How MVMT Watches Optimized Their Cross-Selling Strategy

Mitigating the risk of introducing a new product with testing

When Jake Kassan co-founded MVMT Watches in 2013, he knew right away that optimization would be key to the company’s success. Why? As Kassan put it in a recent interview with Optimizely, testing gives him the ability to evolve and innovate faster than his competitors, and it ensures that every decision he makes (and dollar he invests) has a positive impact on revenue.

“Rather than trial-and-error, which is expensive and very difficult to confirm whether something is really working, we prefer a data-driven approach that allows us to test our hypotheses in a very objective way,” Kassan says. “Whether those hypotheses are right or wrong, we’re able to generate a clear answer, make more educated decisions, and pull the levers that really move the needle on revenue.”

Retention and is a major focus for MVMT because of the nature of watches: customers are only in the market for a watch once every few years. “How do we keep customers returning to our site and shopping throughout the year?” MVMT asked.

Working with BVAccel on the strategy and Optimizely for A/B testing, MVMT successfully introduced the new product feature of interchangeable straps. Adding this new cross-sell item to the site via a series of A/B tests increased conversions by 5.5% for mobile shoppers and 2.2% for desktop shoppers.

To achieve these results, BVAccel used Optimizely to test many hypotheses and decide how to introduce this product in a way that mitigated the risk of disrupting the current buying process.

Choosing where to promote a new cross-sell product

This idea did not come from a quick team brainstorm, BVAccel and MVMT did deep analysis to determine the optimal part of the buying process to promote the cross-sell.

They considered three different scenarios for promoting the straps:

  1. Promote in the shopping cart
  2. Promote right after purchase on the confirmation page
  3. Promote on the product page

Weighing the pros and cons of each, they thought that promoting on the product page most closely reflected the in-person buying experience for a watch and therefore would be the best but feared the straps might be a distraction. MVMT and BVAccel agreed to move forward in a data-driven way testing the straps on the product page to understand the impact on revenue and conversions.


Testing played a crucial role in releasing this product, mitigating the risk that offering straps would hurt important conversion metrics.

"Using Optimizely’s technology and the guidance of BVACCEL, our entire organization has become more data driven about every decision we make,” Kassan said in an interview. “From onsite design to product development, everything we do has influence based on data."

The team approached design for the desktop and mobile pages separately, testing each context uniquely.

Desktop Experience

On desktop, they hypothesized placing the straps next to the watch would fit naturally into the buying process and increase revenue per visitor without significantly decreasing total completed sales. “We felt that, just as if the guest was in a store, they should be able to view all the special features and options for each watch they were considering for a purchase on the product page,” Mariel Bacci, Senior Strategist at BVAccel said.

They tested placement of the strap promotion with three variations:

Test Variation A

Variation A: No cross-sell option

Test Variation B

Variation B: Strap Cross-Sell Under Product Styles

Test Variation C

Variation C: Strap Cross-Sell Under CTA “Add To Cart”

Variation B with the strap upsell below the area of the page that gives the guest a choice of product style produced a 2.2% increase in revenue and an increase in average order value.

They hypothesize that this version performed better on desktop because it is a more natural shopping experience to choose a watch first and then browse strap options.

While the conversion rate, or the number of completed sales, did decline slightly with this test, approximately .4%, the increase in average order value and therefore revenue per visitor more than makes up for the slight decrease in sales. The significant revenue increase can be attributed to an increase in order value from the addition of straps to each purchase.

Mobile Experience

Given that 55% of MVMT’s site traffic comes from mobile, BVAccel dug deep into the data to make sure the test was uniquely optimized for mobile and not simply a transfer of the winning desktop experience.

They looked at mobile scroll maps to understand what percentage of shoppers would scroll deep enough down the page to see the straps if they were displayed beneath the “product styles” section as they were on the desktop site. They found that only 8% of guests would have ever made it to the interchangeable strap section. Armed with this data, they designed an experience that would perform better for mobile visitors.

Mobile A/B Test Variations

The control and two test variations on the original design

Placing the strap cross-sell directly beneath the “add to cart” button performed with a 5.5% increase in revenue and increase in conversions on mobile for this variation of about 1.6%.

The strap upsell directly below the product image is performing better on mobile for revenue because the strap upsell is directly below the product presentation. The version with strap upsell at the bottom is not as effective because most people on mobile will never make it far enough down the page to interact with the straps.


The team gained a ton of insight from this cross-sell test. Here’s a summary of the top lessons learned:

  • Imperfect tests are better than no tests - This test proves, that test variations only need to represent a solid concept to be effective. BVAccel and MVMT implemented this test without properly merchandising the strap options first; Gold watches may have featured silver strap hardware. While this was not ideal, merchandising the straps perfectly would have been a substantial development effort impeding the speed to deploy and ability to learn. If MVMT were not able to find a winning variation, the additional development time to “perfect” the test variations would have been wasted.
  • Know your audience - Knowing that MVMT watches has a more mobile than desktop traffic, they tailored this test to reflect the needs of mobile guests, which generated big gains.
  • Mimic “in store” experience to create integrated buying experience - Doing this helped them to integrate the strap cross-sell option seamlessly into the buying process without making it look “tacked-on” like you might find with something like a “you may also like” section on the product page or shopping cart.
  • Test introducing new or cross-sell products to prove impact - It is valuable to test merchandising on newly introduced products to the site, cross-sell or otherwise. This test proves that with a 5.5% increase in conversion from product pages on mobile and a 2.2% increase in revenue on desktop.


Increase in Sales for Mobile Shoppers


Increase in Sales for Desktop Shoppers

$20 Million

Annual Revenue

Thanks to Optimizely’s technology and the guidance of BVACCEL we’ve been able to make major design changes to the site with confidence.
Jake Kassan

Co-Founder of MVMT Watches

Get Started with Optimizely

It's your turn. How will you experiment everywhere?