How Often Should You Be Optimising Your Amazon Sponsored Ads?
If you’ve ever run an Amazon Sponsored Ad campaign, you’ve probably wondered: how often should you optimise your campaigns? With the surplus of information advising different strategies, we wanted to have our own dive into the data.
At Old Street Media, we reviewed over 2M keyword & product targeting IDs from over 2,000 vendors and sellers to create a set of best practices that will help you optimise your bidding cadence. In this article, we’ll break up our findings into four sections: lookback window for decreasing bids, lookback window for increasing bids, bidding frequency, and final recommendations.
Lookback Window for Decreasing Bids
In general, decreasing bids will grant you less desirable ad placements which will lead to fewer impressions, clicks, and cost-per-click. Advertisers lower bids to improve ACOS or to lower ad spend. We quickly found, however, that most advertisers are taking action too quickly and missing out on sales as a result.
It’s important to remember when decreasing bids that some keywords may eventually convert a sale if you wait for the 7 day or 14 day attribution window to complete. We found that advertisers are prematurely lowering bids on 34% of their keywords and actually killing up to 8% of their unique keywords which may have otherwise led to a sale.
So how do we make sure that we’re not shrinking our valuable keyword pool? It completely depends on your budget and risk profile.
We suggest completing the 14 day attribution if you have a flexible budget and are looking to make highly informed decisions with complete data.
We suggest completing the 7 day attribution period if you’re okay prematurely decreasing bids on 13% of keywords due to cost or profitability constraints.1
In rare cases, we may make immediate changes if you’re okay prematurely decreasing bids on 34% of keywords. Given the materially negative impact on sales, we would avoid this scenario unless it’s absolutely necessary.
Lookback Window for Increasing Bids
Increasing bids will increase your chance of winning higher visibility ad placements which will lead to more impressions, clicks, ad spend, and sales. This strategy is used to increase sales, but it often comes with a higher CPC. It’s always important to track your conversion rate against the increase in CPC to evaluate whether higher visibility ad placements are helping you achieve your goals.
There’s minimal risk of missing out on sales when increasing bids. As a result, we suggest evaluating your high performing keywords and increasing bids based on the prior week’s data.
For example, if today is November 9, 2020, we suggest reviewing data for the week ending November 7 (previous week). This strategy will maximise your sales on high performing keywords by improving your ad visibility for sales-generating keywords.
The risk of increasing your bids is that you might spend more than you’ve budgeted for. Two simple strategies to minimise this risk is to set campaign-level budgets or to daypart your campaigns. Dayparting helps you ensure that your campaigns are live as CPCs decrease throughout the day. We found that 80% of out-of-budget campaigns spend their budget by 11AM, which means that CPCs will significantly decrease as the day progresses.
Bidding Frequency and Other Tips
While your look back window may change based on your risk profile and budget, we suggest optimising your bids regularly. We do not recommend updating your bids daily because ad performance can have a high day-on-day variance. Rather, we recommend looking at a full week of data.
As a first rule of thumb, we generally advise against updating your bids more than once a day. The Amazon Advertising console and API data do not reflect real-time data and have 6+ hour data delays. Furthermore, Amazon has a 72 hour data validation window where they may confirm attribution and click legitimacy, so this data may change.
Our clients also often ask if average selling price (ASP) materially impacts consideration length and attribution windows. We have done extensive analysis on how attribution differs based on ASP.
As shown in the graph below, there is higher 1 day attribution for lower priced products compared to 300+ (USD, EUR, GBP) priced products by ~4 percentage points. However, we found that the differences aren’t big enough to warrant a difference in bidding cadence strategy.
Wait 7-14 days to decrease bids on keywords except in rare cases to minimize the risk of losing out on keywords potentially converting Amazon sales.
Be comfortable increasing bids before the full attribution window is completed to maximise sales.
About Old Street Media
Old Street Media builds automation software and provides marketplace & advertising services for Fortune 500 brands, challenger brands, and media agencies. We work with over 2000 brands and have helped generate 400M in sales in the past year.
Other relevant research & data
One third of keywords that register sales register additional sales after the 1st day. This decreases to 13% by day 7.
Of keywords that generate sales, 79% of keywords generate its total 14 day attribution on day 1 and 89% of keywords generate its total 14 day attribution by day 7.
96% of keywords that register no sales after 1 day will not register a sale by the end of the 14 day attribution period.
By the 7th day, 99% of keywords that register no sales will never register a sale.
Of the 4% that registered no sales on day 1, 79% of the 14 day attribution will be registered within 7 days.
We tried to identify common themes of keywords that have 0 sales on day 1 but lead to an eventual sale by day 14. We found that higher clicks have a higher chance to convert an eventual sale. The average clicks of keywords that lead to an eventual sale are 6 compared to 2 clicks where keywords don’t lead to a sale.
Ad performance variance decreases by 44% when looking at a full week of data compared to daily data. Lower variance means that you’re making more stable and higher confidence decisions.