Making Personalization Personal at Scale

No customer will ever say: “Please, whatever you do, make sure the experience isn’t tailored to my needs. In fact, make sure it’s occasionally frustrating.” 

It’s not a thing. It will never be a thing.

We’ve all waited in long lines, had customer service phone calls that should have ended a long time ago, and dealt with buttons that should’ve worked when they were clicked.

We just want to do the thing we want to do and go on with our lives.

And yet, here we are. Ads that we wouldn’t click on in a million years. Customer experiences that should have been fixed years ago. Promotional emails that we should have never received.

It’s a problem. A big problem.

Over 90% of customers say good service makes them more likely to purchase from a company again, and 80% of customers feel the experience a company offers is as important as their products and service.

But the problem isn’t caused by companies not caring about their customers. They have no choice but to care. The reality is that most organizations can’t scale their personalization efforts because of their legacy IT infrastructure and the overall complexity of personalization when they have thousands of customers.

Here’s an approach to creating personalized experiences at scale


1. Segment your customers by behavior.

Behavioral data drives personalization. You can’t personalize at scale without it.

But, if all of your customers look the same in a traditional database, then you can’t personalize your experiences. It’s that simple. 

We use tools like Mixpanel and Amplitude to understand customer behavior so we can begin our segmentation process and understand which customers belong in which buckets. We recommend starting with at least 6–8 behavioral-based segments.

We group customers with similar needs and behaviors first. 

For example, mothers who shop for only one brand for their children vs. fashion-conscious young women who buy private-label styles. 

They might shop at the same store and maybe even buy the same things, but understanding their underlying behaviors is what makes these segments so powerful. It might seem small, but this makes a massive difference.

2. Understand the customer journey and touchpoints.

This is the only path to getting a true 360-degree view of the customer. It’s great if you know what customers bought, but do you know how they got there? Do you know all of the touchpoints that led to that purchase?

It’s important to take your buyer segments to the next level to understand the customer journey for each segment, from initial consideration to purchase and use to subsequent purchases.  

Amplitude allows us to combine segments and customer journeys to create hundreds of smaller segments for even further personalization. The more precise we are, the more we can start impacting metrics.

3. Build a small, powerful, rapid activation team (a.k.a., pirates).

If I can pinpoint one thing that’s changed in the past five years, it’s a mentality shift from “big-bang, let’s get everyone in the room and change everything all at once” to “let’s start small, build value, and THEN go bigger.”

Most companies are slow to respond to customer signals because their legacy systems simply can’t handle it. 

It can take months to record and process customer information and 6–8 weeks to launch the first campaign. By the time the company has launched the campaign, the signal is gone, and the opportunity is missed again.

It’s a constant issue.

What we found that works is building a “war room” with 3–5 talented team members who you know are passionate about the work they do and are willing to experiment to drive results. 

Fewer large meetings for approval and more action with a small, nimble team.

In the end, large-scale personalization doesn’t have to require millions of dollars in tech and marketing investments. We found that the best brands start small, generate top-line impact quickly—in weeks—and then make further investments once the value has been proven. And they all start with understanding their customers through data as their foundation.

Only then does the fun start.