How could machine learning affect political campaigns?

Machine learning in the SEA

Published by OSG Team May 7th, 2018 Category: SEA

Copyright @ maxuser


Machine learning in SEA is crucial to get deeper insights into the performance of SEA campaigns and to use this knowledge to automatically optimize the important functions. Through a thorough real-time analysis of big data, you can Bids adjusted automatically, create unique and powerful advertising messages and target audiences based on factors such as B. devices, content, geo-targeting, time of day and even the operating system can be precisely determined and addressed.

Machines can collect and process information faster than humans. However, in order to gather the right insights and optimize campaigns accordingly, we as humans need to develop the right strategies. In short, using machine learning without the right strategy leads to failure and, in the end, wasted budget.

The pressure for paid search to drive growth has never been greater. Companies expect their online marketing managers to understand how paid search integrates into the overall business plan. How does attribution affect the performance of other marketing channels and how can you prove that pay-per-click (PPC) has a direct impact on sales and profits? We will elaborate and clarify these questions in the following article

What is machine learning?

Machine learning, or in German machine learning, basically means that a computer no longer needs to be explained exactly what to do. Computer programs based on machine learning can use Algorithms independently find solutions to new and unknown problems. In the past, computers followed commands that were given (coded) by the technician or programmer. Instead, we feed the computers with data and, as it were, animate them to do so independently in this data recognize recurring patternswithout a person having to specify or program them beforehand.

Artificial intelligence enables us to solve increasingly complex problems faster and faster. All this difficult figures, like the data analysis or interpretation of the data is done by the computer, because they are more efficient and faster. Machine learning in the SEA uses information Historical data For example, to make planning recommendations or suggestions for budget setting in AdWords. These can optimize the ROI. So, based on historical data, you can see how adjustments will affect in the future. Before making the switch, you will find out which investment is worthwhile and which is not.

These algorithms do tasks for us that the average human brain is simply not made for. This allows advertisers to concentrate on the essentials, i.e. on them creative solutions to be found in order to properly address the target group.

How does Google use it?

Google takes machine learning very seriously. In 2014, they bought a UK company called DeepMind for $ 525 million. In 2016, DeepMind lost $ 162 million. Part of that loss was due to the development of Google's computer program called AlphaGo. AlphaGo is designed to beat a professional human Go player (Go is a strategic board game in which one player must use moves to occupy more space on the board than the other). In 2017, AlphaGo beat Ke Jie 3 nil - Ke Jie was number 1 in the world!

Why is that important?

I remember a computer program called Deep Blue that beat Gary Kasparov in chess in 1997. The meaning comes from the complexity of Go. After the first 2 moves there are 400 possible moves, but there are almost 130,000 in Go! In a more directly assignable example, Google uses machine learning to support speech recognition. Four years ago, Google had a 20% error rate translating speech to text. Two years ago this was an 8% error rate, and it is now a 4.9% error rate. The more data it uses, the better it gets.

What is a strategy and how is it developed?

Understanding what a strategy is in order to develop an effective strategy is essential. The Duden defines strategy as follows: "Exact plan of your own approachwhich serves to achieve a military, political, psychological, economic or similar goal and in which one tries to factor in those factors that could influence one's own action from the start ”. Strategy is therefore not a tactic, it is bigger. Strategy is the path you take to achieve goals.

A strategy can be broken down into a three-step process:

Step 1: Assess the overall situation

Assessing the overall situation will be the greatest challenge for your company. There are many different methods here. The best known is the S. W. O.T. Analysis. It supports you in uncovering the strengths, weaknesses, opportunities and risks of your company. The Understanding of the overall situation provides the background and context necessary to develop proper goals. Well-defined goals and important performance indicators are the cornerstones of a successful SEA campaign. Every advertiser needs to have a well-defined understanding of what success looks like to them.

Step 2: Develop Account Management Policy

Once the assessment is complete and an in-depth understanding of the overall situation is achieved, it is time to do a Develop a policy for managing the AdWords account. You should ask yourself the question, "How do I want to integrate AdWords into marketing and what is its function?" Depending on whether the AdWords campaigns are intended to increase traffic, profit or increase conversion, or whether it is just a point of contact that leads to conversions via other channels, you have to choose the campaign types carefully.

Step 3: Create an Effective Plan of Action

The next step in the strategy development process is the Development of an action plandesigned to overcome the greatest challenges a company faces.

Let's say you have an AdWords account for a university. Universities always need students. With this information, you can guess that the primary challenge is to find more potential students. Therefore, the guideline could be to use paid search as a driver for lead generation. The action plan in this example would support the strategy focusing on the Focus on increasing reach and remarketing.

The tactic used will now result directly from the action plan. For example, the tactic for increasing reach could be to set up Google Display Network, Facebook and Instagram campaigns. The tactic for remarketing could be: “RLSA campaign with focus on“ shopping cart abandoners ”(in the case of lead generation this would be the premature abandonment in the registration process). In our article on creating an RLSA campaign, we show how such a campaign can be set up and optimized.

Become a PPC strategist

Machines do not make the work of the PPC professional superfluous. So don't be afraid that your job will be taken away from you. However, the advent of machine learning in the SEA and its ability to automate important tasks will be the Change the type of jobs. The value of the PPC professional will be directly related to their strategic ability. Those who see the big picture, fully understand the market in which the company operates, and combine PPC with the overall growth of an organization's overall digital marketing program will win. Machines can analyze data, spot trends, and perform routine tactical tasks much faster than humans ever can. So they take the basic work from humans and provide the basic data for developing the strategy. Because interpreting the data and deriving recommendations for action will be reserved for humans. It is up to you to interpret the machine's data and to derive and optimize the right strategy from it.

Below are some tips to help you as a PPC manager become more strategic. So you can develop strategies to increase the success of your company. Hold the The goals of the company and those of the potential customers are always in view. This will reduce misalignment and provide the understanding needed to gather the right data and use it properly. Also try to understand what the obstacles to goal achievement can be. Is it a strong competitor, high CPC, or low conversion rate? A thorough understanding of the obstacles will help you plan.

Planning resources & selecting technologies

You created the strategy and goals. Now it is important to determine the resources that will be needed to achieve the goals. Depending on the size of the account, the complexity, and the budget, you'll need to determine what resources or technology you need.

Machine learning in SEA and automation currently play a major role. This role in PPC will continue to grow in the future. Paid search and paid social functionality is more complex than ever. The number of platforms and networks that need to be managed is constantly growing. This dynamic is a challenge for PPC managers who need to manage their accounts efficiently.

The right technology provides the ability to automatically analyze large data sets and automate routine tasks to solve complex problems quickly. By providing machine learning in the SEA and automation for PPC accounts, PPC managers are virtually released to focus on the focus on strategic planning and implement strategic initiatives that lead to new growth opportunities.

What kind of technology should I be using?

The answer to this question is, "It depends." Budget and account size certainly play a big role when deciding whether to use a third-party technology solution or the free tools that the advertising platforms provide. There are dozens of technological solutions in the market, ranging from lightweight reporting platforms to complex machine learning and automation solutions.

In addition, the advertising platforms offer automated bid management functions. They also offer the ability to pause keywords, ad groups and campaigns based on specially defined criteria. Scripts can be provided through Google AdWords that allow paid search accounts to be integrated into an organization's inventory or CRM system.

Consider the following points when deciding whether or not it makes sense to deploy technology:

  • Can free automation tools help achieve the goals and execute the strategy effectively?
  • Do chargeable tools offer special functionality that is not possible with other tools?
  • Are you saving a lot of time using technology?

Better performance creates even better performance. Advertisers who use artificial intelligence optimized algorithms for campaign management find that the Continuous improvement in performance. That's because machine learning algorithms are actually getting smarter and can bring this back into their performance. It is this continuous learning aspect that creates the benefit of an SEM marketer. Because if you know how to use machine learning in SEA and work with it, you can competitive stay.

The advantage over humans

Humans are complex beings that are capable of working in the human brain to process immense amounts of data. Still, we cannot devote all of our neural connections to continually improving PPC performance. Even if we could, we probably wouldn't want to, so we made machines. Machine learning algorithms process more data than any human at the same time and use more data than ever before. Another advantage of machine learning optimization compared to the human counterpart is that machines never get tired. You go to work 24 hours a day and keep bringing home results. From this perspective, machines are an obvious choice for certain aspects of campaign management.

4 ways to improve Google AdWords campaigns using machine learning?

1. Target CPA - Whether you are using the AdWords UI or an optimizer like DoubleClick or Marin, you can set up CPA bid algorithms. These algorithms can save you the time and effort of managing your CPC bids and modifiers for things like device, location, and interest.

2. Ad adjustments - Unless you are using ad customization features, this is what you should do. These can be used to build large quantities of keyword-specific creatives. Post dynamic search ads. That way, you can automate the creation of large numbers of ads and provide more data to Google's machine learning algorithms. This will help you decide which creative to look for and when.

3. Similar target groups (similar audiences) - Google uses machine learning to create similar audiences and continuously improve them. Again, the more data Google gets, the better that data gets. Check out your “similar-to-target groups”.

4. Target groups ready to buy - Google uses machine learning in the SEA to predict when its users will be in the market and ready to buy certain products. It is worth testing the campaigns for this, especially on the Google Display Network.

5. Attribution - Google already offers the free Google Attribution tool in the BETA version. This will use machine learning to suggest cross-channel budget adjustments. Make sure to use the tool as soon as possible once it becomes available to your account.

Save time when adjusting CPC bids

Bid and budget management is an area of ​​campaign management that is often tedious and time consuming. Due to the ever-improving algorithm of AdWords bid strategies and AdWords scripts or third-party administration tools, one can therefore use the Make work even easier.

Bid and budget management is the first area that is controlled under the PPC umbrella by machine learning in the SEA. However, there are other tasks that the machine can do for the PPC manager. Machine learning algorithms are also suitable excellent for standardized tasks in SEM such as recommendations for account audits, keyword selection, ad copy, ad testing, audience segmentation, campaign structure and even identifying missed opportunities.

Machine learning in SEA actually offers PPC managers the opportunity to use a Advantage over the competition to get. However, it is based solely on our ability to trust and adopt these technologies.

Obstacles for machine learning in the SEA

Marketers have not yet massively adopted machine learning in SEA and artificial intelligence technologies. But given the technological innovation in this area, we know this revolution is now underway. Google introduced machine learning algorithms to the AdWords market.

Because there is an “invisible” element to machine learning, humans have difficulty designing and trusting machines. Online Marketing Managers Must trust the “invisible”. For example, if a machine is offering $ 5 for a campaign, we don't really know why. We can see it in the system, but the system won't speak and explain why it made this bid decision based on 30 million data points. So, marketers just have to trust the machine learning system to come up with the best deal for them.

Conclusion

Machine learning algorithms don't mean much without humans making, monitoring, and interpreting them. People and machines work togetherto get better performance results. With this understanding, PPC managers don't have to fear losing their jobs if they use machine learning technologies. In fact, they should include technological innovations that can only improve their performance. This can lead to improvements even if the performance exceeds expectations, the campaign structure is complex, or the budgets are small.