Can AI Automation really improve your PPC campaign?

Can paid search giants really promise that AI automation can take away the human aspect of your PPC campaign management?

Search engine titans such as Google are pioneering and encouraging us all to embrace automation. In this blog, I will look at two example Google products where automation is at the heart of optimisation, and weigh-up the benefits and drawbacks of this, as well as highlighting some optimisation tips.

Dynamic Search Ads

The digital marketer is no stranger to Dynamic Search Ads (DSA) and how they can be used to capture the traffic that may have otherwise been missed through your standard keyword campaigns. Across agencies, this type of campaign is often used as a tool for ‘keyword mining’, where the keywords that perform well will be picked up and dropped into keyword campaigns running within the account.

However, Google now strongly suggest that this should not be the case, and that your keywords will perform better staying in the DSA campaign. Google advises that the keywords will run at a 35% lower CPA with 15% more clicks. Croud are currently testing this for our clients to see the plausibility of moving away from ‘keyword mining’, which reduces some of the heavy manual aspects of ensuring relevant keyword coverage.

The way a DSA should be set up has also changed, with specific DSA campaigns no longer needed. Google proposes using the hybrid ad group for DSAs, which run alongside your other ad groups. At this point you may be asking, ‘won’t traffic be pushed through the DSA ad group instead of my exact match type ad groups, leading me to spend more?’ The answer is no, – according to Google, “exact match always wins” in the auction against your DSA, providing you are not constrained by budget, have low-quality scores or have the wrong targeting.

What levers to pull, in order to optimise DSA campaigns

  • The right targeting is key as, since 86% of DSA traffic is incremental, we want to ensure the targets are more broadly defined and that we have enough indexed landing pages to assist this.
  • Keyword negatives and audience exclusions are pivotal in helping traffic be directed to the best-optimised campaigns and ensuring you don’t appear for the searches and audiences you don’t want to (just as you would in your other campaigns).

Responsive Search Ads

The automation of creatives has also become somewhat possible through Responsive Search Ads (RSA). The set-up of the RSA allows 15 different headlines and four different descriptions to be entered, with Google using AI to decide on what the user sees as the ‘best’ combination from their search term.

Since the release of RSAs earlier this year, there hasn’t been much to report about them, as they currently have quite limited metric reports – so far only showing the percentage of impressions each ad combination has received.

Limitations of RSAs

Sure, for good click-through rates and good conversion rates, RSAs are the ad type to use. However, if your client requires specific terms to be shown in their ads, such as brand names and slogans, then you are stuck letting the RSA decide what to show the user because it may not show either of those. You can certainly ‘pin’ your headlines and descriptions to always show in certain positions, but this could then lower the impact of the machine learning, and hands all the manual control back to us – curbing the impact of RSAs and maintaining the human element to creative updates.

What the above two examples – DSAs and RSAs – highlight is that we need to find a balance between AI and manual optimisation. It’s true that machine learning can take care of the laborious, time-consuming, day-to-day tasks a PPC Account Manager must do, such as bidding, reporting and keyword expansions from search query reports.

However, no automation can completely replace how agencies manage PPC accounts; this is particularly the case for Croud, which leverages local market insights and sophisticated skill sets of its network of over 2,000 Croudies onto campaign management– something that AI is yet to be able to do.

In conclusion

To conclude, AI automation really does help PPC Account Managers, by taking care of the repetitive tasks they do daily (and in many instances does it better due to the sophisticated algorithms). This allows more time for Account Managers to focus on strategy, tests and client service, which means working more efficiently to get better results for your client.

However, we are quite some time away from AI automation completely running PPC campaigns to the level that humans do, but that partnership of machine and people is certainly proving beneficial!

To find out more about how we leverage the power of automation, AI, and the Croudie network, contact us.

by Ruby Young
9 January 2019



Related posts