Lightning Lessons: Machine learning in 3 steps

Lightning Lessons: Machine learning in 3 steps

Published
May 2019

Machine learning. We hear the phrase every day. But what does it really mean for your business?

As datasets get bigger and people expect more personalized, relevant content, machine learning is helping marketers find undetectable patterns and predict future outcomes. From collecting and cleaning up your data to putting it into action, this tutorial gives you concrete steps to make machine learning work for you and your business.

0:01

Think with Google presents Lightning Lessons:

0:04

Machine learning for marketers

0:05

[MUSIC PLAYING]

0:09

Machine learning: we hear the phrase every day.

0:11

But what does it really mean?

0:15

And most importantly what does it mean for your business?

0:18

At its core, machine learning is pattern recognition.

0:20

As data sets get bigger and bigger,

0:23

machine learning can help us find undetectable patterns

0:25

and predict future outcomes.

0:27

Here are three steps you can take to make machine learning

0:29

work for you and your business.

0:31

Step one, collect.

0:33

As a marketer, your data probably

0:35

exists across different silos.

0:36

By bringing your data together, you'll

0:38

be able to better understand and respond

0:39

to your customers' needs.

0:41

And that leads to growth.

0:43

Once you've got your data all in one place,

0:45

it's time to clean it up and organize it.

0:48

Remember, the quality of your results

0:50

depends on the quality of your data.

0:52

Step two, analyze.

0:54

Now that you've integrated your data,

0:57

you’ll want to build an accurate model you can trust.

0:58

Start with an outcome you already

1:01

know to test for accuracy. Like identifying

1:04

cats vs. non cats.

1:06

The more examples you provide, the more accurate

1:08

your model’s predictions will become,

1:11

recognizing cats of all shapes and sizes over thousands

1:12

of iterations.

1:16

In your business, you might not need to recognize cats.

1:18

Instead, you might design a model

1:20

to predict the probability of churn

1:22

by using millions of examples of previous customers

1:24

who have churned and those who have not.

1:26

Instead of cats and non cats.

1:27

This model will help your business

1:29

create better strategies.

1:31

Once you've built a reliable model,

1:33

it’s time for Step 3: Activate.

1:36

This is where you’ll take the output and use it to segment

1:39

your customer base into high, medium, and low risk tiers,

1:42

providing each with a customized retention offer.

1:45

This will help you determine where and how to spend.

1:47

No, Machine Learning isn’t magic.

1:50

But as people expect more personalized, relevant

1:53

experiences, machine learning is powering the tools

1:56

that’ll help marketers make better business decisions,

1:58

faster and more efficiently.

Everything a marketer needs to know about machine learning