Learn how to most effectively plan, target, and measure CTV campaigns in today's fragmented landscape.
Technology
Services
Resources
Given the modern data environment, improving your PPC results will increasingly rely on processing quality data. Going from good to great PPC performance involves collecting and processing bigger data sets and using machine learning to find and select the best bids. But, despite what you and many other search marketers may fear, humans are an integral part of this process as well. Machines can’t do it alone.
You’re a talented search marketer, but you want to constantly improve, grow, and stay sharp on the latest techniques in digital marketing. Whether you’re the hands-on owner of a SEM program day-to-day, or you’ve stepped away to think more broadly about your digital acquisition strategy, the goal is still the same: turn your PPC channel into an efficient and predictable revenue driver.
But, once you’ve started hitting the ceiling, it’s time to put more of your data to work.
I’m assuming you’re a pretty advanced search marketer, so you have your ducks in a row regarding your program structure. This means you probably have strong campaign build-out, highly relevant ad groups matched with ad copy and keywords that are tested, kept clean, and aligned with a good landing page experience.
You’re also probably driving some significant volume. If PPC is a truly pivotal channel–generating forty, sixty, or eighty percent of your revenue–you’re willing to spend a lot with search engines. Lots of dollars leading to lots of clicks and revenue.
When you’re spending tens or hundreds of thousands of dollars on search engine marketing, you earn a valuable byproduct: data.
This data is a gold mine! Every click results in some set of actions on your website. Some clicks get further down the path towards revenue than others. Beyond that, every click is not only attributed to the keyword, ad, ad group, and campaign, but also to all sorts of attributes and dimensions that provide details about demographics, geographies, and habits.
The intricate customer journeys of the modern era make a “Conversion” in Google Ads worth less. Your own ability to track the path to revenue is so much more meaningful. Your data tells a better story, you just need to capture it, unify it, and process it.
Hip hop artist Drake makes the claim “I don't carry cash 'cause the money is digital”. This resonates with many of us search marketers these days, hosting most transactions (and therefore conversions) online. But an offline conversion in-store is part of the customer journey for a great many buyers. And the ensuing revenue may have come from your search ads. The customer journey isn’t as cut-and-dry as one may suspect in Google Ads 101.
Search marketers generating less volume from their PPC can probably get by adequately looking at impressions, clicks, and conversions in Google Ads for their optimization efforts. However, when you’re working towards being best-in-class for a scaled-up search program, the publisher metrics aren’t good enough. Performance marketers running huge programs require more to truly grow.
Collecting the data we’re discussing here will very quickly generate “Big Data”, which is all the rage these days. #BuzzwordAlert. But big data actually enables exactly what we’re looking at discussing here: a better optimization approach.
Big Data means you get the full picture. You get statistical correlations. You get to see reality, and no longer need to ask yourself “why” something is happening. You have the information necessary to tell you what is most likely to happen under any given circumstance. Seeing the actual numbers-based correlations between objects, segments, decisions and their outcomes is the real gold mine derived from your data! More information shows more precise trends from input to output. Which ultimately enables more specific and accurate decisions for search campaigns.
Generating a usable data set, though, isn’t that easy. You’re likely collecting quite a bit of data, and with some additional tags or tools you can collect even more. But it’s the unification and integration that is key. Unifying data to attribute the customer journey, the conversions events, and the revenue values back to the campaigns, keywords, and clicks that sourced them is required before any meaningful optimization actions can be used. It’s a necessary step to accomplish our goal: maximize performance.
Human beings, despite extreme intelligence and ability to reason through situations, aren’t capable of taking advantage of this big data set on our own. We need other tools as a “middle-man,” and I’m not talking about spreadsheets. We need something that can process data every day, note trends, and learn how the relationships are situated from click to revenue. I’m talking about letting the machines take over.
I’m a human, and if you’re reading this I presume you’re also a human. So you and I probably share the desire of not wanting machines to take over the world. Or take over our jobs. However, I feel very strongly that I do want machines to help make my life easier, through automation.
Machine learning is the type of artificial intelligence that should take over here. And it should automate specific parts of your program:
Machine Learning is the obvious choice when dealing with big data in a field like search engine marketing. However, other automation is super valuable to put those calculations to work. SEM automation crunches your online and offline data to learn. Then, it pushes bids, bid modifier adjustments for different dimensions, and a host of other efficiency tactics.
Don’t be fooled by all the discussion here of collecting data, processing it, and automating through machine learning. Nothing here can happen in a meaningful way without human intervention.
Human experience has been, and will be, the glue that keeps SEM programs running effectively. Machines enter to help save time, run more robust calculations, and execute better decisions from the available data. Machines take grunt work off of our plates, but not the strategic and directional elements that only humans can accomplish.
Frankly, it’s something we have to be thankful for! We getting better results when the machine is processing data and making bid calculations from the trends. And we’re also freed up to think more creatively about the art and strategy for hitting our marketing goals.
It can feel unfamiliar to not be “in control” of every bid decision. However, relinquishing that control need not be too difficult once you reason through the cost/benefit. It’s a simple calculus: algorithmic bidding executes far better for at-scale programs with big data sets. And better results benefit your business!
As an experienced search marketer, you likely focus on improving performance and increasing your bottom line. PPC has tremendous room for growth given the amount of data we can now capture about every stage of the customer journey. Your integrated data, particularly data pertaining to revenue, should be working for you.