Updated: Jul 18
Like it or not, automation and machine learning in pay-per-click advertising are here to stay. Smart Bidding, Performance Max, Broad Match… these are just the beginning.
And there’s a good reason Google is so bullish on their machine learning capabilities.
Technology like Broad Match uses oceans of information across thousands of data points and signals. Google’s algorithms analyze these in real time; first when someone searches for a topic, and then again when ranking matching ads for the auction.
It’s not just more information than the human mind can analyze; it’s more than we can even comprehend. But it allows Google to find us searchers who are likely to convert with a relatively high degree of accuracy, especially where the alternative is purely human-driven.
But neither people nor machines are at their best when working alone.
Automation and machine learning reach their full potential when paired with human oversight—an arrangement that allows us to focus on things that machines still can’t do: creativity, contextualization, and problem-solving for example.
What you need to know before you use machine learning in Google Ads
The finicky nature of lead generation PPC makes it an especially good fit for one specific combination of Google Ads technology: Smart Bidding with Broad Match keywords. However, there are some things to keep in mind before you press any shiny buttons.
It requires a certain volume of conversions before it’s viable. If your account gets fewer than 30 conversions per month, apply micro-conversion events instead (engaged visitors, link clicks, page visits).
It’s rarely feasible for smaller accounts, such as those getting around 20 or fewer clicks per day. Some advertisers focus on budget as a threshold, but cost per click plays a bigger role than outright spend.
It’s expensive to get rolling. All of Google’s machine learning technologies require a learning phase that lasts around 4-6 weeks. You need to be okay with treating your ad budget for this period as an investment in data gathering.
It requires accurate first-party data to be fed back in. This is the only way to fully optimize the system’s ability to make the right decisions for a business. This data takes many forms including lead qualification, bidding against sales, and Enhanced Conversions.
Is the effort worth it? Absolutely.
Is it going to be easy? Anything worth doing rarely is.
Level up your Search campaigns: A step-by-step guide
When launching new campaigns on Smart Bidding and Broad Match, you can’t simply apply them from day one and expect results. That’s like jumping into the middle of the ocean for your first swimming lesson.
Build up to an automated campaign state by prioritizing these goals in the following order:
Source as many clicks as possible to acquire data.
Get those clicks to begin converting in order to refine your data.
Optimize for overall performance and profit by leveraging that cleaned data.
Phase 1: Gathering data
Bidding strategy: Maximize Clicks Clicks or Manual CPC
Keyword match type: Phrase or Exact
Goal: Get as much traffic (and data) as possible.
Duration: Typically around 2 weeks starting from when your campaign begins to spend its full budget and drive consistent traffic.
The objective here is not to have an efficient campaign, but to build a functional data-gathering operation. Hence, consider how to structure your google Ads account in the age of SMART bidding. As more people click on your ads, you will need to watch search terms closely. Block themes aggressively with Phrase Match negative keywords.
Here’s a starter negative keyword list to help you get started.
Phase 2: Refining data
Bidding strategy: Maximize Conversions
Keyword match type: Phrase or Exact
Goal: Turn clicks into conversions and collect enough accurate data for Smart Bidding to work correctly. This is typically 7+ conversions per week.
Duration: This phase is less about time lapsed and more about conversion volume gained. You’ll want to look for an average of one per day, or ~30 in a month.
At this stage, you should be getting enough clicks to understand which queries bring traffic to your landing pages. Use the search terms report to exclude everything irrelevant, and make changes to your landing page copy and offer so that more people hit that “final” conversion (typically a phone call or form submission for lead generation).
Phase 3: Optimizing data
Bidding strategy: Target CPA or Target ROAS
Keyword match type: Phrase or Exact
Goal: Turn your campaign into an efficient growth lever and hit client KPIs.
Duration: As long as it takes.
When applying a CPA or ROAS goal, start by setting a value on the higher end of your target range based on the last 7-14 days of performance.
Then layer in Broad Match keywords and Dynamic Search. This combination gives Google’s system more flexibility and freedom. Your ads will qualify for more auctions, CPCs will typically be lower, and you’ll discover more opportunities for growth.
Phase 4: Cleaning data
With an efficient campaign humming along and picking up new keyword opportunities along the way, you’re finally ready to achieve peak performance. Layer in Smart Bidding and integrate your customer data in three major ways.
Replicate your lead pipeline to show Google how you do business. The graphic above is an example of a lead generation business pipeline, with each stage weighted in value based on how likely it is to progress further downstream.
Import your offline conversion data so that Google knows how each of their in-platform conversions (e.g. form submissions) perform once they engage with a revenue team. This can come from an IVR, CRM, or other system of record. Google Ads has native integrations with HubSpot and Salesforce, while nearly every other tool can connect via Zapier. Or you can simply go old fashioned with a spreadsheet or CSV.
Build a process to maintain data hygiene. While no data set will ever be 100% clean, you still want to maximize accuracy so that Google has the latest and greatest information when it comes to your client or brand.
Don’t run forms as a conversion event without a way to filter out spam, such as reCAPTCHA or Honeypots.
Continue to monitor for and apply negative keywords. Use the search terms report in Google Ads at first; when your volume gets into hundreds of conversions per week, n-grams will make this easier.
Run standalone Search campaigns for your brand terms. Not only will the considerably lower CPCs and different search behavior influence Google’s machine learning, but by targeting impression share, you can avoid overpaying for your branded terms.
Exclude customers from the main campaign; you don’t want to target someone who’s already bought from you and therefore no longer in the market. Update your customer list every week or two to keep this information up-to-date.
Navigating the nebulous nature of Google Ads
You’ll face many challenges as you navigate the Smart Bidding and Broad Match combination—some within the platform and others outside it. One of the biggest will be managing expectations with clients and bosses.
My advice is simple: Be honest.
Lay out the benefits, explain where you need to be careful, and ask for the support you need. Above all, be clear about the level of investment required. If you can’t get buy-in from the people who set the numbers, it’s not worth the stress.
Another concern is what happens when you layer automation technology. With Performance Max active in the same account, you run the risk that some of your Search campaign’s traffic will get diverted there. These are the new problems PPC managers need to become good at fixing.
And what if Google removes the ability to target keywords at some point? How will you adapt? Do you understand other targeting options well enough to sculpt traffic like you would with keywords?
Our industry is constantly evolving, which means our jobs are fluid too. Don’t spend your time worrying about ad engine automation taking your job; it’s not happening. Instead, shift your focus to creating value by overseeing the machines.
There’s no better insurance policy for a career in PPC advertising.
About The PPC Expert
Menachem Ani is an expert in Online Advertising and eCommerce with over a decade of success developing high-impact marketing strategies for online retailers and lead-generation clients.
Some of his core skills include SEO/SEM Marketing, Google Shopping Management, Campaign Management, Inbound Marketing, and Conversion Optimization.
In 2009, he founded JXT Group to provide full-service management for online advertising campaigns. His agency is a Google Premier Partner – only 3% of partner agencies have this designation. It means they have multiple team members certified in advanced AdWords knowledge, by Google’s own standards. It also means their company has demonstrated high levels of activity delivering solid overall revenue and growth and developing our client base.
To ensure that you are alerted to when more juicy pieces of content gets released subscribe to our weekly newsletter to get PPC news the will unleash your true PPC potential.