How We Combined AI with the Human Touch to Achieve 30% YoY Growth for Our Client — Part II

We increased YoY top-line Amazon sales by 30% for one of our clients during 2023’s T5 period – Turkey 5, the five biggest sales days of the year from Thanksgiving to Cyber Monday.

In Part I of this series, we delineated the technology we used to achieve this.

But technology alone would not have created such impressive results. In Part II, we’re going to describe the human touch that caused this campaign to excel so magnificently.

The Human Element

For this premium brand, we needed to ensure that we drove the market dynamic to our advantage.

Amazon Advertising is auction-based. If your bid isn’t high enough, you won’t win the placement. If it’s too high, your ACoS — Amazon Advertising Cost of Sales — will be too high and your ROAS (Return on Ad Spend) too low. If this gets out of control, you could exhaust your budget and end up in the red.

Because T5 contains such a tidal wave of sales and traffic, waiting even a few hours to react to minute changes can result in the loss of significant amounts of revenue, new customers, and returning customers for large advertisers. This is where the human element comes in — to react fast enough to data-driven metrics so as to maximize ROAS for the client.

Manual searches for core keywords

The first step of our strategy was to conduct human searches for core keywords at a regular cadence. We wanted to determine if our brand was showing up on TOS — Top of Search, Amazon’s prime real estate placement for product searches.

Core keywords are the main keywords for your products that drive the majority of traffic. There are usually no more than a dozen of them.

It would be impossible to manually monitor hundreds of keywords every hour — automated bid strategies can help with those keywords. But core keywords are so crucial that we chose to monitor them manually and take action within minutes.

Conducting manual searches allowed us to obtain immediate insights, about two hours faster than Amazon Marketing Stream.

If our core keywords weren’t at Position #1, we immediately increased bids for those keywords to keep them at TOS.

Daily budget targets + pre-game campaigns

Before T5’s kickoff, we had already prepared several “backup” campaigns and targets that we could leverage if we were not hitting our daily budget goals. Because we prepared these campaigns all pre-game, all we had to do was activate them if we weren’t hitting our hourly budget targets.

There are certainly nuances, but this is a fairly linear strategy. If, by midday, you’re far from 50% through your advertising budget, you need to take action.

These pre-prepared campaigns are typically competitor ASIN targeting campaigns that usually see a lower return but can improve awareness, particularly if the products are on promotion.

Measuring success — Marketing Stream data lag

Amazon Marketing Stream data lags by about two hours so it definitely isn’t real-time. But analyzing hourly data from other days gave us insight into:

• At what hours of the day traffic ramps up
• At what hours of the day ACoS improves
• When traffic ramps down
• When ACoS deteriorates.

Based on this data, we were able to make fairly good assumptions about the T5 traffic patterns.

But it still wasn’t enough.

Although we set up our campaigns to increase and decrease bids according to the expected daily and hourly traffic patterns, we additionally performed regular manual checks to discover:

• Were we winning TOS for our core keywords? If not, we adjusted bids based on this data set.
• Were we within the ballpark for our budget? Did we need to bid more aggressively?
• What game-time decisions were needed? Did we need to feed more budget to a campaign based on performance thus far in the day?

The results speak for themselves

By leveraging technology and combining it with the human touch as described above, we were able to win core placements 65% of the time for our client. And by ceaselessly monitoring ad spend, we were able to come within 1% of the client’s budget.

The result was a YoY revenue increase of over 30% in ordered product sales.

Modern tech and AI is about how you use it, but it’s no replacement for human acumen. Those who are going to win with tech are those who learn to combine its strengths with the strengths of human sentience and excellent pre-game planning.

To find out how PointStory can optimize your Amazon advertising campaigns and increase your ROAS by combining tech with the human touch, contact us for more info!