Photo illustration for "2017 review" and "Paid Search"Paid search in 2017 really kept everyone on their toes, as we responded to all the changes brought about by AI and machine learning, programmatic advertising, and the growth and maturing of several channels and platforms. Here are the top 5 changes we observed in 2017.

#1 – Machine learning proliferates through paid search, for better and for worse

AI and machine learning were definitely the biggest trend in paid search marketing in 2017. Both Google and Bing introduced options to maximize conversions and clicks, respectively, using machine learning to determine optimal bids. The two channels also premiered tools to generate ad copy and images automatically based on searcher queries and intent.

While these changes boosted effectiveness and efficiency for paid search, some machine-learning-based developments made things trickier and more time-consuming. On Oct. 4, Google announced that AdWords campaigns could now spend up to twice their average daily budget on days with “lots of high-quality traffic.” Advertisers wouldn’t be charged more than their monthly charging limit (the monthly equivalent of the average daily budget), so the net spend would come out the same.

However, for clients with very defined budgets and fluctuations in daily spend, such as B2B companies with much lower user response on weekends, managing spend was already fairly complicated. This change added to the complexity of managing those accounts, especially because it wasn’t optional and went into effect the same day it was announced.

Fortunately, we were able to respond immediately to course-correct the spend increases. We also added additional monitoring to these accounts to ensure the right amount of responsiveness, so increases in engagement can be boosted by additional bidding, while retaining enough budget to generate conversions throughout the month.

#2 – Programmatic ads still need human oversight

As we noted earlier this year, “Programmatic can be a good investment; the real-time bidding algorithms can save money by responding instantly to fluctuations on various ad exchanges and by reducing management costs.” But as we also noted, “By nature, programmatic algorithms don’t judge the quality or purpose of a site before placing an ad there.” And that led to some seriously brand-damaging placements for major companies.

In early March, one of our B2C clients learned through Twitter that their ads were appearing on offensive and inappropriate sites. We immediately paused their programmatic advertising before any further brand damage could be done, then quickly resumed placing ads programmatically using a blacklist/whitelist approach to ensure optimal placement while retaining as much of the efficiency of programmatic advertising as possible.

#3 – Mobile: No longer eating at the card table

For most companies, mobile is now a fundamental part of paid search, rather than a separate thing. Mobile costs per click (CPCs) were significantly lower than desktop CPCs for many of our clients, bringing mobile costs per conversion within range of their desktop counterparts. On average, the return on ad spend (ROAS) for mobile is still about 40% lower than it is for desktop, but we expect to see improvements in revenue and returns in the coming year.

#4 – Shopping: Not just for B2C anymore

Our B2C clients continue to see impressive growth in the returns from their product listing ads (PLAs) through Shopping channels. However, this year, Shopping became a profitable channel for B2B as well, with conversion rates and RoAS on par with other channels, including search and display ads. We’ve increased investment in Shopping for most of our clients, and have seen it pay off well.

#5 – Bing keeps maturing as a channel

Our ecommerce clients saw a noted improvement in results from Bing; overall, it beat Google for ROAS on non-brand terms (the all-important terms that attract new customers), and for one major client, Bing and Yahoo consistently beat Google for ROAS on brand terms as well.

The Shopping and Bing trends dovetailed nicely for several of our clients, who saw their ROAS from Bing Shopping overtake ROAS from Google Shopping in Q3 and Q4. As personalization and audience identification improve, we expect to see even higher returns from Shopping and PLAs.

From everyone on the (un)Common Logic paid search team to you and your loved ones, we wish you happy holidays!