During the industrial revolution humans acted like parts in a machine, stationed along assembly lines, trying to make themselves work as mechanically as possible. As we enter what is now being called the “4th Industrial Revolution” we have come to accept that machines are far better at being mechanical than humans.
In the bustling world of search advertising, where campaign managers balance their time between creatively building campaigns, and mechanically managing and updating them on a daily basis, we are once again spending most of our time filling a role that makes more sense for a machine.
A generation ago, we made the change from a manufacturing to a services-based economy. This shift again changed the nature of the workforce – and marketing in many cases helped lead that transformation. Now, once again the role of the marketer is evolving, and in this case, it is being upgraded.
Many forward-thinking marketers are excited about this transformation, when we can focus on what we do best – innovate – while machines will step in and do what they do best – analyse large amounts of data in order to rationally identify and exploit patterns.
“Big Data and Machine Learning, is the infrastructural beginning to an exciting new era that will enable brands to communicate with consumers through new digital channels in a more humanized way using modern technology,” writes Ranee Soundara for Medium.
While some are still reluctant to embrace new marketing technologies, many marketers are beginning to understand that machine learning is vital to higher efficiency campaigns and stronger results, the next step is finding the right solution.
How Machine Learning Works in Search Marketing
In 2014, venture capital investments in Artificial Intelligence startups, including machine learning, deep learning, and predictive analytics has multiplied nearly sevenfold, from $45M in 2010 to $310M in 2015 according to CBInsights.
As investments in AI and machine learning continue to gain momentum as a consequence of the “4th Industrial Revolution,” the centres of power in the enterprise have shifted accordingly. Functional leaders are now equally responsible for the budgets and the outputs of innovative technology. As Gartner Research famously predicted, by 2017, CMOs will spend more on IT than their counterpart CIOs.
This shift is happening because marketers are getting swept up in a tsunami of data. This labour intensive work of digging through reams of unstructured datasets to try and understand the bigger picture is impossible to do with 130 exabytes of data persisting in the digital universe (that’s 18 zeros for us common folk). Humans are capable of processing at most 1000 terabytes (12 zeros), and we process numbers a lot slower, with something we call human error. Believe it or not, this applies perhaps more to search marketing and campaign automation as much as it does any other area of marketing.
When it comes to accuracy and performance, machine learning is playing in a whole different ballpark, and all those marketers still batting in the little leagues will find it increasingly difficult to stay competitive as their competitors leverage machine learning algorithms more frequently.
What is Machine Learning, Exactly?
Machine learning is a vast subject with many methods and applications, but it is typically used to solve problems by finding patterns that we simply cannot see ourselves, according to econsultancy.
For example, the ad auction is a murky place, where marketers are unsure of where to set bids, how to make adjustments for mobile, and ultimately how to get as many conversions for the lowest spend possible. On top of that, there isn’t enough time to devote to each campaign to ensure it’s maximizing its performance relative to its potential. Using machine learning, AdWords and third party vendors are offering technology solutions that closely follow the ad auction, and learn how to update and adjust bids automatically using historical data to predict the best bids to set according to budget, quality score, competition, and changes in the auction over the course of the day.
The old way of managing ad campaigns reminds me of the old Simpsons episode when Homer Simpson set up a drinking bird to do his work for him. In this case, machine learning algorithms don’t just press the “Y” key over and over, they constantly adapt using the information collected and work to improve performance beyond what humans are capable of.
You can step away from those day to day responsibilities and focus on taking on new clients, developing creative, and improving performance in a more human way.
Two Birds with One stone
The problem most marketers face when running search campaigns is two fold, there’s not enough time or manpower to sit there and adjust bids and budgets for all accounts and campaigns (which reduces the ability to scale up), and second, marketers are struggling to achieve greater results in an increasingly more competitive auction.
In a nutshell, people want to do things faster, better and easier, and the only way to do that is to hand it over the the machines.
Acquisio provides what we believe to be a unique solution for the search market, that allows marketers to focus their time on more productive and strategic initiatives while leveraging the investment we’ve made in advanced machine learning to manage paid search bids and budgets. The outcome is significantly greater improvements not just in productivity, but in campaign performance as well. It’s called Bid and Budget Management (BBM).
“Our machine learning-based, proprietary bid and budget management algorithm is the only high-frequency trading model for AdWords and Bing, adjusting bids and budgets as soon as they are updated by the publisher and predicting what the next bid is going to be – which we can prove drives better campaign performance than other predictive algorithms” explains Bryan Minor, PhD, Chief Scientist at Acquisio.
How Bid and Budget Management (BBM) works
Just as a self-driving car is able to recognize both driver patterns and behaviour in the moment, and adjust to its surroundings on the road, BBM is always conscious of the auction environment, processing millions of calculations and adjustments relating to changes in the auction, time of day and more, to keep your campaigns running smoothly. This results in a better overall campaign performance, all while you take a back seat and let the algorithms drive for you.
In the PPC auction, if you set a bid, that you think is reasonable, and then leave it, the constant fluctuations in prices throughout the day means you’ll likely come back to your account tomorrow and be disappointed with the results. What’s worse, you’ll likely have overpaid for some clicks, and missed out on others.
Many predictive algorithms adjust bids as infrequently as hourly, daily or even weekly. By predicting and adjusting bids every 30 minutes, Acquisio participates in the auction more often than any other optimization solution, and makes more accurate adjustments. This helps drive CPC/CPA down and clicks/conversions up.
In fact, our solution is proven to lower cost per clicks by an average 40%, when looking at more than 20,000 accounts powered over the course of one month by Acquisio. And, with algorithms running to properly pace budget across the full day and the entire month, accounts using BBM were 3x more likely to maximize the full budget without overspend.
And when it comes to time saved, a division of WSI – which boasts one of the largest digital marketing networks in the world – was able to cut out hours, if not days, from their typical campaign management process using BBM.
“We saved so much time with the automation we could shift focus to the quality of our campaigns,” explained Heitor Siviero, Project Coordinator at WSI Brazil.
With marketers focusing on improving campaign quality, and machine learning algorithms running daily to improve performance, clients often see what we call, “x-graphs,” where there’s a noticeable spike in clicks and drop in average CPC after setting up our machine learning algorithms.
With results like these, it’s easier for businesses to attract new customers, and with the time saved on manual campaign management tasks, they are in a better position to take on new clients and scale their operations where they matter: strategy, creative and execution.
The great thing is, our technology allows us to deliver differentiated campaign performance for even the most difficult-to-optimize accounts, including those with very low volume or low spend, a chronic challenge for anyone managing search campaigns for smaller businesses.
Take the Next Step
Whether you’re part of a small local business or a Fortune 500, it’s time to embrace the age of machine learning for search marketing.
This post was originally published on Marketing Tech Blog by Acquisio on July 5th