Five surprising insights Brinks Home learned through AI testing

As a retention marketer, Lauren Gengler is no stranger to A/B tests. Brinks Home™, a leading home security company, offers homeowners contracts for monitoring and protection. Lauren tries to make sure that Brinks Home’s customers renew their contracts. Like any lifecycle marketer, she experiments with email campaigns to see what resonates with customers. And like most marketers, she grew frustrated with the slow pace of A/B testing.

In an effort to accelerate testing and learning, Brinks Home turned to OfferFit's AI testing. OfferFit’s platform uses a type of machine learning called reinforcement learning. Lifecycle marketers use OfferFit to experiment along the dimensions they choose – such as subject line, creative, sending time, and cadence – to discover the best offer for every customer. 

The campaign was a smashing success – Brinks Home grew contract value 457% over their “business as usual” campaign, a $5M annual benefit. And they learned a few things along the way. Here are 5 insights Brinks Home gained through AI testing.

Don’t @ me – names in the subject line lower open rates

Brinks Home found that putting the recipient’s name in the subject line lowered open rates. “People think it feels personalized to use their name,” Lauren points out. “But it’s been overdone.” These days a name in the subject line doesn’t feel personal, it feels like spam. It’s a clumsy way to try to make a personal overture. 

“Marketers need to expand their ideas of what it means to personalize,” says Lauren. “If you send an email when someone wants to read it, that’s personalization.”

The right time of day depends on the day

And when do customers want to read an email? Well it depends on the customer, of course. But it also depends on the day. Brinks Home found more success once they started varying sending times based on the day of the week. On weekdays, send the email early – the customer might read it before work. On weekends, send the email later – people aren’t ready to consider their home security contract first thing on Saturday morning.

This insight might seem like common sense, but it’s important to remember that Brinks Home isn’t guessing – they experimented and found what worked best for their customers.

Emojis: do it, but don’t overdo it

Emojis are fun, but do they increase open rates? And how many emojis are enough? Here common sense can’t guide us, the only way to figure out what customers will respond to is to experiment and discover what works empirically. 

Brinks Home found that emojis in the subject line improved open rates, but some emojis work better than others, and they need to be used in a way that seems natural. Too many emojis drove open rates down. “We needed to stand out, but continue to be ourselves,” Lauren explains.

We’ve all figured out what $9.95 means – just say $10

It seems to be a law of marketing that every price needs to be quoted one cent or five cents below the dollar. But Brinks Home found that their customers didn’t respond well to $24.99, and “10% off” didn’t perform well either. After all, who wants to solve a math problem just to learn how much a discount offer is worth? Just give the dollar amount – a whole number, no percentages. 

Now when will gas stations get the memo?

You can raise prices, sometimes

The most surprising finding from AI testing was that offers which raised rates sometimes performed better than discount offers. The most profitable offers for Brinks Home were renewal offers with a rate increase – a raise in rates together with a promise not to raise them for a fixed term into the future. Many customers understand that paying to lock in a rate sometimes makes sense – just ask anyone who’s gotten a mortgage while interest rates are rising.

“Using OfferFit gave us the confidence to test new ideas,” Lauren explains. “If it doesn’t work, we learn from that and try something different. If you’re A/B testing with your own time as a marketer, this risk is greater, because it takes so long to find out what isn’t working.” Without OfferFit’s AI testing, the team at Brinks Home might have never tried the rate increase strategy, because it seemed too risky and too unlikely to work. 

Use caution when implementing this blog post – your customers are not a monolith

The insights above were generally true for Brinks Home’s customers, but they might not be true for yours. And more importantly, your customers are individuals. Brinks Home has learned that most customers are more likely to open an email with the right emoji, but that doesn’t mean everyone will. Using OfferFit, Brinks Home is able to give each customer the offer most likely to resonate with them. Learning from general trends is important – Brinks Home has retired subject lines with the recipient's name – but heuristics are no substitute for making a 1:1 decision for each customer.