ManoMano, the leading European player in online retail within the DIY, home and garden sector, has chosen Snowflake to design a user-oriented data platform with a data self-serve strategy. The result? A solution tailored to the needs of the company, and to users who are both autonomous and empowered in their data use.
ManoMano: A tale of massive growth
ManoMano has grown extraordinarily quickly since its inception in 2013, raising funds of €725 million. A vertical marketplace specializing in DIY, home and garden, ManoMano brings together over 4000 sellers and hosts a vast 16 million products. The company employs roughly 1000 employees, who work out of offices in Paris, Bordeaux and Barcelona, tirelessly attending to the needs of 7 million active customers across Europe. The ManoMano platform—which receives over 50 million unique visitors per month—expanded into the B2B market in 2019, with the launch of ManoManoPro.
Given its position as an online-only platform, ManoMano is more than just « data-driven » — data lies at its very core, inextricable from the marketplace’s activity and success. Every ManoMano employee utilizes data by necessity, whatever their role. This is particularly true when it comes to scraping, which entails automated collection of data about competitors, and monitoring user behavior. The objective: to understand users’ browsing habits, remain competitive in the market, expand the product catalog, identify new leads and consolidate ManoMano’s position as the retailer of choice among tradespeople.
When internal data (sales volumes, promotions, and so on) are taken into account, the volumes of data being stored and processed are enormous — around 2 million select queries are processed every week on average at ManoMano. In turn, every employee is empowered, within the scope of their individual role, to develop new data, forging a path toward improving the company’s performance. Overall, no fewer than 200 KPIs—including around 30 « golden KPIs »—are reported to management on a daily basis. The final elements of the ManoMano data strategy are ML and data science, which encompass aspects such as categorization, automated extraction of product specifications, product recommendations and cross-selling.
Data, a driving force for business performance
In light of such massive growth, data management has steadily become more complex, to the point of introducing tangible risks. « We had two problems to address — a deterioration in the data quality and performance of our existing data warehouse; and increasing technical debt because there were just too many different tools being used. How the data was organized was becoming more and more misaligned with our business needs, which were growing exponentially, » recalls Cédric Cormont, Senior Lead Data Architect at ManoMano.
So, in 2020, ManoMano turned to Snowflake to deploy a data platform that was suited to both its then-current needs and, more importantly, to any potential future needs, in line with its significant growth objectives. According to Cormont, the project was part of a four-step process: « the first port of call was to improve technical performance, and then to quickly establish a data organization strategy commensurate with the growth in data volumes. After the third step, which involved rationalizing and stabilizing the platform, the project needed to be finalized in a way that allowed data to be organized within the new data warehouse, so as to enhance data quality. »
The migration from the existing data warehouse to the Snowflake platform took six months, with both being run in parallel during the last month. In the first phase, ManoMano chose to copy over the legacy data, with no specific optimization or remodeling work. This solution was sufficient, initially, to solve our performance issues, » said Cormont. The deployment of the Snowflake solution was, most importantly, an essential step in ManoMano’s complete overhaul of its usage-oriented data governance and strategy.
Before the deployment, ManoMano had been reconsidering its organizational approach. Up to that point, the organizational structure had been empirical and centralized, with multi-functional teams, and there was a risk of being overwhelmed by technical issues. Now, however, the structure is specialized, with technical teams on one hand and business teams on the other. « Such organizational structures just weren’t possible before the Snowflake deployment, » said Cormont.
Self-service data with Snowflake
Leveraging the findings of the Snowflake 360 audit, which was performed by Snowflake teams at ManoMano over the course of a week of observations and analysis, ManoMano has adopted a FinOps-style approach with a very clear objective: to take control of cloud costs by empowering users—especially the data scientists themselves—through dashboards and usage alerts.
Data governance issues can then be addressed. « With Snowflake, we planned from the outset to involve employees so they can utilize data subjects, » explained Cormont. « In a company, the purpose of data is not to please the data teams, but rather to serve the business itself, which must be able to make use of it in a self-service manner. » The Snowflake platform therefore includes a data catalog that all users can access, with the aim of enabling employees to discover and understand key aspects of the legacy data, such as: who’s responsible for this? What’s the SLA? How should incidents be handled, and by whom? And so on.
It didn’t take long for this to bear fruit — BI usage, of course, is burgeoning, as are data discovery, data visualization and dashboarding (with SnowSite). Projects involving ML and data sharing are growing too. To drive the project even further, ManoMano has turned to Snowflake’s AirFlow orchestrator to develop event-driven architecture. « With this approach, Snowflake has not only helped us to break our data monolith, but also, and most importantly, to design microservices capable of publishing business events, which eliminates the risk of breaking pipelines when data schemas are modified, including with SaaS tools, » said Cormont.
To round off its self-serve data strategy without running the risk of the data platform becoming too complex, ManoMano has implemented Snowflake’s design layers, enabling it to create business tables, document them and transfer them to the catalog for re-use later on. Meanwhile, Snowflake’s data academy supports and certifies users in their use of self-service data.
Now fully stabilized and widely used by all employees, Snowflake has unlocked ManoMano’s potential to take its data strategy further and harness the data in its ecosystem.