Three well-meaning paid search optimizations that might kill performance
That recent adjustment to your paid search campaign seemed like a good idea at the time, but now you’re seeing a drop in performance. How could this happen? Columnist Andy Taylor notes some common issues that arise from well-intended optimization choices.
I only recently learned of a Wikipedia page dedicated to inventors who were killed by their own inventions. That page is amazing, and if you end up leaving this article and just reading that list for the next half hour, I totally understand.
But it got me thinking: what are some ways in which paid search managers end up killing their own campaigns as a result of their own optimization attempts?
Bid increases in budget-restricted campaigns
Optimally, paid search campaigns are allowed to spend as much as efficiently possible, with efficiency judged by the return on ad spend (ROAS) generated and how that aligns with profitability. However, many advertisers are constrained by necessary budget restrictions, and use campaign budget settings in AdWords, Bing Ads and Yahoo Gemini to prevent spend from going above their spend limits.
The downside of this management style is that any keyword bid increases actually reduce the total amount of traffic a campaign can garner if it is already being held back by budgets prior to the bid changes, since the price paid per click increases with higher bids. Greater cost per click means it takes fewer total clicks to reach budget caps.
Thus, bidding more aggressively might get ads further up the page, but traffic would actually go down, since budgets would prevent the ads from showing as often. I’d say this qualifies as an optimization that ends up killing performance.
Limiting paid search campaigns to only target RLSA or Customer Match lists
Remarketing Lists for Search Ads (RLSA) and Customer Match audiences provide advertisers with the ability to target ads to high-value groups who have either visited a brand’s website or provided the brand with their email address, respectively.
With such demonstrated affinity for a brand, these audiences naturally outperform non-RLSA and Customer Match audiences when it comes to measures like click-through rate (CTR) and conversion rate, as shown in many case studies.
As such, advertisers might be inclined to only target these audiences in search, and indeed such a strategy is a necessity for some brands in highly competitive industries, since they simply can’t afford to bid enough for keywords to reach non-RLSA and Customer Match audience searchers and still maintain an effective ROAS. Outside of these rare situations, however, targeting only users familiar with your brand has obvious consequences.
For one, most advertisers use non-brand keywords in order to attract new customers. Limiting ads to just users who have already been to a website or provided their email address eliminates the possibility of reaching customers that are totally unfamiliar with a brand.
Additionally, non-brand search ads help to grow RLSA and Customer Match lists by adding new users who are likely to click/convert further down the line. Thus, turning ads off to non-audience members may result in RLSA and Customer Match lists slowly dwindling in size over time.
And while advertisers could potentially use other channels in the hopes of growing site visitors and email lists, search query is often a very strong signal of who will ultimately buy from you, compared to weaker signals from other channels like paid social.
So while targeting campaigns solely to RLSA and Customer Match lists may result in higher overall account CTR and conversion rate, it can also lead to negative impacts on new customer acquisition and the health of these audiences further down the line. For these reasons I’d say such an optimization might be another example of an optimization dealing death to the campaigns it was meant to help.
Deleting/pausing keywords and Google Shopping product targets that aren’t performing well
A great man once said, “There are no bad keywords,” only bad bids. The logic behind this is that, assuming your keyword list is reasonably targeted and relevant to some aspect of your business, keyword bidding should be used to pay the appropriate price for traffic.
If a keyword isn’t providing an adequate ROAS, the bid simply isn’t low enough, and keywords that will never produce an adequate ROAS should eventually be bid down to a level so low so as to shut off traffic.
Brands that try a keyword and find that the ROAS is poor might be tempted to pause or delete the term. However, the first step should be to try lowering the bid, while also examining whether the ad copy and landing page for the term are as effective and targeted as they can be.
Similarly, product targets for Google Shopping campaigns shouldn’t be immediately paused if they don’t initially perform as desired. Can the target be broken down into smaller targets with fewer products each in order to set more granular bids? Can bids be adjusted to hit ROAS goals? Are there queries driving poor-performing traffic to the products within the target that should be prevented with the use of negative keywords?
These are all questions that advertisers should look to answer before pausing product targets that are producing poor ROAS.
Deleting or pausing keywords and product targets without attempting to adjust bids and other variables can result in valuable traffic being written off for good, to the detriment of paid search campaigns.
Don’t accidentally kill your paid search campaigns
These are just three of the ways in which well-meaning optimizations might end up coming back to haunt paid search managers. And while this post is no Wikipedia page on inventors who were killed by their own inventions, hopefully it will help get you thinking about some of the adverse effects of optimizations you’re considering or have already undertaken.
Some death traps, much like the parachute coat, can be avoided.
Here are a few examples that come to mind, with some suggestions for avoiding self-induced paid search failure at the hands of well-meaning optimizations.