As 2017 draws to a close, let’s take a moment to catch our breaths and look back at the whirlwind that was PPC in 2017.
There wasn’t a big change that dominated the landscape like enhanced campaigns of 2013 or expanded text ads of 2016, but multiple trends created an atmosphere of constant, incremental change this year. However, if we were to dub 2017 the year of something in search marketing, it would clearly be the year of the machine. While machine learning and other forms of artificial intelligence aren’t new to search marketing, their use became pervasive in 2017.
Here’s a look back at the big developments and key trends that happened in PPC in 2017 that will continue to inform and influence our work in 2018.
Finally past the year of mobile, this was the year of AI in search
Sure, there is still work to do in improving mobile experiences and conversion rates, and we’ll continue to see Google, in particular, push its initiatives in this area: AMP for ads and landing pages, Purchases on Google and more. This year, the big shift was the extent to which machine learning and other forms of artificial intelligence permeated all things search.
Here are eight highlights of ways the search engines ingrained machine learning into their products. They cover everything from keyword matching to ads to audiences to spend pacing to attribution:
In March, Google made putty of the meaning of ‘exact’ in exact match, stretching it to include close variants of a keyword with different word order and/or function words.
Ad rank thresholds got a machine learning infusion to take the context of a query into consideration when setting the bid floor.
Google’s Smart display campaigns are nearly entirely powered by machine learning.
Google’s data-driven attribution methodology is entirely AI-powered. It’s been in AdWords for more than a year, but it gained new attention with the introduction to Google Attribution.
Google and Bing released new automated bid strategies: Bing’s Maximize Clicks and Google’s Maximize Conversions.
Google’s move to let daily spend exceed up to 2x the budget? Yep, that, too, relies on machine learning to try to predict spend trends throughout the month.
One flavor of Google’s custom intent audiences on the GDN uses machine learning to automatically create audiences based in part on inferred characteristics of an advertiser’s target customers.
Bing Ads is testing AI-powered chatbot extensions in search ads.
Dynamic Search Ads in Bing Ads came to the US and the UK.
Forget A/B testing, because machine learning
Another big, if more subtle, shift was in ad testing methodology. All year, Google has pushed advertisers to move away from the A/B testing model of running two ads per ad group and manually assessing performance.
If there was any doubt Google was serious about this, the move to limit ad rotation options in Augustput that doubt to rest. The change makes the push for advertisers to choose “optimize,” letting the machines choose the best ad to serve, that much more forceful. Google’s Matt Lawson laid out in a column last month the argument for having at least three ads in an ad group: Overall impressions will increase as Google’s algorithms will serve up the best ad based on the specific query. Advertisers shouldn’t even be evaluating individual ad performance under this new rubric, but rather at the ad-group level of performance, says Google.
To this end, Google rebooted its Ads Added by AdWords pilot in September. The ads suggestions test automatically generates additional text ads (for approval) in some ad groups. Again, the goal is to get more advertisers running more ads in their ad groups, even if Google has to do the work for them.
Original article by Ginny Marvin on http://searchengineland.com