recently ran a three-day sale on its website. During the sale, revenue on Amazon dropped 70 percent, and it returned to normal levels after the sale ended. While the promotion produced the intended result, its true economic value was much lower than expected because the additional revenue generated at the website was offset by lost revenue on Amazon.

Order volume and SKU variations. When a company rolls out a new product line at retail, one common issue is to forecast the correct quantities of each variation—size, color, pack size, etc.—to ship to stores. Bricks-and-mortar retailers have limited shelf space, so ordering the correct amount of each variation often improves revenues significantly.

On the other hand, shelf space in e-commerce is virtually unlimited. To get a sense of how popular each product variation will be at retail, marketers can offer their full product lines at their websites and on Amazon for several weeks before launching retail distribution to assess consumer preferences, including regional differences. Retail partners will be thrilled to receive the proper quantities of each product variation and will reorder more quickly, since they can avoid overstocking unpopular versions.

Keyword data. With a 44 percent share in North America, Amazon is the nation’s largest product search engine—and marketers should pay attention to the keywords consumers use when shopping with it. But the only way to access Amazon’s keyword data is to sell through Amazon Retail, and that requires marketers to surrender control over product pricing, content, and support, as well as order data and insights.

You can get the best of both worlds by selling through Amazon’s third-party marketplace and leveraging Amazon Sponsored Products. Sponsored Products is a keyword-driven, pay-per-click (PPC) platform that serves ads inside Amazon.com and drives traffic to your listings. Third-party merchants can use Sponsored Products to measure keyword performance metrics such as units, sales, click-through rate, conversion rate, and average cost of sale.

Optimizing keywords on other platforms. Running a PPC strategy in Amazon is a great way to generate extra revenue, but it’s also a great way to optimize PPC campaigns on platforms such as Google’s Adwords. Combining data from multiple PPC channels is especially valuable for long-tail searches, but it can take a lot of time to generate significant data. Combining data from multiple PPC platforms allows you to optimize your campaigns more quickly.

Amazon also offers an auto-targeting program where its system picks the keywords it thinks are most relevant to your products. This can help a PPC specialist discover new keywords to bid on in other channels.

Optimizing descriptions and SEO. PPC has always been closely tied to organic SEO, and since Google doesn’t provide organic keyword data, PPC is the only source of accurate keyword performance data. Provide your merchandising and copywriting teams with the top-performing keywords from Amazon Sponsored Products, and they can turn it into more revenue from organic search channels. Assign great importance to the keywords that perform well on Amazon; searches in Amazon carry clear buying intent.

Customer feedback and product development. There are three ways to collect customer feedback through Amazon: customer reviews, product questions, and order feedback. Collecting customer feedback from these sources is an easier and more structured source of customer sentiment than social media, and is cheaper than running focus groups and customer surveys.

To quickly generate ideas for improving the quality and features of a product, read the reviews and questions your customers write on Amazon to see what they really like or dislike about those products and those of your competitors. One client that experienced high numbers of returns discovered that its stitching was falling short of customer expectations by reading the customer feedback on Amazon; it improved the quality of its stitching— and lowered its return rate, as well.

Questions and product descriptions. Customers often have a question they wish could be answered before they purchase a particular item, and in many cases, they will put off a purchase or avoid making the purchase at all if they can’t find the answer immediately. But in some cases, they’ll reach out to ask a question.

Amazon offers a system for collecting and answering customer questions before and after a purchase, and it publishes those questions on the product detail page for anyone to answer. On the Wolfgang Puck Pressure Oven page, for example, customers often have a specific dish they want to cook and need to know if a particular model is right for the job. “How big a turkey can actually fit in this thing?” one shopper asked. A user responded, “A 13-lb. turkey did fit—best turkey I ever cooked!”

Offer data and distribution problems. Monitoring the offers for your products on Amazon can tell you a lot about the health of your distribution channels. If you have an interest in controlling distribution, want to keep prices profitable, care where consumers buy the product, and want to ensure good post-purchase support, Amazon is the first place you should look. Anyone—even unauthorized sellers—can list a product on Amazon. Monitor product offers to find problems in your distribution chain.

Amazon is also a good place to find sellers who might be hawking counterfeit versions of your product. Let’s say you sell a product for $6.00 at wholesale, and someone is selling it on Amazon for $6.50, including shipping. With the cost of fulfillment and Amazon’s 15 percent fee, that seller is either losing money on every sale, or they sourced counterfeit product at a cheaper rate than your wholesale price.

The Seller’s Advantage

The data you get from Amazon (and how much) depends on how you sell your products on Amazon. As a wholesaler,  you have limited access to the data, since Amazon sells the product directly to the consumer and it owns the data. But selling as a third party allows  access to more data because you can reach the consumer directly using the Amazon.com platform.

Ranjit Mulgaonkar is founder and CEO of DNA Response and a veteran of Amazon’s e-commerce services, and Nathan Grimm is director of marketing and product, Grimm leads the e-commerce and corporate marketing teams. Contact DNA Response at (206) 995-8080.