PPC Search Bid Management: Humans vs. Machines

Despite there being an almost religious fervor among certain PPC campaign management teams at advertisers and agencies, the truth is that both human capital and automation combine in a variety of ways to deliver an optimal overall campaign.

The issues of human vs. machine PPC campaign management aren’t black and white, and there are many factors that will dictate exactly what level of automation, bid management, or campaign management automation is appropriate for you. Before we get into some predictive factors that help determine if and when automation is right for you, let’s look at the areas where automation and technology excel.

Automation really helps when you want to:

  • Maximize revenue, profit, and/or mixed metrics only when a strong foundation of keywords, match types, ad creative, and landing pages already exists.
  • Deal with and optimize in a media and bidding landscape where a multitude of simultaneous choices exist.
  • Manage around competitive reactions to changes made within your campaign (market elasticity).
  • Do heavy data collection and lifting, including with segmented campaigns.
  • Merge product datasets/feeds into semi-automated or fully-automated ads.

Similarly, there are areas of campaign management where humans excel. Humans can think! Humans are supposed to make sure that whatever technology is in use has the strongest possible (yet constantly evolving) foundation.

Humans really excel with:

  • AdGroups structure
  • Keywords/match types
  • Testing and changing targeting levers (geography, syndication, mobile, etc.)
  • Ad creative (particularly on high volume AdGroups)
  • Landing pages (test, create new, modify)
  • Identifying and fixing Quality Score issues
  • Exploring tradeoffs between ROI and volume
  • Planning and execution of experiments
  • Allocation and prioritization of resources

Given the above strengths of humans and machines/automation, one can easily see that for larger accounts it makes a great deal of sense to have humans highly active in campaign management in conjunction with the machines.

Many search campaign management technology vendors give the impression that technology will replace human involvement almost entirely, when in fact, by taking over the heavy lifting in the areas of bid management, budget optimization, cross-media allocation optimization, and time-consuming campaign adjustments, automation simply frees up the human capital to focus on other tactical and strategic initiatives. When a campaign of sufficient size (clicks and conversions) is moved from a manual process to an automated process, there is an immediate lift in efficiency. However, without human attention to the tasks listed above, the campaign efficiency gains will begin to plateau. In some cases, where there is a highly competitive keyword set, efficiencies may even begin to disappear over time, as competitors in a given vertical work on improving Quality Score, conversion rate, ads, landing page conversion rate, and average value of a conversion. All these factors provide bidding leverage (increases in the maximum reserve price on a click that the bidding algorithm can deploy into the marketplace).

Campaign automation can extend beyond bidding in some cases. Automated ad creation, automated AdGroups breakouts (separating keywords into their own groups), automated application of negative keywords, and automated match type adjustments are all possibilities. However, in most cases, human oversight can improve upon the suggestions made algorithmically, and certainly human teams may want to moderate the suggested changes before pushing them live.

Smaller spenders (or, in particular, those with not a lot of clicks) often can justify a fully manual campaign for a variety of reasons, the most significant being the lack of data. In order to know what is “working” at the keyword, time of day, and geography levels, you need a lot of clicks and a lot of conversions.

Similarly, many businesses, including professions and B2B advertisers, have more complex buy-flows. Even trackable telephone calls and a process for feeding that data back into a media optimization algorithm may not fully account for the value delivered when certain visitors from certain keywords, engines, and geographies arrive to learn more.

Use of proxy variables (also known as micro-conversions) can sometimes take a niche advertiser with lower budgets and provide a campaign management automation solution with enough data to optimize on an automated basis.

Machines can have their place in almost any size campaign, but an algorithm is only as good as the human driving the machine. That’s why keeping yourself trained on best practices – both strategically and tactically – is so important.

Related reading