We’ve been watching the new 2018 Google Academy on Air webinars eagerly. The first session covered ‘What is Machine Learning?’ Here are the highlights of the session…
What is Machine Learning?
Machine Learning (ML) means that software and programs don’t have to be told what to do in every scenario. They learn for themselves what to do by looking at historical data and analysing trends to become a powerful helper. It’s a disruptive technology that some people are fearful of, but there are lots of benefits to it too.
- It helps make decisions a lot quicker and solve more complex problems
- It makes us more efficient and stops wastage
- We can apply our brains to creativity rather than crunching data
Examples of Industries impacted
- Health care – for example helping doctors spot anomalies in scans, meaning people get diagnosed and treated quicker
- Financial services – for example tracking trading volatility and market trends in real time, meaning investments can perform better
- Retail – for example Amazon personal recommendations, meaning you see more relevant products and services that you may be interested in
How does it work?
If you want a computer to recognise a dog on YouTube you can program several layers that become more refined as it progresses through the process.
- The first layer has multiple inputs that help define a dog such as looking at colours and edges
- The third layer conceptualises what these shapes mean
- The final layer, connects all the information to give a probability over whether the image is a dog or not
You might already be using it...
Machine Learning is already part of many of our daily lives. For example, it is used in:
- Virtual Assistance for example Google Home – the more you use it, the more it understands you. It recognises and learns the patterns in your language and voice
- Traffic Predictions – apps like Waze uses experiences of other drivers to give better advice
- Online Fraud Detection – PayPal uses ML to look for businesses that are more likely to be affected by fraud. Pattern recognition helps the service get better.
- Delivery Services – Domino’s predicts, based on past experience, how long it usually takes for the pizza to arrive
- Unique recommendations – Spotify recommends you songs based on your listening history. It can even recommend accurately new music that you probably will like.
- All Google products – they already use it – Chrome/YouTube/Gmail/Photos etc…
Machine Learning is now driving the success of Google Ads. With so much data collected, the power of machine learning means we can crunch huge volumes of data and develop better ways of doing things. This leaves us more time to spend on analysing that data to add business value.
4 examples of this functionality are…
- Google Similar audiences – finds you similar audiences to target by capturing interests, behaviours and likenesses of users
- Smart bidding – determines the most important customers and serves ad at optimal bid and position
- App Campaigns – identifies valuable users and targets where they see the best value
- Data-drive attribution (DDA) – Determines how much credit to assign to each click in the user journey based on performance and conversion paths
How can I get started with ML?
- Work with a Programmer who can use opensource software to develop Machine Learning applications to help your business
- Use systems that are already set up with it, such as Automated Bidding to make your Ads more effective etc
Next Google Academy on Air session…
Automated and Smart Bidding
If you want to know more about how Machine learning features in Google Ads can enhance your campaigns, contact us today.