PERCENTIL case study - Miros

PERCENTIL case study

International fashion marketplace rids itself of discounting


Beyond words







Visitor retention


  • 153K unique visitors
  • 50%-50% randomised split test

A unique business model

PERCENTIL is a well-established secondhand clothing marketplace operating in Spain, France and Germany for the past 10 years. Since the company was founded, they’ve been mostly loved by mothers and women in their 30's and 40’s, as well as young adults as of recently.

This loyalty comes from PERCENTIL’s unique approach to buying the used clothes. Clothes are purchased directly from private sellers. The company handles the logistics of picking them up at no cost to the sellers. Once a sale is made PERCENTIL pays a percentage of it to the original seller.

Having such a specific way of attracting sellers they provide premium quality second hand garments to their customers, resulting in more demand, which in turn drives more supply from the sellers.

Challenge: The Paradox of Choice

The company was providing too many options, however great those options were, leaving their customer unintentionally facing the paradox of choice - too many options that makes the buying decision harder. People were getting lost in their buyer journey.

Similar to the rest of the industry, they tried solving the issue by using incentives such as discounts. It only ended up motivating their customers to look for the smallest price tag rather than what excites them the most.

They also tried out different technology vendors, but failed short in accommodating the speed a good search and recommendation engine requires.

Left and right road sign on the road

Sometimes we discount items and they're still not getting sold. Not because the price is too high, but because they're not being found. And with Miros, we are able to help customers find those items.


Lourdes Ferrer


Seamless Integration

After unsuccessfully testing 3 solution providers, they finally came across Miros.

Having a fast and non-stressful integration from Miros, PERCENTIL’s team started getting more confident in the project’s success as the data started showing an uplift in numbers.

PERCENTIL’s team loved Miros’ ability to precisely recognize a shopper's browsing behavior and accurately predict what they truly intend to buy. They implemented Miros to help their shoppers discover fashion items they were actually excited about, each session leading to success in about 60 seconds on average.


Miros relied on their insight that 80% of shoppers never touch the search bar. By implementing Miros’ Wordless Search, PERCENTIL ended up exceptionally improving their buyer’s journey, leading to a significant GMV uplift of +7% after only a single month of full deployment. They also witnessed a +4% uplift for AOV and +7% for customer retention.

The buying journey finally became fun, which made visitors return more often. And the fact that the shoppers were able to find more exciting items, allowed PERCENTIL to increase the amount of items purchased per session.

PERCENTIL was able to place exactly the right products in front of the right customers mirroring their buying intent, which finally let them stop relying on discounts as much, and become able to provide noticeable value again, both for its shoppers as well as sellers.


Wordless Search

What is it?

Just like the name suggests:

Search without using a single word.

Buying visually complex items like fashion, clothes, footwear, furniture, art, design pieces, decor… is a function of style and beauty, not features. So why do we keep making our shoppers buy these like they’re buying a book or a laptop?

Wordless Search is an AI technology that relies on shopper behavior. It recognizes browsing patterns based on which it mirrors the buying intent your shopper has, without them having to input a single word. It gives off the impression that their minds are being read.

Book a demo

Learn about Wordless Search

Give your shoppers the experience they were always willing to pay a premium for. Book a demo to see how.