fbpx

Gemini Sound’s Conversion Rate Improves 73% with Predictive Personalization

by | Sep 20, 2022 | eCommerce, Success Stories

 

“I would describe Jarvis as a tool. It’s a tool that really works. You’re clearly seeing the click-throughs and conversion rates are higher with Jarvis. It’s not even a question. There’s no reason not to use it.” 

 

For Music Lovers

Gemini Sound is an innovator in the design, manufacture, and sale of high quality sound equipment. David Cabasso is their VP of Product, who runs Marketing and manages (as he phrases it:) “the impression of how our products are displayed and shown to the world.”

Their customers are like you and I: musicians and music enthusiasts who desire high quality equipment to create, edit, share, and play their favorite sounds.

In addition to their direct-to-consumer (DTC) business, Gemini Sound also provides equipment to retail stores, other ecommerce stores, and partners with various distribution channels.

And, well it’s music to our ears… so we’ll just spoil the surprise right now: Gemini launched predictive personalization on their eCom site and grew their conversion rates by 73% – and revenue-per-shopper by even more. 

But please, read on!

Challenges

The Gemini Sound store is, like 1.75M other merchants, built on Shopify. 

Gemini acquires their fair share of traffic (just like we all do… right?) They have been investing significantly in SEO to capture high-intent shoppers. Plus, like the rest of us they rely on paid traffic from the normal gambit of platforms & publishers.

So the weekly focus and challenge for a brand like Gemini Sound is turning their website traffic into purchases; revenue; loyal customers.

And, of course, doing so in an ROI-positive way: without inefficiently spending their money & time.

Thus, zeroing in on the formula for transforming a website-visit into a paying customer is the real gold mine.

As a fairly technical marketing leader, David has been responsible for defining the logic for their conversion rate optimization (CRO), website personalization, and the types of recommendations they display to shoppers.

“There are a few different areas of our site that have a featured product section. But historically, those types of personalized recommendations are something we’ve controlled manually.”

But there’s a problem here…

And by that, I mean two problems.

(1) It takes too much time.

Defining A/B tests, building segments & cohorts based on attributes, and setting up rules-based optimization funnels are important to get started optimizing. 

But they’re time intensive. 

Even without other responsibilities in the business, a human brain can only manage so much complexity.

(2) Honestly, the results should be better.

Managing personalization via rules-based flows or manual tests and tweaks doesn’t produce the best results.

Provided a store has enough product SKUs, transactions, and customers (in short: enough data), this is one of those scenarios where an intelligent algorithm can do a job better than a human.

But you’ve heard that before.

The paradigm of CRO shifted from split-tests to segment based personalization to IF / THEN rules-based workflows… and we are now in the era of machine learning and predictive personalization. 

Why should we settle for worse performance when there’s a better way?

Prioritizing Conversions

One of the challenges of owning an ecommerce storefront with a small marketing team is managing performance optimization. 

David helps to paint the picture:

“Over the past few years we’ve been tweaking and modifying our site to try and increase conversions. We’re a higher-ticket item, so it’s not a quick purchase for most people. What has helped is Jarvis stepping in and showing people what it expects users want to see.”

Conversion rate optimization is a high priority for the Gemini team. They have been regularly redesigning website pages & elements to streamline the customer journey: from product discovery to purchase flow, and even enabling better technical research (they sell sophisticated audio equipment after all)!

What optimization tactics has the team focused on? 

David says it’s a combination of different things: “Designing site elements in a variety of different ways. Redoing layouts across pages. Changing copy. Working on SEO a ton. Building out additional marketing flows. Working to bring customers back: via email or SMS. Some things work and some don’t work. We’ve just continued trying.”

Introduce Jarvis ML – Predictive Personalization

During his hunt for new tools and technology to improve site performance, the Shopify App Store (wisely) recommended the Jarvis ML app.

But, oh man, unfortunately the setup process was just extremely difficult and super taxing:

  1. Install the App
  2. Explore Dashboard Insights 
  3. Turn On the Predictive Recommendations Widget
  4. Place the Widgets Wherever You Want

That’s it. 

As David says, “It’s straightforward. Plug and play. Zero questions about that.”

When asking David when he started to notice results, he told us it was pretty much immediate. 

We’re talking about hours to days, not weeks. 

“It was so quick and so impressive that it made so much sense to move forward to a full-on plan.”

Deploy!

Launching the experiment was a breeze. “As we redesigned and optimized the various pages on our site, Jarvis really just fit into what we wanted to do. It all just clicked.” David recalls.

After going live, we sat back and let the predictive personalization engine run, learn, improve, and ultimately figure out what makes Gemini customers tick.

David gives the breakdown: 

Jarvis introduced personalized widgets across our Shopify store to show users not what we wanted them to see, but what they wanted to see. Each unique person sees something unique to them, based on Jarvis’s predictions. This in turn led to more clicks, more people staying on the site longer, more conversions, and it affected a variety of other KPIs that were all getting better and better.”

Did Jarvis do a good job? 

“Honestly, my expectations were very low based on the things I’ve done in the past for conversion rate optimization. Jarvis was a lot better than I expected – it was pretty impressive. The data really showed that it works.” David recounts.

When this point came up in our discussion with David, we had a quick side-bar about the strange, almost inexplicable nature of optimization tools nowadays:

It’s surprising to both of us how well these tools can work. There tends to be almost a critical, cynical predisposition towards predictive personalization and optimization technology. And frankly, it’s pretty difficult to explain exactly how it works in any deep detail. I would summarize it as: statistics, data science, well-trained machine learning models, and well-designed data infrastructure. But we laughed and agreed on one point: the proof is in the results.

And as my dad used to say: you can take that to the bank. 

David shared that the core metrics he reviews on a daily basis are pretty much the same as other ecommerce stores:

  • Sales & Revenue
  • Conversion Rate
  • Click-through Rate
  • Average Order Value
  • Return on Ad Spend

Thus it is on these metrics that we focus.

The results of a predictive personalization engine

Jarvis ML’s performance was calculated by running a side-by-side comparison with a control group. 

The relative difference in results was pretty jarring:

  • Click-through Rate: 101% higher with Jarvis ML
  • Visitor Conversion Rate: 73% increase with Jarvis ML
  • Average Order Value: 28% larger with Jarvis ML
  • Revenue Per Visitor: 128% more with Jarvis ML

Just to ensure you caught that, the revenue Gemini generated for each site-visitor more than doubled. The same hard earned and expensive traffic was producing 2.3x the amount of doll hairs.

And all it took was a few clicks, enabling Jarvis.

Predictive personalization via machine learning (our favorite buzzword) automates the CRO process, reduces manual time wasted, and does something magical that we sometimes have trouble explaining: it just creates exceptional results. 

“I would describe Jarvis as a tool. It’s a tool that really works. You’re clearly seeing the click-throughs and conversion rates are higher with Jarvis. It’s not even a question. There’s no reason not to use it.” 

It seems we’ve made a champion out of David and the Gemini Sound team! 

John Rayburn, a product expert on our team, had a suggestion for Gemini that he shared on one of our review calls: “Just don’t spend all that extra cash on some fancy radio!”

 

Interested to hear how we can help you stop losing sales using predictive personalization? Let’s talk!

And if you’re looking for an awesome waterproof speaker, this one on geminisound.com is my favorite.

 

Check Out Our Other Posts

Gemini Sound’s Conversion Rate Improves 73% with Predictive Personalization

Gemini Sound’s Conversion Rate Improves 73% with Predictive Personalization

  “I would describe Jarvis as a tool. It’s a tool that really works. You’re clearly seeing the ...
Spencer Rascoff’s warning to VRMs

Spencer Rascoff’s warning to VRMs

Spencer Rascoff’s warning to VRMs As summer comes to a close and 2023 is right around the corner, the world’s ...
Decked Increases Revenue from Ad-Click Shoppers by 21%

Decked Increases Revenue from Ad-Click Shoppers by 21%

Starting with The Big Picture It’s not a secret that eCommerce merchants are suffering from falling returns on ...