How to get better leads and conversions with Google’s AI
If you’re looking for ways to modernize your PPC optimization, you’ve probably come across value-based bidding (VBB). This technique revolves around teaching AI systems at Google and Microsoft what types of conversions you value most. Together with automated bidding and ad formats like responsive search ads (RSAs), the ad platforms can then prioritize getting you more of the best conversions and significantly improve the results from your ad budget.
VBB can make successful advertisers better, and it can even be a solution for advertisers who’ve tried and failed at PPC because they were unhappy with the quality of the conversions when leads were low quality or buyers made too many returns.
In this article, you will learn how to deploy VBB for three different types of advertisers: pure-play e-commerce, hybrid retail and lead gen.
The principle behind VBB
The idea of value-based bidding is that automated bids should be based on the value the resulting clicks and conversions add to your business. That’s not so different from the idea of bid management in general. But rather than achieving this goal through the manipulation of CPCs or targets like tROAS or tCPA, it’s achieved by teaching the machine the true value of conversions.
The reason VBB is so important in PPC in 2022 is that automation is now the standard way new campaigns operate and when you give automation bad or incomplete goals, you risk creating a vicious cycle that leads to poor results in those campaigns.
One problematic scenario is when advertisers give the ad engines an incomplete picture of what their goals are. Is the conversion they’re reporting to Google truly the conversion the CFO of the company cares about, or is it just some intermediate goal that happened to be easier to set up?
It’s similar to a problem you may face with people. When you hire someone for your PPC team, you can only expect them to drive great results if you tell them what results you’re after. If you tell your new teammate to get as many leads on the landing page as possible, don’t be surprised if those leads aren’t all of the most reputable origins.
If, on the other hand, you tell your coworker that the leads on the landing page will go to the sales team and they expect those leads to be well qualified, they will likely change how they go about generating leads and the quality will go up. If you tell them they will be judged not just on the volume of leads but also how many turn into paying customers, results are likely to get even better.
And so it goes with machine learning too. The machine will only do a great job if you teach it what you’re really after!
So let’s look at how you can teach the machines what a conversion really is and which type of conversions are the kind you’d like to get more of.
Optimizing PPC with better conversion data
There are two levels of sophistication when it comes to teaching the machine about the value of your conversions. Let’s start with the more sophisticated and precise method first. For every click or order, we will teach the machine what happened in the weeks after the original conversion event.
For lead gen advertisers:
The most sophisticated method of teaching the ad engines what you value relies on offline conversion imports (OCI), a method that depends on capturing the gclid or msclkid, passing it through your CRM and then feeding it back to the ad engines within 90 days as the value of the ‘conversion’ becomes more clear.
Recently Google introduced Enhanced Conversions for Leads, a simpler method with many of the same benefits but without the need for storing the click id in your own system.
E-commerce advertisers don’t need to grab the engine’s click ID but can instead send their own unique order ID with the conversion. As the true value of the sale becomes clear, advertisers can restate values to the ad engine within 55 days. Look up conversion value adjustments to learn how this works.
If you haven’t implemented one of the three methods above, it’s probably not because you weren’t aware of them, but rather because there is a technical limitation within your team that’s made it hard to implement. So let’s look at a new, simpler alternative to optimizing PPC with your conversion data.
It’s called Conversion Value Rules and lets you tell Google more about how to value different conversions based on a common attribute, like location, device or audience. While not as precise as the other methods, it’s a much easier way to teach the machine so it can start to prioritize the types of conversions that matter more to you.
Questions to help determine the true value of conversions
With Conversion Value Rules, advertisers create rules to adjust conversion values based on attributes like location, device, and audience.
When setting Conversion Value Rules, advertisers should focus on elements of a conversion that Google may not be able to observe like lifetime value, average deal size, lead-to-sale conversion rate, returns, etc. Google already knows about conversion rate differences between different locations, but what they may not know is what happens to conversions from different locations after they start to engage with your business.
Let’s look at some example questions to guide yourself to an initial set of Conversion Value Rules.
Conversion Value rule questions for lead gen advertisers:
- If you generate leads for HVAC installers, do prospects in certain zip codes have bigger houses and spend more on a typical installation?
- If you generate leads for education, do prospects in cities that are closer to campus tend to stay in the program longer?
- If you generate leads for plastic surgery, do prospects who read your article about rhinoplasty tend to become repeat customers and have higher lifetime value?
Conversion Value rule questions for pure-play e-commerce advertisers:
- Do purchases made in a hurry on mobile devices lead to more items being returned for refunds?
- Do purchases from people who read your blog with tips for runners tend to be more frequent repeat buyers of running shoes from your brand?
- Do purchases from those who engage with your social media platforms tend to lead to a bigger brand impact when they share their own images of their purchase with their friends?
Additional Conversion Value Rule questions for hybrid retailers:
Hybrid retailers can ask the same questions as pure-play e-commerce retailers but refine their Conversion Value Rules further with additional questions like these.
- Are customers in California worth more because it’s the only state with physical stores?
- Are customers who shared their email address when they shopped in-store worth more because they make fewer returns?
Now that you have an idea of what types of questions to ask to get an idea of conversion signals Google may not be able to detect on its own, it’s time to create rules for your most important traffic segments.
Which segments to score for Conversion Value Rules
The sample questions above can get you thinking about Conversion Value Rules to create, but you may quickly get stuck on deciding for which locations or audiences to answer these questions. That’s where a good PPC management tool like Optmyzr can help.
Optmyzr’s new tool for Optimizing Conversion Value Rules starts by asking advertisers to rank the typical value for each of the highest volume locations and other segments detected for a site.
The tool also helps solve the challenge of deciding a good value for each rule. It helps with a question like: if a customer from California is worth more than average, exactly how much more valuable are they? The good news is that VBB will work even if your answers are not precise. Just creating a Conversion Value Rule that says a conversion from California is a bit more valuable than typical will help steer the engine’s AI automations in the right direction. It’s like giving it a nudge that says if all else were equal, it should try to get more conversions from California.
To make this scoring process easier, Optmyzr asks advertisers to rank every segment on a scale of 1 to 5. It can be a bit jarring as a data-driven marketer to be asked for a gut-based judgment call, but like Google’s mantra of “don’t let perfect get in the way of good enough,” the beauty is that this type of optimization works well as an iterative process rather than a quest for instant perfection.
After ranking around 30 segments, the tool will have enough data to create an initial batch of Conversion Value Rules which will teach Google’s AI how to get better conversions for your company.
Determining the right Conversion Value Rules
After you’ve thought about the relative value of different conversions for a business, the next step is to translate those insights into rules. Remember Conversion Value Rules can be for a single attribute, like just location, or for combinations of segments, like location + audience, or location + device.
These combinations can be complex to figure out and cumbersome to maintain but Optmyzr’s tools can help with this too. Using the principle of the wisdom of the crowds, it uses scores from you and your team to come up with a sensible set of Conversion Value Rules. For example, an advertiser who values conversions from California a lot and who also sees more value from mobile conversions may see a value adjustment of +20% for that combination.
By setting Conversion Value Rules like this in Google, Smart Bidding strategies like Maximize Conversion Value with an optional tROAS can go to work to find more of the highest quality conversions.
In modern PPC, where bids, ads, and so much more are automated, advertisers can still get an edge over their competitors. This requires taking true-and-tried principles like solid bid management and knowing the new ways to optimize these levers. Value-based bidding is the modern way to improve bidding. And thanks to innovations from Google and Optmyzr that make optimizing Conversion Value Rules easier than ever, better-performing campaigns are now well within any advertiser’s reach. If you’re interested, you can try Optmyzr free for two weeks.
The post How to get better leads and conversions with Google’s AI appeared first on Search Engine Land.