In Google’s latest Academy on Air presentation that discussed the basics of machine learning, the audience gained several insights into all of the following:
- What is machine learning
- How machine learning works
- How machine learning helps marketers
- How businesses and customer use machine learning
- How machine learning is involved in Google’s products
- How your business can take advantage of machine learning
Overall, Google wants everyone to become more familiar with this trend, despite them using it everyday without realizing it. They also wanted to point out the amount of benefits when taking advantage of the opportunity, whether it is improving the ROI for businesses or adding more convenience to customer’s lives. With this presentation, it was Google’s goal to increase the overall awareness of machine learning while urging users to take advantage of the amount of opportunities created from this concept.
Machine learning is an application of artificial intelligence (AI) that provides systems the ability to automatically learn and improve from experience without being explicitly programmed. The term isn’t as young as we think it may be. In fact, machine learning has been involved in our lives for years without us realizing it.
A few examples of machine learning that exist nowadays include:
- SoundCloud’s “The Upload” using machine learning to find new tracks that you may like.
- Intuit using machine learning to make Expense Management for the self-employed more convenient.
- Microsoft using machine learning to help their customers deal with the latest trends in their industry.
- Business startups learning how they can apply machine learning to help their customers make better decisions.
Many of us are already aware that machine learning is designed to help people but knowing how it works is something different. Google provided the example of having the viewers figure out the pattern between a set of numbers. If we can figure out the relation between the numbers, we can figure out the next number in letter D.
After concluding that the answer was 81, Google explained that machine learning operates in a similar way, but the problem is often more complicated. Similar to us, machines learn from experience.
Machines learn when we give them mountains of data, past experiences, and “ask” them to find any patterns within them. Then when we provide them with a new situation, they are able to use these past experiences to handle the new one.
Google has conducted research on machine learning for years. It is a growing trend that is affecting our economy, both on and offline. In a nutshell, it is designed to make our lives easier, such as analyzing data and doing the “heavy lifting”, while humans can focus on other tasks, such as creativity.
Machine learning has already been implemented and benefitted a number of industries, including healthcare, financial services, energy among many others. Machine learning helps us to be increasingly efficient when it comes to problem-solving, allowing businesses and customers to save time and money.
Teaching Machines to Recognize What is in Videos
While humans can easily look at a picture or video and recognize what is happening, teaching a machine to recognize the same concept can be more complicated. But it is done through a technique called neural networks.
A neural network is comprised of different layers of pixels, each layer being smaller than the one below it. There are also neurons, or nodes, that recognize how the pixels relate to each other, making the connection between each one. The nodes then condense the pixels onto a smaller layer, continuing the process until the machine recognizes what is in the picture.
Why is this important? For those researching machine learning and contributing to its overall growth in each industry, it’s essential to know the details. But for others who are just looking to apply the technique to their business, they can use this concept to briefly explain to customers how machine learning works if they asked the question.
Humans are using machine learning every day. But the crazy part about it is that they don’t even know it! Machine learning is involved in all of the following:
1. Virtual Assistants
Having a virtual assistant is almost like having a personal secretary in your home. While it may not understand your schedule or voice immediately after purchasing it, the more you use it, the more it will learn about you personally.
An example that Google had used was talking to it when you are in the other room. It may not recognize a muffled voice when you are further away from it at first, but after using it over time, it may recognize your muffled voice later. This is because the virtual assistant is recognizing the tones, language, and patterns from your voice.
2. Traffic Predictions
For years now, machine learning has been saving millions of drivers time when arriving to their destinations. Instead of simply choosing the shortest route to a destination, the GPS system may suggest a longer route that has significantly less traffic. This way you can skip the traffic and arrive earlier.
How does the GPS know this? It detects traffic on different streets throughout different times of the day, detecting the patterns of traffic. The GPS finds a longer route with less traffic that would allow the driver to arrive earlier.
3. Online Fraud Detection
PayPal uses machine learning for protection against money laundering. This is done through pattern recognition because the most common types of fraud occur in a similar way. So when it detects that a human is involved in this pattern, it asks “did you just spend $x at x place?”
Not only has machine learning notified fraud victims seconds after it happening, it doesn’t require human involvement.
4. Delivery Services
Domino’s uses machine learning for predicting how long it takes for a pizza to be delivered at a particular time and day. This is done through using past experiences to calculate the amount of time it will take for a pizza to arrive.
Customers have also found this method to be incredibly accurate and have become satisfied with the company as a result of the introduction of machine learning.
5. Unique Recommendations
Companies like Spotify use machine learning to create playlists of songs according to the preferences of the user. It learns by keeping track of what you are and are not listening to, what you liked, disliked, and for what you searched.
So these lists can consist of different songs that may have no relation to each other, but they each relate to you because of another song you listened to before.
In the healthcare industry, CAT scans were used to help doctors pinpoint cancerous cells in organs by picking out “dark spots.” But other than the doctor studying the scan carefully, there was no other way to detect cancer and there was a risk of untreated cancer if any “dark spots” were not found.
Nowadays, a new program called computer-assisted diagnosis (CAD) is used to study each pixel within the scan, interpret the data, and direct the doctors to the problematic areas. As a result, everyone is benefiting because doctors are now able to spot these diseases in their early stages, attacking them while they are still vulnerable. Doctors are even saving hours of time when not having to study the scans.
Financial Services Industry
Within the financial services industry, there are mountains of data from the stock market that can easily be fed into machine learning models. The machines are then able to predict the direction of the stock market based on past experiences.
People can then rely on these machines as experienced, honest traders to react to current market trends, making evaluated decisions in real-time. As a result the overall wealth for corporations and stockholders can increase.
In the retail industry, companies like Amazon are able to provide recommendations for customers, based on past purchases. Machine learning also compares other customer profiles that are similar in order to recommend another product to that customer.
So when a customer buys a phone, Amazon will recommend a specific case for that phone because it matches that customer’s profile to others who have purchased the same phone and have also bought that case.
Machine learning is already involved of Google’s products, including YouTube, Chrome, Google Maps, the search engine, etc. It is at the core of everything they do.
1. Google Clips
If you are unfamiliar with Google Clips, it is basically machine learning in a camera. Simply clip the camera to your shirt and it will recognize people and animals. After walking around with it, the camera will have taken pictures of people smiling and laughing, taking pictures of life’s best moments.
2. Google Photos
Google Photos app will create beautiful collages, choosing the best pictures to create memories on your phone. You can also search for pictures and it will show you all the pictures taken that relate to that search term. Google’s app will show you pictures from your query without you telling where or when the picture was taken!
If you have an e-mail with Google and you get a ton of e-mails every day, Google can recommend replies based on determining the meaning behind the previous message.
Machine Learning in Digital Advertising
For marketers using automated bidding, machine learning is customizing customer’s experiences to show ads that are most relatable to them – based on their keyword search history – in the top positions. Machine learning also takes into account purchasing behavior, determining if the user would be a good fit for what the business is selling.
Google had an example of an older and younger brother. Both brothers use the same search terms in Google, but based on past behavior, Google would not show ads in the same order. This is because the older brother have been looking to purchase a product whereas his younger brother only wanted to find information about the product. This allows businesses to save money from wasted clicks.
How Machine Learning is Used in Adwords?
- Google Audiences
Whenever a consumer explores sites and content online, signals are generated which can be gathered and interpreted while maintaining the consumer’s anonymity. These daily interactions across the web, which include searching, browsing content, watching videos, and interacting with social networks, help to understand and categorize audiences. Refined audience data enables you to hone in on the precise user who is most likely to care about your brand or product.
- Automated bidding
Automated bidding takes the heavy lifting and guesswork out of setting bids to meet your performance goals. Machine learning uses millions of signals to determine the optimal bids and continually refines models based on performance.
- Data Driven Attribution
Data-driven attribution gives credit for conversions based on how people search for your business and decide to become your customers. It uses data from your account to determine which ads, keywords, and campaigns have the greatest impact on your business goals. Machine learning algorithms create attribution models based on past performance and paths to conversion.
- Universal App Campaigns
As an app advertiser, you want to get your app into the hands of more paying users. So, how do you connect with those people? Universal App campaigns streamline the process for you, making it easy to promote your apps across Google’s largest properties including Search, Google Play, YouTube, and the Google Display Network. Just add a few lines of text, a bid, some assets, and the rest is optimized to help your users find you.
Google provides more detailed benefits about machine learning and how it helps marketers to save time and money while increasing ROI.
With all of this information and examples about how machine learning works for each industry and business, it’s important to think how you can implement it in your own. Google explains the steps on how to apply it to your objectives and the results you are looking to achieve with it.
The goal is to allow the machines to do the “heavy lifting” while you focus on your business objectives, customers, and results provided to you from machine learning.
Here are some ways that Google’s tools involving machine learning can help you:
- Google similar audiences. Machine learning can connect you with qualified customers using their search trends to help you grow your business.
- Smart bidding. Google will help you to determine who is and isn’t likely to buy from you.
- Universal App Campaigns. Google will help you to invest money only where there is value coming from the customer.
- Data-driven attribution in AdWords. Google strongly encourages marketers to use the data-driven attribution model if it is available to them.
Google Products to be Launched in 2018
At the end of the presentation, Google announced that there would be a number of new products coming out in 2018 that would leverage machine learning to help companies not only increase their online visibility, but grow their business, both online and offline.
Among the new products are in-market search ads, a form of advertising that differentiates customers from information-seekers. It is designed to help businesses minimize wasted clicks, improving their ROI and getting the most out of their marketing budgets.
Work with a Certified Google Partner
After watching the presentation, it’s easy to see how much machine learning can help both customers and businesses. Below is a summary of all the benefits provided:
- Increase ROI for businesses
- Increase customer satisfaction
- Eliminate time spent doing the “heavy lifting”
- Improve overall calculations and solutions
- Provide convenience for customers
- Present growth opportunities for businesses
With so many benefits from machine learning, taking advantage of this new trend is crucial in order for businesses to stay ahead of competition. But of course each business is different with its own goals, target market, services, opportunities etc.; therefore, using machine learning for your business can be challenging.
Proceed Innovative is a proud Google Partner that has years of experience using Google’s tools and products to help businesses of various industries achieve their goals. Our digital marketing professionals can help you leverage smart bidding that allows you to get the most out of your advertising budget. By having the machine learning do all the “heavy lifting,” we ensure to measure detailed results while implementing new strategies.
If you need help making Adwords and machine learning work for your business, contact Proceed Innovative at (800) 933-2402. We will make sure that your business will reap the benefits mentioned above!