I am here to help you shift your mindset on PPC automation and AI. Firstly, to help you understand the role automation and AI play in PPC currently and secondly, to help you to start thinking of automation and AI as collaborators to your role rather than competitors. Finally, I share some tips on how to get started.
What is PPC automation?
Automation is all about doing things more efficiently. It is nothing new, over the years many industries have been revolutionised by automation. From robots doing manual labour jobs like building cars, to ATMs dispensing cash and communicating to banks. It is evident to see that automation is all around us.
In our everyday lives as consumers, we embrace automation without a second thought. From the moment we wake to the moment we go to sleep; automation shapes our everyday lives. We automate our heating schedules; our payment schedules the list could go on. We benefit from automation coupled with AI to find new shows to watch on the likes of Netflix.
PPC Automation vs AI
Automation and AI work hand in hand, but are fundamentally different.
Automation is a process to remove manual input and works with data to take repetitive action based on pre-defined rules that don’t require human input.
Whereas, AI or artificial intelligence is an evolving beast that learns and develops and works to understand the data. AI can change a decision and outcome as data evolves.
Both automation and AI can help businesses and marketers to streamline their processes.
The fear of automation & AI
So why are so many of us afraid to let automation into our working life?
- The fear of job replacement
- Being redundant in our roles
- Losing an element of control
- Untrusting of the results
But automation doesn’t have to mean these things. If you embrace and work with automation it leaves you time to focus on less menial tasks. Furthermore it has the power to enhance the results that you see.
Components of automation in PPC
Automation is commonplace in PPC, it is used to streamline processes such as account setups and management. As with most automated processes, automation in PPC requires a level of human input to get it working, but if you can map out the steps, the likely chance is the task can be automated.
Examples of automation in PPC include:
- Bidding – use automated bidding strategies such as maximise clicks and target impression share.
- Optimisation – automate optimisation actions using Google recommendations or performance based rules.
- Rules – set rules to make adjustments up to once per day. These rules can apply to bids, budgets and statuses.
- Scripts – use scripts to automate actions up to once per hour. Scripts can generate reports or take actions in accounts.
Components of AI in PPC
AI’s presence in PPC has increased exponentially over the last 2 years, and if you have not adapted to use it then you are likely already a step behind your competitors.
- Bidding – use smart bidding strategies such as maximise conversions or maximise conversion value.
- Targeting – benefit from Google’s algorithms to automate search term and audience targeting in campaigns such as dynamic search ads and performance max.
- Creative – with responsive search and display ads Google dynamically decides at each ad auction which variation shows.
- Optimisation – Using data signals not available to humans for each ad auction or the likes of automatic adjustments and predictions of trends.
Pros and cons of PPC automation and AI
As with everything there are pro’s and con’s to the use of automation and AI in PPC. Ultimately it comes down to each business and each account element to determine whether the value brought in things like efficiencies or increased results outweighs the sacrifice in elements such as loss of control.
- Increase efficiencies
- Reduce time spent on tasks
- Spend more time focusing on elements in your control
- Faster learning
- Improve results
- Accounts for wider unseen factors from Human’s
- Brand guidelines are not adhered to
- Less control
- Short term campaigns don’t always perform as well as there is no time for the machines to learn
Tools that support Google automation
There are a number of tools that you can access within Google Ads to support how you utilise automation and AI and predict its impact on results.
- Performance planner – use this for search or performance max based activity to predict the results of campaign activity based on a set of criteria such as spend amount or target revenue.
- Optimisation score – this is Googles indication of how well it believes your account is setup. Whilst it is important to review these suggestions, it is even more important to not just blindly apply them. Review each recommendation in line with your business objectives and branding and apply and dismiss as required.
- Portfolio bid strategies – you can use portfolio bid strategies in multiple ways. Firstly if data is spread across multiple campaigns, using a portfolio strategy can help Google to learn across multiple campaigns. Secondly, you can use bid strategies in line with target CPA and target ROAS bidding to set maximum CPC limits which you cannot do within the individual strategies in campaigns.
- Seasonal bid adjustments – you can help guide the bid strategies for any expected changes to performance for small periods of time (up-to 7 days). The benefit of these is that it does not then kick the bid strategy back into a ‘learning phase’ if there is an unexpected shift in performance. This can be used for events such as Black Friday where typically conversion rates increase.
- Data exclusions – data exclusions are again a way for you to guide the bidding strategies. These essentially tell Google to ignore data during the period specified. These can be used in instances such as broken tracking.
Embracing smart bidding
If you are still a smart bidding sceptic you are at risk of getting left behind. Whether you opt to use Google’s algorithms or a 3rd party tool, if you aren’t doing either you are likely either missing opportunities or wasting spend.
Smart bidding happens at auction, meaning additional signals that human input cannot factor in for each and every search.
Smart bidding – best practices
- Portfolio bid strategies
- Enables you to set max CPC bid limits for target CPA and target ROAS bidding
- Smart bidding learns from multiple campaigns at once
- Ensure campaign goals in the same strategy are aligned
- Targets are important
- Ensure targets are achievable or you risk hindering the bid strategy
- If you already have results keep the targets within 10% of your current results initially and gradually adjust
- Pace any changes you make
- Reduce the need for learning periods by making smaller incremental changes
- Allow the system time to learn before making further changes
- Accurate tracking is crucial
- Ensure your conversion tracking is set up accurately
- Measure the right conversion points
Deciphering the good from the bad
When it comes to automation and AI there are elements that still benefit from human input.
What is right for one account, may not be for another, there are certain elements I would always advise to disable/steer away from automation:
- Auto-apply recommendations for ad creative
- Auto-apply recommendations for removal/pausing
- Auto-apply recommendations to add new keywords
- Overall avoid anything that may compromise or go against your marketing strategy or brand
Automated campaign types in Google Ads
Automated campaign types are not new, some have been around for many year but Google is always developing new initiatives
DSA (Dynamic Search Ads)
- No keywords, matches to searches based on website content
- Google automates ad creation and landing page selection
- You set budgets and bidding model
- Automates multiple parts of the marketing funnel
- You feed the campaign information – website URLs, product feed, audience signals, assets
- Refine and optimise the campaign still – use google reps to add negatives
- You set budgets and choose smart bidding model
What to automate?
When deciding what to automate focus on a matrix of 2 questions:
- How time consuming is the task for you to complete?
- How frequently do you have to complete it?
This will lead you to:
- What automations will drive the most value?
Shifting the way we work
As has always been the case before automation existed, our roles as PPC managers are always evolving. Automation and AI are just another twist in the road and require a larger shift in the way that we work. So, what is changing and what is staying in terms of our involvement in the setup of campaigns?
There will be changes to:
- Campaign structures
- Less granularity – SKAGs are out
- Keyword match types
- Targeting capabilities
Where we will still have the biggest input:
- Content and creative
- Budget planning
- Source data – for example product feeds
The new era of PPC management
In order to stay ahead you have to continually adapt your strategies to create new best practices and approaches. The role of a PPC manager must also evolve, and you must shift focus and you can add value by contextualising, thinking creatively and challenging Google’s decision making.
The PPC controller
AI is smart, but not above errors, it also (at least in its current state) cannot understand your brand as well as a human
- Monitoring systems
- Prevent bad actions from automation
- Work as a stop gate to QC the automation and AI
- Ensure context is upheld
- Ensure core data aligns to the business
- Ensure branding is upheld
The PPC pilot
AI can’t function without guidance. Now rather than managing complexities within Google, we need to inform Google better about our business
- Provide base information
- Teach and guide the machine
- Optimise base data
- Experiments and testing
Whilst automation works on the tedious elements the PPC Pilot can be adding value to the base information to guide the AI and continue to drive and elevate your results.
- Account structures – organisation and objectives based on keywords, audiences and products
- Creative assets – imagery, video, messaging and landing pages/websites
- Structured data – product feeds and data
- First party data – customer emails, offline conversions, post conversion data
- Conversion tracking – point of sale/lead tracking
- Targets – determining the optimum targets
It also allows you more time for analysis, so you can strategise and pivot based on live data
Conversion tracking is the most crucial element to campaign smart bidding.
Think about how as a human you optimise campaigns, you might reference other data sources when assessing a campaigns success.
Very often in B2B for example we will track a lead or trial sign ups within Google Ads, but we don’t always pass that offline information back in. Or for ecommerce we may track revenue but not account for profit
So if we move to an era of AI and smart bidding, where Google is deciding which conversions are valuable, we need to ensure we are teaching Google what valuable looks like.
So whilst tracking initial leads and enquiries was okay before, it is becoming more crucial to take a view of the full sales process. Because while yes leads are valuable, ultimately as businesses what really matters is those leads becoming sales.
When it comes to PPC your creative assets (ads and landing pages) have always been crucial, but no more so than right now.
Collaboration is now your best friend. If you aren’t already you need to be thinking as an integrated marketing team. From SEO to content to UX, each channel supports the other.
You need to focus on testing, improving and optimising:
- Site speed
- Calls to action
- Call to action
When it comes to ecommerce in Google Ads, your shopping is the single most important element.
The feed determines which user searches your ads can match to so it is crucial to ensure you are providing accurate and full product data. Understand both what is important to Google’s algorithms and what matters to potential customers.
- Required fields
- Enhance titles & descriptions. This could be through adding in key search terms or product attributes.
- Ensure accuracy as these elements determine how Google matches to users searches. Whilst adding certain things may increase your clicks, if they are reflective of your product then they will increase bounce rates and lower conversion rates
- Recommended fields
- Attributes – colour, size etc.
- Custom labels – use to provide supplementary data e.g. profit margins
First party data
From customer email databases to sales and lead data your first party data is more and more important.
- Email databases such as your customer emails or newsletter signups. Just remember to ensure these are all GDPR compliant and you have permissions to use these for marketing purposes. Once uploaded you can use them as observation signals, targeting or exclusions.
- Offline sales data can be passed back into Google Ads. This can either be through platform integration with tools such as Salesforce, API or through a Google Sheet upload. It requires some technical changes to how your tracking is set up to ensure the GCLID tracking parameter is captured and passed into your lead/sales database to enable Google to match data back.
- Profit data can also be passed back into Google to help better guide which conversions are most valuable.
Test, analyse, test again
As new features release embrace them. If they don’t work, it doesn’t mean they won’t ever work. As with everything automation and AI in paid media evolves and gets better over time.
The smart bidding we see today, is very different to the smart bidding we saw 5 years ago. So, if it doesn’t work the first time, circle back in 6 months. See what updates have been made, read up on other use cases and test again.
The evolving landscape
As the nature of AI and automation, it is an ever evolving and changing beast. There are new features and tools dropping both directly in the platforms and externally to push the market forward.
Some of the most recent changes that have come to the world of PPC (and digital) are:
- Rollout and updates of performance max campaigns in Google Ads
- Auto-apply recommendations
- ChatGPT – this along with likewise tools could be one of the biggest revolutions in digital marketing. Its use and capabilities stretch into PPC from campaign plans to ad drafts, people are already feeling its benefits. Ann Stanley shares how she sees ChatGPT and OpenAI changing the world of digital marketing in her webinar and subsequent blog.
New to PPC automation? Where to start
If you have been reluctant to test automation and AI, or are just not sure where to start, then I would recommend:
Automated rules – simple logic can be implemented up to on
- Logic based bid rules to increase or decrease bids (when using manual CPC bidding) based on performance data – e.g. if ROAS is >10 and clicks are >100 then increase the bid by 10%
- Logic based budget rules to increase or decrease budgets based on performance data
- Run as an experiment a/b test and review the results. It will depend on your business which will be the best strategy to test initially, but if you are ecommerce and track revenue for example it will make sence to test target ROAS bidding.
- Dynamic search ads are an easy place to start. You retain a higher degree of control as you can easily add exclusions to both searches and also which pages Google references on your website.
- Automation and AI is the future and it is here to stay
- AI works best when guided by human input
- Embrace automation but don’t just blindly follow Google’s recommendations – they don’t know your business or have context. Disable auto-apply recommendations and manually review
- Ensure you continue to understand the building blocks of PPC to know what is powering the automation and to retain the ability to identify when things aren’t as they should be
- Keep a close eye on your data
- Don’t treat all campaigns or clients the same – learn what works for each
If you are new to using automation and AI then all the changes and options may seem overwhelming. We are here to help, from training to consultancy and management. Get in touch to discuss how we can support you in this shift.