There are a few data-related considerations specific to a marketplace:
- Unlike a SaaS model, Ankorstore does not have recurring revenue or guaranteed repeat customers, so there is no churn or lifetime value to calculate.
- The concept of "cart" differs from the perspective of the brand (one "order" from one retailer) and the retailer (a "checkout" that may include products from multiple brands).
- Ankorstore differentiates the retailers between signed-up and first buy, as their level of commitment to the platform can vary significantly.
The Ankorstore data team organization
Unlike many companies, where data analysts are expected to be full stack and handle everything from data engineering to proper analysis, Ankorstore believes in specialization. This strategy allows them to grow faster and makes it easier to recruit specialists in specific areas.
As a result, the data department at Ankorstore is split into three distinct teams:
- data and analytics engineering (collect, clean, model data)
- product data (work with the CPO, A/B test, search, rank)
- business data (work closely with business users to help them with their data-related requests on a daily basis).
This structure allows analysts in the business data team to have a strong business sense and answer business questions, while the data engineering team handles the technical aspects, producing dashboards for example.
Gautier Chenard, Lead Data Analyst of the business data team, manages 10 analysts, including Tristan Comte, Senior Data Analyst. During a webinar, they shared with us how they use Husprey and why they find it beneficial.
How to use Husprey for an analysis
Ankorstore demonstrated how they use Husprey for an analysis of their loyalty program. This analysis was to be shared with the marketing team so that they could decide whether to keep the program or to cut it off.
First, the data engineering team prepares the data in the warehouse, ensuring that it is clean, harmonized, and up-to-date. This allows analysts like Tristan to easily access the data and make simple SQL queries.
"Then I went into Husprey, I created a data set called "raw data" which aggregates all the dimensions I will use in my notebook" — Tristan Comte, Senior Data Analyst
Tristan uses Husprey as an aggregator of data that will be used in the notebook. He uses basic SQL langage to publish each charts making it easy to explore the data and uncover insights.
Husprey or Jupyter notebook?
Gautier identifies 3 main reasons to use Husprey notebook instead of Jupyter notebooks.
"The main asset is the visualization. Creating bar charts on a Jupyter notebook is painful. On Husprey it is super easy, with just a few clics you can transform the query results into a clear and engaging visual representation of the data." — Gautier Chenard, Lead Data Analyst
The second benefit is the ability to connect to databases and explore the data directly in Husprey. "On the left side, I can go into charts, explore the fields connected to our dbt or to our data warehouse. It makes the work easier"
Last but not least, the ability to easily share the analysis with business users is considered as one of the main advantages of Husprey notebooks. Unlike Jupyter notebooks, which can be complex and difficult for non-data people to understand, Husprey has a clean and intuitive design that allows anyone to comment and interact with the analysis.
"We work with non technical team, we need them to be able to read and understand easily the analysis we share through the notebook. Husprey has a nice UX, which allows people to easily comment, share links, it helps us a lot" — Gautier Chenard, Lead Data Analyst
How did Husprey facilitate data analyst's work vs another tool?
Sure, you can use Notion or Google slide to present an analysis, but Husprey allows you to create SQL queries directly in the tool instead of just taking screenshots. "If I have to work on the project V2, I can refresh this analysis and reuse the queries. I don't need to go back to my sandbox or to compute queries from A to Z. I only have to copy-past the queries I need. All with a single tool", Gautier said.
"This is the true difference. We have a single tool to tell a story, write queries, go into our database, all this with a UX that allows "non-data-people" to comment and interact" — Gautier Chenard, Lead Data Analyst
To sum up, Gautier considers Husprey closer from Notion than from a Jupyter notebook, with the SQL dimension in addition.
