Personalization: it’s pretty much the holy grail of modern shopping.

Customers spend more when they get it. Businesses that excel at it can see a 6-10% revenue increase. When it comes to automotive, a Cox study found that a whopping 90% of car shoppers prefer a customized journey. And yet, Business Insider found that across industries, only 22% of customers are satisfied with the personalization they receive.

There is incredible opportunity to grow and earn customer loyalty by personalizing the entire car buying process, and it starts with your website.

How do you personalize a digital experience?

There are essentially two ways: rules-based, and artificial intelligence- based. We’ll run through these two different methods and explain why AI has so much potential to revolutionize personalization.

What is rules-based personalization?

Rules-based personalization means responding to customers according to pre-set rules for different segments. For example, I might segment my online shoppers into groups: first-time site visitors, or viewers of a particular VDP, or service shoppers. Then I can set up my rules system to respond to those visitors in relevant ways. First-time visitors will see a store-wide sale offer, or be invited to chat with a sales rep. VDP viewers can be directed to a special offer on that vehicle. Service shoppers get $10 off their next oil change. You get the idea.

Rules-based can also be used for email marketing. I might segment my client database by pre-sale or post-sale, or by sale or lease, and then target those segments with appropriate content. For example, post-sale customers can receive service reminders at pre-set intervals, which can be helpful for them, and help you retain more business.

So far so good, right? Rules-based personalization certainly can do a lot in moving the shopper forward by showing them relevant content.

But let’s look at our email customers again. It’s all well and good if I have five shoppers who all bought the same truck within the same week– I create a segment and target them with the same follow-up. But what if I sell 20 cars that all have very different service needs? Or, looking back to our website visitors, what if my VDP viewers are not all the same? Maybe one is doing initial research, and maybe a different one has visited that VDP three times, is only interested in leasing, and found my website by clicking on a Facebook ad. Those two shoppers don’t want the same thing- and you would never treat them the same way if you were talking to them in-person.

You can set up a rules-based system to recognize a lot of factors and combinations. But there’s a limit to how many permutations you can take into account. That means rules-based personalization works in groups, not down to the individual.

That’s where artificial intelligence comes in.

What is AI-based personalization?

Artificial intelligence, simply put, is when a computer does the work of humans. One of the major applications of artificial intelligence is machine learning, which is when machines are trained not only to do the work of humans, but also to continually learn and improve themselves. So a machine learning system uses data to learn patterns and behaviors, and improve its performance based on this input. One more definition: predictive analytics, a machine learning system that uses data to learn, make predictions, act on those predictions, and learn from the results. Predictive analytics is what we’re really talking about when we say AI-based personalization.

Here’s an example. Let’s say the visitors to my site are not converting on a particular offer when I show it five seconds after arriving on-page, but they are converting when I show it after 10 seconds. Machine learning can understand this and optimize timing for future customers. Another example: our shopper above, the repeat visitor interested in leasing. Predictive analytics can optimize the best interaction based on all these factors together.

Here’s the crucial point: AI can analyze huge amounts of data and respond based on them– all in real time. It doesn’t need much input from humans and it can personalize at a massive scale.

Practically, what does this mean? Here are three examples of what AI-based personalization can do on dealership websites:

1. Fuller understanding of customers for more relevant and specific interactions

Every person is unique, and machine learning can understand and respond to this. Instead of taking into account one or two or five factors when figuring out how to respond to a customer, it can take everything into account– every piece of information about that person that it has, whether it lines up with other customers or not.

2. Constant optimization based on shopper behavior

AI-based personalization takes into account all the data it has about past customers to predict what future customers will want. It then observes and reacts to those new customers’ responses, constantly fine-tuning its interactions. In this way, an AI-based personalization system on your website is constantly improving its own interactions with customers.

3. Follow-up interactions

When a rules-based system presents a customer with an offer and they convert, that generally ends the conversation. But because AI can respond to customer behavior in real time, it can keep the conversation going, providing value and moving the shopper down the funnel. For example, if a customer converts on a vehicle offer, a predictive analytics system can respond to that and encourage the next step, such as starting their trade-in valuation, or booking an appointment. And if the customer takes that next step, it can encourage the next one. This keeps the engagement going, building your relationship with that customer as you help them complete steps along their car-buying journey.

AI offers amazing potential to reach customers. If you want to learn more about our AI platform for dealers, check out the video below. And if you just want to play with a fun AI site, check this out.

 

Leave a Reply

Your email address will not be published. Required fields are marked *