Algorithmic Bidding: Looking Back at the Holidays
To no one’s surprise, the holidays flew by this year. As advertisers, most of us were happy to see them go. The holiday season can make or break a company’s year, this is particularly true for those selling products online. Expectations for commerce performance are higher than ever – especially with the ability to track the full-funnel journey from first click to purchase. Compounding the pressure, the core shopping season was a full week shorter in 2019 compared to the year prior. The tightened timeline was expected to hamper sales for most retailers.
The time spent managing campaigns on Amazon, Google and Bing always increases notably in Q4, with constant optimizations needed to ensure successful results. For search marketers, the list can be daunting. The need for keyword, budget and bid adjustments skyrockets. There’s hardly enough time to make every necessary optimization, and there exists a fine line between maximizing returns and pushing sales volumes, thus companies who use automation to tackle the small decisions have more time to make the big decisions.
Machine learning and algorithmic bidding are a reality for most advertisers these days. At Empower, we embrace and harness machine learning and automation. With the flurry of the holiday season now behind us, let’s review the do’s, don’ts and lessons learned around automation for Amazon Advertising and Search.
Turn Off Bidding Folders
During the rush of Black Friday and Cyber Monday, or even leading up to Christmas Day, it’s important to turn off third-party bidding. The majority of third-party platforms only update bids once a day and have trouble reading the swings in the market. A certain level of seasonality is built into each algorithm, but one can’t expect it to fully adjust to the season’s relative unpredictability.
In the case of manual bidding, if your team goes in and makes multiple keyword-level bid changes, active bidding folders will override these adjustments based on a historical data set, as opposed to recent surges. Most algorithmic bidding folders have guardrails to curtail drastic changes, but the holiday swing can be so extreme that these guardrails can hamper performance.
Spikes in online traffic can be unpredictable, this year in Amazon we saw a drop in volume between Black Friday and Cyber Monday of over 50% across our commerce client base. During that two-day period – to match these drops – many manual bid decreases were needed. Without pausing algorithmic bidding, it’s likely that bids would not decrease fast enough – causing excess spend and inflated cost-per-click during some of the busiest shopping days of the year.
Historically-driven, one-off algorithmic updates are just not enough during the holidays because performance can flip quickly. In some cases, it may be better to ignore certain dates for bidding folders. Most paid search and commerce platforms have the option to set blackout dates in which the algorithm will not use data to inform bidding. Simply put, turn off folders to make room for manual changes.
Using Automated Rules
Setting up automated rules can help during these chaotic timeframes. Google, Bing and third-party solutions like Kenshoo and Marin have automated rules that can make automatic adjustments or notify users when certain conditions are met. This past holiday season, Empower used automation to make daily budget updates to clients’ most popular campaigns, those meeting a minimum ROAS were automatically increased during peak times. The same can be said for campaigns falling below ROAS goals or spending without converting, which were decreased automatically via automated rules. Keyword-level changes through automation can be more difficult as campaigns likely have different objectives. Using structured bid changes for keywords with similar cost per conversion will allow for more control.
The 2020 holiday season is a ways away, but it’s never too early to understand the technology and prepare campaigns. Through a combination of automation and manual changes, you can maximize performance in topsy-turvy purchasing windows including, but certainly not limited to, the holiday shopping season.