``````
nodes_teal_fade.png

The Granify E-Commerce Blog

Behavioral Targeting: A Throwback to Old School Sales

behavioral-marketing-computer-loop.gif

Behavioral targeting, machine learning, and artificial intelligence are a bunch of futurist buzzwords that get thrown around a whole bunch. Businesses use these terms to describe the strategies of the future: using data, and a lot of it, to do business. But is it really so futuristic? I think not.

Rather, behavioral targeting backed by machine learning is the new and improved version of old school sales.

Don't get me wrong—what we’re talking about does use groundbreaking technology, but the intent of behavioral targeting, machine learning, and artificial intelligence is, at its root, just the classic sales strategy "know your customer."

Claim your free consultation today to learn more about machine learning e-commerce optimization.

Before Machine Learning, Sales Had to Get Less Personal as It Scaled


Getting the right message to the right customer at the right time has always been part of the core sales cycle. Long before the web, a good salesperson got to know a customer. Once you knew a person, you knew what they actually needed; and a great salesperson fulfills a need for a customer.

But this does not scale. If you have a potential market of a million people, it’s impractical to hire enough salespeople to get to know every single person.

One of the earliest ways to scale sales and marketing was with mailers: collect a bunch of data about certain markets and flood every single person with a marketing mailer or coupon. It was a numbers game. If you send out 100,000 coupons and you get a conversion rate 2%, you have made 2,000 new customers. Need another 2,000 customers and find a similar market and send out another 100,000 coupons.

Over time, direct marketing got smarter. Through market research they were able to segment populations. This lead to better targeting, which most of the time lead to higher conversion rates. So now you can send 50,000 coupons and get a 7% conversion rate making 3,500 customers. More customers, less cost? Yes, please!

Slowly, Targeting Got Better, but it Has Had Big Challenges


This successful strategy translated well to email and the web. When users register on a website, they are ripe with data points, like location, purchase history, age, and whatever a user would offer up. You could easily bucket users into a segment and send them a message to bump conversion rate.

But all of this segmentation and bucketing can’t compete with the personalization powers of a regular salesperson. Yes, it solves the scale issue and can improve all sorts of numbers, but it also leaves so much on the table.

Think about it. You could segment and bucket an urban high-income male to assume he likes expensive cars more than laundry detergent coupons. But if you really knew him, you would know that this gentleman despises the high cost of cleaning clothes.

Another problem with this kind of customer engagement is that it depends on the user to give you their information. If they don’t actually log in, fill out a profile, and continue to log in every time they visit your site, you don’t even have enough information to effectively segment them in the first place.

So we return to the original problem: how can we scale getting to know a person? And how can we do this when most of the traffic on a website is anonymous?

This is where you’ll want to start using some fancy buzzwords like behavioral targeting, machine learning, and artificial intelligence.

Understanding Machine Learning Behavioral Targeting

A good way to begin understanding behavioral targeting is by thinking back to the brick and mortar retail experience. A person walks into a store, and the salesperson sees them and has no initial information about them. As the customer browses the store, the salesperson watches them and learns what items catch their interest, those they show little interest in, what they put in their cart, and the many nuanced details in between.

The salesperson notices that the customer has just less than $100 worth of stuff in the cart and that they have been looking exclusively at the sales rack. Being a price-conscious shopper, this person would probably be happy to hear about a special that would save them 25% on purchases over $100. The salesperson acts accordingly, and the customer’s satisfaction is high.

Or maybe the discount is buy-one-get-one-50%-off in exchange for filling out a survey. A customer walks in and declares they need to buy just one jacket, and the salesperson observes they are in a hurry. Telling this shopper about the discount is a moot point. The salesperson acts accordingly, and the customer's satisfaction is high.

This is what behavioral targeting does, but at the scale of all your website visitors, 24/7. Machine learning, behavioral marketing technology like Granify uses the customer’s behavior, not just the information they hand over in their profile, to build a model. This model determines whether or not a certain message would drive a person to buy today, and what that message should say. 

Scaling the Old School Sales Method


Machine learning allows us to scale that personal "getting to know your customer” sales tactic. We can watch behavior and learn if, when, and what messaging will be effective. This way, the customer has their need filled, is happy with the purchase, and will come back for more.

 

Request A Free Consultation

SHARE THIS STORY | |

Looking for more? Our readers also liked...

Want weekly insights in to the world of e-commerce optimization?

Recent Posts

Search