The Precision Playbook Series: A/B Testing Your Way to Profit
Win in the Amazon DSP. Is your investment working or are you overpaying for traffic?
In the digital marketing world, everyone talks about the tactics, the clicks, the impressions, the bids, the brand safety, etc. But a select few are playing a more advanced game realizing that the secret to winning isn't to be the loudest or to spend the most, but to be the smartest.
This series is for the marketers who are tired of the same old advice. It's for the ones who know that true value of a campaign is measured in profitability, not just traffic. I’m going to pull back the curtain on some of these that separate the pros from the rookies.
Each installment will challenge a conventional wisdom and provide a step-by-step guide to a tactic that will change the way you think about your campaigns. Hopefully you will be able to find the signal in the noise and prove your true incremental value.
Let’s begin.
Win in the Amazon DSP: Is your investment working or are you overpaying for results?
In the wild programmatic west of the Amazon DSP, most advertisers are locked in a bidding war. The real pros are here playing a much different game realizing that the biggest advantage isn't a bigger budget but a better buying strategy.
An out-of-the-box and simple way to learn if you’re Amazon DSP campaign is worth it, is to run a low-bid, control campaigns alongside your main, high-bid campaign. This is something too many media buyers overlook, as their focus is almost always on winning the most competitive auctions.
The point is to solve a key business question: Is this campaign efficient and effective? This strategic test is not just optimizing bids. It’s about holding our strategy accountable and forcing the algorithm to find new pockets of efficiency. It ensures every dollar spent is working as hard as possible.
Set Up a Low-Bid Campaign
Create a second campaign with the exact same targeting, creative assets and landing pages as your primary campaign. The only difference is that you set the bid to the absolute minimum you can. This low-bid campaign is not meant to be a primary driver of sales here but more of a silent observer.
The idea is to set up two identical campaigns that compete against each other. Something I always frowned upon as a rookie move. But stay with me. Promise it will all make sense.
Primary High Performance Campaign uses your typical, aggressive bids to capture valuable traffic and drive sales.
Control Campaign (Low Bid) uses the lowest possible bid. Usually the floor bid on the Amazon DSP. You want to capture the least desirable, cheapest traffic.
By running these two campaigns side by side, you can see if your high bids are actually driving new, incremental value.
Here is a step-by-step guide to get started:
Start with your main campaign, then duplicate it to make sure the targeting and audiences, placements, etc are identical.
Give the low bid campaign a smaller, but still meaningful, budget. You don’t want to starve it because enough budge allows it to run consistently and gather enough data for a fair comparison.
This is the most important step: Lower the bid on the duplicate, control campaign to the absolute minimum allowed by the platform.
Fun part. Let both campaigns run for at 2-4 weeks at least. For smaller budgets, you may have to wait longer for actionable data. Then compare the performance metrics of the two campaigns: ROAS, eCPA, ACOS, CPC…
The difference in performance between the two campaigns shows you the incremental lift of your high-bid strategy. If your high-bid campaign has a significantly better return on ad spend (ROAS) than your low-bid campaign, you know your investment is working. If they have similar returns, you’re probably overpaying for traffic you could get for much less.
This simple test works for two key reasons. 1) It forces the algorithm to find new pockets of efficiency, and, 2) It provides a clear measure of incremental value.
The Amazon DSP's algorithm is a machine learning system designed to find the best opportunities within your budget, theoretically. By forcing it to work with a very low bid, you are essentially telling the algorithm: I need you to find me the most unbelievably cheap, yet still valuable, inventory you can. The low-bid campaign will find a long tail of highly efficient placements that your main campaign, with its higher bid, would never even see because it's always focused on the more expensive, competitive inventory.
By comparing the performance of your low-bid campaign to your high-bid campaign, you can see if your higher spending is actually driving truly incremental results. If your low-bid campaign delivers conversions at a $10 cost-per-acquisition (CPA) and your high-bid campaign delivers conversions at a $25 CPA, you can see the premium you're paying to win more volume.
This allows you to make smarter decisions about your budget and whether the extra spend is worth the additional conversions.
“Okay, but what about crap inventory”?
Buying low-value inventory from sources like background apps is one of the most common ways to waste ad spend. The key is to move from a reactive, defensive strategy to a proactive offensive one.
Here are your usual suspects to safeguard against pure junk:
Instead of wack-a-mole or trying to block every bad app and site in the world, the most effective strategy is to only bid on a pre-vetted list. You build a targeted list of safe apps or sites that ar relevant to your audience along with choosing only 2-3 supply vendors. You then set your campaign to only bid on inventory from these specific placements.
Why this works: It removes the risk of your ads showing up on fraudulent sites or apps you've never heard of, ensuring that every impression is on a placement you've already approved.
While inclusion lists are great for a focused approach, you still need a system to manage the open exchange: Block Lists
Why it works: Background apps and garbage pubs often have super high CTRs or low viewability. These are other huge red flags too.
So while everyone else is still fighting to win the most expensive impressions, you'll be the one quietly uncovering the most valuable opportunities in a place they'd never think to look.
A little bit of tactical discipline and a simple test campaign can be your most powerful weapons.
So, there it is. The first rule to leaving the bidding war behind.
This is the kind of stuff that separates the serious from the curious so thank you for reading and continual support. Now go ahead and put this to work and tell me what happens.
The truth is in the data, and I'd be honored to hear what you uncover.