Automation and machine learning in PPC

/ PPC / October 03, 2019 / Comments

Digital transformation has continuously showcased and proven that specific mechanical tasks are more efficient and effectively done by automation or machines rather than humans. The search advertising industry is an excellent example of the so-called 4th industrial revolution that our society is currently undergoing.

Numerous tasks would improve time efficiency and accuracy if done automatically. This is where machine learning comes into play as well, crunching big data and exploring user patterns, while PPC marketers are gaining time for strategy and solution-finding.

In the past few years, Google has rolled out several “smart” automated features, which further confirms the growing influence of machine learning in pay per click advertising.

It is not likely that machine learning will ever wholly replace PPC advertising specialists. Machine learning and automation tasks enable PPC advertisers to focus on high priority and more in-depth analytical tasks, while AI completes routine tasks, analyses data, and spots trends that can be implemented in the PPC strategy.

Among the tasks that can be automated are:

  • Reporting
  • Competitor analysis
  • Workflow

Areas in which machine learning can bring valuable input:

  • Bid adjustment
  • Ad copy testing

A great starting point for PPC campaign automation is installing or even creating your scripts that can be pasted into Google Ads to improve productivity, workflow and performance. Use Google’s free guide to getting started with scripts.

Let’s have a closer look at the main areas where machine learning can contribute to your PPC strategy.

In-Market Audiences

In-Market Audiences was the first of Google’s machine learning technologies, revealed in 2017. It can be used for building brand awareness, as it targets users that are actively searching for specific information, product, or service.

The AI studies prior search and conversion data auction-by-auction and concludes the market audience that is most prone to convert. This is done by determining the search intent and considering the user’s search history. Manually, it would be impossible to process such a vast amount of data.

However, a growing amount of PPC specialists are increasing the use of automation in their strategy. It is important to remember that setting up any automation requires a strategy, so you would not end up scaling the wrong thing. It saves time long-term but requires a time investment in the staging phase.

In conclusion, do not fear the AI takeover, PPC specialist will remain there to strategically decide what tasks to allocate to the machines. Instead, understand the technology and how you can leverage it and spot opportunities for your integrated PPC strategy.

Google Smart Bidding

Smart Bidding is a Google Ads feature that maximises conversions and enhances PPC advertiser target CPC (cost per click) and ROAS (return on ad spend) bidding strategies at the auction level.

It combines signals like geolocation, user’s device, and time of the day of the query.

The machine learning will lean towards the most profitable position or the defined goal as in the Smart Display campaigns.

As manually, the task of analysing large amounts of Google Ads data exported to Excel would take a lot of time, automation speeds things up, and provides a more accurate prediction.

As Google describes it: “In bidding, machine learning algorithms train on data at a vast scale to help you to make more accurate predictions across your account about how different bid amounts might affect conversions or conversion value. These algorithms factor in a wider range of parameters that affect performance more than a single person or team could compute.”

Bid management has a high time-saving score and increases performance. Keep in mind that this is not an overnight solution, as the system must process the data and optimise for new strategies before you see any results.

Smart bidding:

  • Adjusts bids for every auction.
  • Analyses search context to predict the likelihood of conversion.
  • Incorporates historical data about user behaviour and auction specific signals such as user’s browser, language, device, and search query.

Smart Display Campaigns

Smart Display campaigns can be broadly used in almost all formats in the Google Display Network. The campaigns can target users at all stages of the purchase funnel, from initial interest to ready to buy.

Smart Display campaigns are entirely machine-controlled. Firstly, the account manager has manually entered target CPA and campaign budget, as well as ad headlines and images. Secondly, the campaign is left to the automated system to achieve the given goal.

The system learns from data and optimises the campaigns to reach targets. Therefore, the more data it has, the abler the machines are to optimise the campaigns.

There are three automation aspects involved in Smart Display campaigns:

  • Automated bidding – at every auction your bids are optimised based on the likelihood of converting.
  • Automated targeting – with continuous optimisation and using dynamic prospecting, your ad increasingly is displayed to the audience with the highest potential to bring you revenue.
  • Automated ad creation – the entire ad is generated automatically.

Smart Shopping campaign

Smart Shopping campaigns combine automated bid adjustment and product ad placement across networks.

How does it work? Google’s machine combines your existing product feed and assets, and tests different image and text combinations to achieve the most relevant ads shown. The ads can be shown on the Google Search Network, the Google Display network, Gmail, and YouTube.

A requirement for using Smart Shopping campaigns is providing a remarketing list with at least 100 active users , as well as to set up conversion tracking with transaction-specific values.

Each Google Ads account can have a maximum of 100 Smart Shopping campaigns.

Smart Campaign

Smart Campaign is a campaign type for small businesses. The advertiser creates an ad called “smart ad”, writes a business description that includes information about the business’s product or service, sets a budget, and leaves the rest to the AI.

This is a great way to build business awareness and get your name out there. The machine learning algorithm shows the ad to potential customers looking for terms related to your business, your geographic area, or for terms relating your business whilst physically being outside of your target geographic area.

Ad copy optimisation

To optimise your ads, you can use either Responsive Search Ads or the “Optimise: Prefer best performing ad” rotation option to optimise ad delivery.

These ads are expected to perform at a higher level than other ads in the ad group. If you have multiple ads within an ad group, Google will test each ad and will learn which ad performs the best. It will then show the best performing ad more often than the other ads.