Book a Meeting
Referral & Loyalty Program for All Types of Businesses
Book a Meeting

Share this article

Table of Contents

A lot of decisions are made during the initial setup of a referral program—the language used in the call to action, which graphics to include, the referral offer, and more. Although it is possible to nail it the first time, it is always worth testing a few variants to get the most from your referral campaigns. The power of A/B testing allows you to optimize your referral funnel so you can maximize return on investment (ROI) and grow your customer base.

The problem? Keeping track of all of the variants along the referral process can start to get out of control. One thing leads to another and next thing you know you're elbow deep in spreadsheets and an arbitrary conversion rate.

That's where Friendbuy comes in.

Friendbuy's A/B testing feature will allow you to test different offers, creative, design, messaging, and much more to understand what customer touch points drive higher engagement and conversions for your referral program.

Friendbuy has simplified the process of creating variants, measuring against specific KPIs, and implementing your champion. We have removed the barriers to A/B testing so you can refine your referral funnel like a pro.

Let's go through how A/B testing works in referral marketing and how to achieve statistically significant results (AKA more sales and revenue).

What Is A/B Testing?

A/B Testing, also known as multivariate testing or split testing, allows you to create and test different variations of your marketing efforts to see which will give you the desired outcome (such as more signups). Then each variation will be split randomly between website visitors, where the term "split test" comes from. Split testing allows you to randomize your variants so that you can get clean, reliable data based on visitor behavior in real time.

All the elements of your referral program can be tested. These include CTA placement, CTA copy, graphics, referral offers, and more. A/B testing allows you to change individual elements of your campaign to see which version performs best. Each version is referred to as a variant.


So for example, you can find out which referral rewards get more signups, such as cash rewards or percentage discount coupons. You can also test copy, such as dedicated referral program email subject lines or on your referral landing page.

This way, you can be sure that you're getting the biggest bang for your buck with your referral program.


See how to set up A/B testing in your Friendbuy dashboard here.

In general, the best practice for A/B testing is to limit the number of changes you wish to test. For example, try either a copy change or a referral offer change, but not both at the same time. Once you identify what you wish to test, you then set up the new variant and compare it against the default variant, which is typically the current version in production.

It is possible to create and test more than two elements in a single A/B test with our solution, but it may be more difficult to identify which factors contributed to the success of one variant over another. This will create "false positives", which will muddy the test results and potentially lead you down a path that doesn't increase revenue.


See how to set up A/B testing in your Friendbuy dashboard here.

After creating variants and having the test run for a select period of time, you can then choose which variant is the champion based on results and determine if you wish to make the champion the default version for production. These results could be an increased click through rate, share rate, or conversion rate.

The duration of the test will depend on a few factors, such as how much traffic you get. The more traffic you get, and larger the sample size, the quicker you can collect data.

For example, a high-traffic site will generate faster results, whereas a site with lower traffic will take a while to collect enough meaningful data. A/B testing helps you optimize your entire referral funnel to get the most from your referral program.

Once all the data and your test results come through, you can adjust your referral programs to reflect the "winning" test variations. And with Friendbuy, you can be certain that you're working with reliable data, meaning your referral program can be tweaked and improved with confidence.

Easy peasy.

New call-to-action


A/B Testing for the Top of the Referral Funnel

The top of the referral funnel is where the most people will interact with your referral program through impressions and clicks. This is where your loyal customers (Advocates) and new customers will enter your referral program for the first time.

Let's start by looking at how we can test different elements that your loyal customers will see first.

Our top recommendation to our customers is to include some kind of share button or website ribbon on their homepage so that all website visitors see your referral program when they come to your website. The key to a successful referral program is letting your customers know that your referral program exists.

Once your share button or website ribbon is up, you can optimize with A/B testing by making changes to the button or ribbon, such as:

  • Size
  • Position on the web page
  • Color
  • Call to action (CTA)

For example, flower delivery company The Bouqs Company is testing the CTA on its homepage by comparing “Refer, Get 30% Off” to “Get 30% Off” to see which offer resonates most with customers.

Screen Shot 2021-03-22 at 1.33.39 PM

Screen Shot 2021-03-22 at 4.28.17 PM

When promoting your referral program through email, test factors such as:

A/B Testing for the Middle of the Referral Funnel

At the middle of the funnel, you can use A/B testing to optimize the Advocate share rate and the friend click rate. 

To optimize the share rate, choose one element at a time to test, such as:

  • Primary CTA to the sharer (referrer or Advocate)
  • Widget header
  • Background hero image
  • Ways to share
  • Advocate rewards

DTC nail care company Olive & June used A/B testing to determine which sharing method was most compelling to customers. After verifying that most Advocates wanted to share referral offers with friends and family via personal referral links, the company made the Advocate PURL more prominent in the widget.

You can also use A/B testing to optimize referral visits and increase the friend click rate by comparing variations of:

  • Friend rewards
  • Images in referral share (email, social, text)
  • Copy in referral share (email, social, text)
  • Subject line, copy, and imagery for email messages

Premium scrubs brand FIGS tested its shared email subject line with two different variants:

  • Control: {[sender}} shared FIGS with you
  • Test: Your friend thinks you'd love these scrubs - here's $20

The control variant won by a landslide with almost twice as many clicks and four times the revenue. This is a perfect example of how much of a difference a few words can make when it comes to refining your referral program.

A/B Testing for the Bottom of the Referral Funnel

At the bottom of the referral funnel, you can use A/B testing to optimize conversions by experimenting with what referred customers see on your web pages, such as:

  • Friend rewards
  • Product recommendations
  • Referral codes that require the user to copy/paste versus direct access to the offer via a link that applies the discount at checkout

This is where you can test the value proposition to find what gets your target audience to convert. And once the referred Friend clicks on the referral link, you can test all the elements on the referral landing page too.

Little by little, you can test the entire referral process to get more referrals with every iteration of your referral program.

How Casper Generates Greater Return With Referral Marketing


Friendbuy’s Simplified A/B Testing Process For Successful Referral Programs

Running tests is Friendbuy's bread and butter. You can perform A/B testing at any stage of the referral funnel in just four simple steps:

  • Select the referral widget or share button you wish to test—this will be your default variant.
  • Create a second variant by duplicating the selected widget or share button.
  • Decide which element you will be testing and edit the variant.
  • Assign weights to each variant—typically we recommend setting the default to 80 and the new variant to 20 for a period of time.

With Friendbuy, no developer resources are required for A/B testing. Marketers can easily complete the process in the platform, giving you total freedom to optimize your referral funnel. Any element of a referral widget or landing page can be tested, including:

  • Heading
  • Subheading
  • Preview content
  • Hero image
  • Product graphics
  • Call to action
  • Share buttons

We’ve been laser-focused on referral marketing for a long time, and in our experience, applying these pro tips will help you get the most from A/B testing:

  • Test one element at a time. 
  • Work on big changes first and then refine details.
  • Start at the top of the funnel and work your way down.
  • Have reasons for testing—don’t just test random ideas; test things that you genuinely think will work better than what you’re currently doing.

With Friendbuy, you can concurrently A/B test your referral program to see which offers and copy resonate best with your customers.

One thing to consider when A/B testing is the rule of a hundred. If your product's original price is under a hundred dollars, then a percentage based discount, such as 30% off, will appear more attractive to customers. Likewise, if your product's original price is above a hundred dollars, then a numerical discount, such as $50 off, will impress customers more.

A jewelry company called Get Back Necklaces tested this and it made a difference of 700%. They originally offered their customers Give $10, Get $10, but they wanted to A/B test that against Give 50%, Get 50%. As a result, their referrals spiked higher than ever before. Referral revenue of their percentage based offer doubled their dollar based offer, resulting in a 700% increase in referral revenue.

Another example is  Casper, who recently tested their offer strategy to see if:

  1. A higher Friend incentive of 25% off and a lower Advocate offer of $75 off worked better than
  2. A lower Friend incentive of 20% and a higher Advocate offer of a hundred dollars.

Drum roll for the results please...

They found that A performed better, with the higher Friend incentive and lower Advocate reward. When they made that change, they saw a 26% lift in shares and a 16% lift in conversion rate.

That's an important reminder to not rely on what you think is the best offer for your customers. Let the data guide your strategy.

In terms of timing, testing your referral offer strategy and creative now is a great way to get prepared for the holiday season ahead to make sure your best converting creative is live during peak times.

Keep in mind that you want to encourage repeat referral behavior. Some of your customers will refer a lot, so provide an incentive that encourages these high performing Advocates to keep referring.

For example, consider offering an account credit towards a featured purchase or maybe a third party gift card if your product category calls for that. As a counter example, a free t-shirt with your brand logo might sound nice, but how many of those does someone need and will they keep referring after they get one? Probably not.

Finesse Your Referral Funnel with Friendbuy

Whether you’re a referral marketing pro or just getting started with this channel, Friendbuy is here to help. In addition to all of the technical support you might need, our referral experts will also help you refine your referral marketing strategy. We want to see you succeed, so if you feel like your referral campaigns need a little attention, just ask for our advice. For more pro tips on maximizing your referral program, download The 2023 Referral Marketing Blueprint for E-Commerce.

New call-to-action