Understanding the Challenges of Amazon Marketing Optimization

Amazon.com is crowded. Most Americans skip Google and go straight to Amazon when they shop online, and as a result it has become one of the key product marketing battlegrounds. Today, improving visibility, keyword dominance and digital shelf space on Amazon requires an increasingly high level of active, real-time campaign management, putting a lot of pressure on brands and marketers to rise above the din.

Marketing optimization is a key component of a successful Amazon strategy. Used effectively, it can improve the return on investment of your advertising campaigns while also taking the tedious, time-consuming manual labor out of the equation. That being said, you can't just slap a "Marketing Manager" title on an artificial intelligence algorithm and wait for the sales to roll in. Successful marketing optimization requires a multilayered approach, active input from human beings, and a thorough understanding of the challenges and limitations of the different technologies available.

Amazon as Marketing Platform

In the second half of 2018, Amazon began to roll out Amazon Attribution. This tool, similar to those used by Google and Facebook, allows advertisers (initially limited to first-party vendors) to measure the impact of ads based on how consumers discover, research and purchase products both within and outside of Amazon. Combined with Amazon's huge database of consumer shopping behavior, it will allow the e-commerce giant to target consumers with uniquely high accuracy and relevancy.

Amazon is already the third largest digital advertising platform in the U.S., worth about $125 billion, and eMarketer predicts a growth of 50 percent through 2020, which would bring digital ad spending on Amazon to about 7 percent of the U.S. total. This drives a sharp increase in marketing costs for vendors; for instance, keyword cost per click (CPC) has risen by roughly 56 percent year-over-year across all categories as of March 2019.

If you didn't already consider Amazon one of the key product marketing platforms, the fact that it's currently investing in and expanding its role as a major advertiser means that performance-driven marketing is no longer going to cut it. Similarly, the resources required to market effectively on Amazon are becoming unsustainable. This is where marketing optimization comes in.

Types of Marketing Optimization

Don't make the mistake of assuming marketing optimization is a one-size-fits-all solution to your e-commerce campaign management problems. A completely manual approach to selling on Amazon may require an unreasonable amount of hours in the day, however, relying entirely on AI algorithms also has its drawbacks. While it saves you the most time, AI can never do quite as good a job as a human being can.

Marketing optimization is a broad term that encompasses numerous tactics and technologies, each having their own strengths and weaknesses. In most cases, an effective approach will contain a mix of different tools along with manual intervention and regular campaign audits.

1. AI/Machine Learning

The big topic in marketing optimization today is artificial intelligence — i.e., the use of computers to mimic the cognitive functions of the human brain. Many marketing optimization tools use machine learning technology, a subset of AI which can receive a flow of data and "learn" for themselves, updating their algorithms based on new information. ML allows intelligent marketing tools to respond to changing market conditions, search behavior and purchase patterns, a crucial feature for maximizing profitability.

On Amazon, an effective AI tool needs to use various signals to build optimization models for keyword bidding. The challenge comes from the data required. Because of the nature of AI, its intelligence is a result of the amount of data it has. To respond to a trend, AI needs at least 15 conversions on a campaign per day, which may be challenging for longer tail keywords. AI can be effective for optimizing high-volume, high-frequency keywords, but secondary and long tail keywords (key for targeting specific consumer segments) typically don't get enough traffic.

2. Programmatic Rules

Striking a balance between AI and a fully manual approach, programmatic rules have the potential to be the most effective form of marketing optimization on Amazon, though they require a lot of time and knowledge to implement properly.

A programmatic rule executes actions based on pre-defined requirements or parameters. The user creates a logical "if/then" statement; one or more "if" conditions are assigned to one or more specific keywords, and if the keywords meet those conditions, the tool executes the "then" action, such as increasing or decreasing the bid on those keywords.

The purpose of rules, as opposed to AI, is that rather than letting a computer decide the best course of action, you're programming a tool to respond to certain conditions in the same way you yourself would. For example, you can bid on a certain keyword but set a rule to reduce your spending if it turns out to be a dud, freeing up budget for more efficient keywords. Unlike AI, however, a bad rule doesn't correct itself. Rules are only as smart as the person who programs them, and to define a really effective set of rules requires a lot of knowledge and experience with your target audience.

3. Dayparting

Perhaps the simplest form of optimization, dayparting at its most basic level is a form of scheduling, consisting of a timed "on/off" switch for your campaigns. Amazon doesn't include this option in the Advertising Console — all campaigns begin and end at midnight unless manually paused — so a third-party tool is required.

While it's still the norm to run campaigns 24 hours a day, dayparting is growing in popularity. It allows vendors to divide the day into several distinct parts and adjust their campaigns based on the time of day. Smaller vendors with limited resources may use their marketing budgets more effectively by focusing their keyword bids on high-traffic hours and reducing their spending during off-peak times.

Different technologies each have their strengths and weaknesses, and an over-reliance on one form of marketing optimization is never going to be very effective on Amazon. Fully relying on AI algorithms to manage your campaigns may reduce the amount of labor required, but may not produce the desired outcome depending on your goals and KPIs. A strong multifaceted strategy that mixes different types of marketing optimization based on their unique applications and uses human intelligence to oversee every element is key to a competitive Amazon marketing campaign.

Melissa Burdick is the president of Pacvue, an e-commerce SaaS platform that empowers brands, sellers and agencies to programmatically manage and optimize their advertising on Amazon.