How Libra's team reached 3x sales per agent in 90 days

Your renewal notice goes out. The customer opens it, sees the new price and goes looking for a cheaper quote.

By the time someone on your team calls back, they've found one, or they've stopped picking up. That follow-up depends on your team finding time for every customer, and some weeks they don't.

That was the problem Libra Insurance set out to test. Here's how Libra set up the test, what the model did in each conversation, and what changed in 90 days.

TL;DR

• Libra Insurance ran a controlled A/B test: 2 teams of comparable size, same product, pricing, segments and lead source, for 90 days.
• One team kept its existing process. The other used Sellence for personal text conversations about renewals.
• After 90 days, Libra's team using Sellence reached 3x the sales per agent of the control team, with the same headcount.
• That team started the test behind.
• You can run the same kind of test on your own floor: 2 comparable teams, the same leads, one number, 90 days.

Who is Libra Insurance?

Libra is a digital insurer in Israel, listed on the Tel Aviv Stock Exchange. It sells insurance online and runs its own sales team, much like yours.

A big part of that team's work comes after the quote: renewals, price questions and customers who want another conversation before they decide.

What problem was Libra trying to solve?

At Libra, the work kept going after a customer got a quote. People compared prices, put off the decision or needed one more conversation before renewing.

Following through depended on the team having time to come back to each customer. On a busy day, some customers left with a price or cover question nobody answered.

You've seen this at a good local garage. They'd call every customer back if they could, and the ones they miss go to the place down the road.

How did Libra set up the test?

Libra ran a controlled A/B test on its own sales floor.

• Team A (control) kept its existing process.
• Team B had the same headcount and used Sellence for personal text conversations with customers.
• Everything else matched: the same product, the same pricing, the same target segments and the same lead source.
• The test ran for 90 days, and the number that counted was sales per agent.

What did the model do in each conversation?

It handled the follow-up and the qualification, including the questions that came back while a customer was still thinking about their renewal.

The conversation kept going past the first reply. A customer comparing prices could explain what mattered to them, get a relevant answer and come back to the decision days later without starting again.

Here's what that looks like for your customer. This is an illustrative conversation, with names and details invented, and it's the same example shown on our Libra customer story.

Illustrative conversation

LibraHi, it's Maya from Libra. Your renewal is coming up. Anything you'd like to check before you decide?

Customeri've found a cheaper quote

LibraIs the excess the same on both quotes?

Customerno, it's higher on the cheaper one

LibraThat means paying more yourself if you claim. Want to review your current renewal before deciding?

Customeryes, send me the link

The customer raised price. The reply asked one question that turned a cheaper quote into a fair comparison, then brought the decision back to the cover they needed.

What changed after 90 days?

Libra's team using Sellence reached 3x the sales per agent of the control team, with the same headcount. It started the test behind.

• 3xSales per agent
• +84%Outbound reply rate
• +50%Client engagement

The same test also showed Libra's outbound reply rate up 84% and client engagement up 50%.

"Within 3 months, the Sellence team had tripled their performance."

Put simply: same people, same leads, same product, and Libra's team got a 3x difference in what each person sold. Picture that on your own renewals.

What changed in how follow-up worked?

The test measured sales per agent. Here's what was different about the follow-up itself, and what you'd look for on your own floor.

• Every customer got a reply. Follow-up stopped depending on who had a free hour, which is the first thing to check on your team.
• The reply answered the real objection. A price question got a straight comparison, such as the excess on each quote.
• The conversation remembered. The customer's question stayed in the thread between messages, so nobody had to ask twice.

That's the work Sellence's persuasion model does: it keeps one conversation going for weeks, handles objections like price and timing, and remembers what the customer said along the way.

What can your team take from Libra's test?

Run your own. A controlled test is the clearest way to see what changes on your floor, and it's simpler than it sounds.

1. Pick 2 comparable teams. Similar headcount and similar results before the test.
2. Give them the same leads. Same product, pricing, segments and lead source.
3. Choose one number. Libra used sales per agent, because it's what the business pays for.
4. Run it for 90 days. Long enough for renewals and slow deciders to come through.

If you sell insurance, renewals are a good place to start, because the customer is already yours and you already have their details.

Key takeaways

• Libra ran a controlled A/B test. 2 comparable teams, the same product, pricing, segments and lead source, for 90 days.
• The result was 3x sales per agent. Libra's team using Sellence reached 3x the sales per agent of the control team, with the same headcount.
• That team started behind. The gap opened during the test.
• Follow-up stopped depending on free time. Every customer got a reply that answered their real objection, in a conversation that remembered them.
• You can run the same test. 2 comparable teams, the same leads, one number, 90 days.

Frequently asked questions

What did Libra Insurance test?

Libra compared a sales team using its existing process with a team of comparable size using Sellence for personal text conversations, in a controlled A/B test over 90 days. Product, pricing, segments and lead source were the same for both teams.

What were Libra's results with Sellence?

Libra's team using Sellence reached 3x the sales per agent of the control team over 90 days, with the same headcount. Libra also saw its outbound reply rate rise 84% and client engagement rise 50%.

Did Libra add sales headcount?

No. Libra reported the result with the same headcount. The measured result is sales per agent, not the number of messages sent.

What did the model handle at Libra?

Personal follow-up, qualification and renewal conversations, including price objections, in one text conversation that remembered what each customer had said.

How do I run a test like Libra's?

Pick 2 comparable teams, give them the same leads, choose one number such as sales per agent, and run it for 90 days.