Combining AI and Design within the world’s largest Online Marketplace

Every month, OLX helps 322 million people worldwide to buy, sell and exchange products and services through their 20+ brands. To lead product innovation and keep finding new ways to meet customers' needs, they needed fast innovation, product development, and experimentation. This is where Koos came in, founding OLX² together with OLX's in-house talent. The result is a collaborative innovation lab that combines the power of data science and design to develop future-proof, human-centred solutions with AI at its core. By helping to foster an innovation mindset in the world’s largest online marketplace, OLX² created three AI-by-Design concepts aimed at different customer segments.

Client

OLX Group

Challenge

Helping future-proof the classifieds business through innovation, product development, and experimentation by exploring AI solutions to improve customer experience and create new business opportunities.

Impact

Three disruptive AI-by Design concepts aimed at different customer segments while inspiring an innovation mindset within the organisation.

An innovation lab creating AI-by-Design solutions

OLX Group is a large-scale network of trading platforms operating worldwide in over 30 countries with over twenty brands. As a complex system with a rooted legacy in the classifieds industry, shifting towards experimentation and radical innovation can be challenging next to running the day-to-day business on team and organisational levels. The collaboration between OLX and Koos has been facilitating this new direction. How? By setting up OLX² – an innovation-dedicated lab, founded to boost product and business innovation within the organisation. Our focus is on designing, testing, and validating user-centred business ideas, which can fundamentally change the way people use classifieds within a two-year horizon.

The unique point-of-view of the team does not stop here. We are an international and multidisciplinary group: from design thinking, data science to business. With this diverse expertise, they aim to create the most advanced yet acceptable applications (following the MAYA principle – ‘Design for the Future, but Balance it with Your Users’ Present’) of AI that radically solves real customer problems and spike utterly new business opportunities, to future-proof classifieds.

A multidisciplinary team at its core

Since January 2020, Koos has provided two full-time consultants to OLX²: a UX designer and a Service Designer. In addition, on the OLX side, there are another five members: a project lead, a data scientist, a researcher, an engineering manager, and a frontend developer. They report directly to the Chief Data Officer of OLX and are advised by the CEO of OLX Group on a quarterly basis.

The name OLX² comes from Google X, Google's innovation lab. Double X makes a squared for exponentially accelerating change.

AI by Design

This collaborative approach was named AI-by-design. It took on a combination of disciplines that can make for the perfect love story: Data Science and Service Design reinforce each other and potentially lead to original user-centred solutions.

 

But the value of OLX² goes further than such individual developments. Since the team aims to go through the process autonomously while amplifying innovation on a cultural level from the inside out, it can accelerate growth continuously. In other words, our ultimate goal is to create an in-house self-sustaining team in which Koos has, at some point, potentially no role anymore.

However, the reality is that existing product teams don’t often have the resource capacity to give life to entirely new services. The team’s strategy is to have a supporting role by offering in-house development capacity that ensures the solutions can be applied to all departments. As an add-on to other teams, they share radical innovation capabilities and, consequently, accelerate growth without affecting their roadmap, KPIs, and OKR’s. Moreover, the way OLX² is established and operating in the organisation allows working cross-siloed and cross-markets while delivering long-term impact and success. For example, we are currently validating and adapting an innovation designed for the Indian market to replicate it successfully in the European market.

Mission-based working

So far, OLX² has worked on three different AI-by-design solutions in three other domains in 2 years:

  • AI-based video selling classifieds to improve trust and safety
  • Radically personalised product recommendations
  • High-trust online self-inspection for cars

These involved more than 12.000 end-users through interviews, surveys and big-scale experiments.

Currently, our focus mission is on the latter: the automotive industry, developing a global self-inspection tool for second-hand cars that will significantly disrupt the way the condition of vehicles is determined.

Our latest mission is focused on increasing trust in second-hand cars.

OLX² missions not only drive innovation for the topic at hand, but they also have the potential to amplify innovation for the whole organisation by showcasing and reinforcing user-centric problem-solving.

— Steven OLX² Project Lead, OLX Group

Inspiring innovation within the largest Marketplace company worldwide

The innovation culture within OLX Group is scattered. The very nature of the company, harnessing many brands across the globe in various industries, is behind this diffused character. Brands like Otomoto (Polish car website) and Avito (Russian classifieds) mainly operate like separate businesses, focusing on their own goals. This results in a work practice primarily focused on incremental innovation since that directly contributes to the KPI’s of that specific business unit.

While this type of innovation, which generates better solutions to enhance existing products, was already nurtured, exploring and developing new technological directions was still slow. Along with the emerging of a new wave of tech native competitors, OLX saw this as an opportunity for improvement. To keep thriving as a leader in product innovation and maintain their position as one of the world’s fastest-growing networks on the trading market, they wanted to better facilitate radical innovations with AI at its core.

Aiming for non-incremental opportunities required OLX² to have a new skill set, a diverse team and added responsibilities. The initiative started as an experiment to understand whether we could help scale the innovation and develop solutions that can change the future of classifieds. In parallel, we also share experiences and best practices regarding experimentation across the Group.

OLX² is the OLX speed boat exploring the waters of innovation. Scouting for opportunities helps the vessel find the right direction to lead the online marketplace scene.

Merging Service Design with Data Science

The team uses the service design Double Diamond approach as a foundation to identify key customer problems. Along the way, the approach is iterated, tweaked, and extended into a new way of working. It consists of six steps:

The OLX² approach highlights:

1. Data scientists and designers work together in the process. We see that data science comes as an afterthought in several companies when the solution is already found. This means that they do not design the answers for AI, which can cause a lot of challenges. At Squared, data scientists are involved from the start to exploit the potential of AI more efficiently, thereby creating better future-oriented innovations. This way, designers are educated and steered along the way on realistic AI solutions that are human-centred and data-driven, which leads to leveraging the power of Machine Learning and AI.

 

2. We don’t stop after testing: we want to deliver. We want the solutions to be implemented and used by real customers, which requires development capacity. Not only that, but we learned that it is hard for other teams to implement our solutions and free up their (already complete) backlogs if we just have a pitch deck and the results of some prototype tests to convince them. So, we wanted to take our developments one step further and do everything we could to ensure a smooth implementation. Now, we have in-house development capacity, allowing us to build our solutions. This progress provides the resources to create working demos, MVP’s, and experiments on the platform, which validate our solutions from the business perspective instead of just the desirability and to set up effective collaborations with leaders and teams across the group.

One of our most successful examples of a large-scale experiment is the Fake Door experiment. We set up a banner on the most significant car selling platforms, offering a free report download. Although it was a fake gate, we ended up having half a million people clicking on it and getting a survey pop-up instead!

What is to come?

Innovation in a big corporation is not an easy process. We continuously refine our way of thinking, approach, reassessing needs, and discussing whom we should collaborate with to improve and shape future innovation.

OLX² wants to set an example of AI-by-design innovation and inspire other teams in the organisation to learn from their journey. They host monthly events to share their learnings and mistakes, design training programs around experimentation and AI and build their team website to make knowledge sharing more accessible. On this website, other employees can access all the research docs, AI demos, and prototypes for their own benefit.

The team’s current focus is on scaling one solution towards different markets.

We are in the process of running sprints to research, design, and develop an MVP for this solution.

Want to know more?

  • Based on our experiences designing, developing, and managing AI, we have summarised the key insights in this white paper.
  • In collaboration with us, our colleagues from OLX have been telling the AI-by-design story in their Prosus AI blog series.

The brainpower behind this project

Ask them more about this case!

Get in touch

Erik van Zoelen

Our UX'er gets energised out of developing a product until the last pixel and loves to design digital products that allow him to be amazed by people.
Related case

Or just keep reading

Back to all projects

Disrupting the Polish jobs market: designing a new strategy for OLX using qualitative and quantitative data

Gaining insight into the needs and dreams of users to shape the future of the job market.