precision and recall
categories linked to products
- Efficient and fast onboarding of new vendors.
- Acceleration of product range growth.
- Lack of industry-wide product data standards.
- Product catalogs in different formats, attributes, structures.
- Various levels of data quality and accuracy.
- Different product categorisations used at vendor/ZAGENO.
HOW ONEDOT HELPED
- Increased revenue based on expanded product range.
- Speeded up and simplified vendor onboarding.
- Improved conversion through better product data quality.
- Automated product categorisation, cleaned attributes.
- Improved product search results via better filter attributes.
- Identified product duplicates, removed unnecessary content.
ZAGENO is an online marketplace for life science products, with an innovative solution for procurement departments to streamline ordering. The idea is to eliminate bottlenecks to breakthrough science and research through curated connections linking scientists and their institutions’ procurement offices with leading life-science vendors. With offices in Cambridge, USA, and Berlin, Germany, ZAGENO is currently comprising about five million products and established itself as a unique online biotech marketplace.
The online biotech marketplace ZAGENO exceeded its planned growth trajectory for the year 2018, mainly because of the company’s customer-centric culture. Dedicated ZAGENO teams of experts in science, E-Commerce, systems integration, and customer support, work directly with customers. This enables smarter, faster processes which allow more time and resources for value-added science and better research results.
ZAGENO’s database encompasses 1’800 brands and 2+ million stock keeping units, making ZAGENO the #1 online marketplace for life science and biotechnology products. High-quality product data and attributes are crucial for researchers to find the right product matching their experimental needs.
ZAGENO build the first comprehensive E-Commerce platform for biotechnology products featuring a cross-manufacturer product comparison. However, industry-wide standard product catalogs are non-existent. Therefore, incoming vendor datasets and the present ZAGENO database have different formats, attributes, structures and also various levels of data quality and accuracy making it difficult to integrate.
“Onedot supports ZAGENO in our rapid market expansion by automating the vendor onboarding and making products more searchable through better product data quality. This ultimately leads to a faster growing global customer base and higher customer satisfaction.”
David Pumberger, CEO & Co-Founder at ZAGENO Inc.
ZAGENO uploads the raw product catalogs of new vendors to Onedot’s cloud. In a predefined interval, Onedot then processes the uploaded product catalogs as a batch and presents intermediate processing results to ZAGENO for validation and feedback. One batch contains up to 600`000 SKUs.
The feedback from ZAGENO is collected and then used to retrain the Onedot’s machine learning models. By repeating this process, the performance (precision and recall) of the processing increases over time and the feedback effort required by ZAGENO is continuously reduced.
Each processing includes the following results
- matched incoming products to ZAGENO’s marketplace categories.
- mapped vendor attributes to ZAGENO’s target structure.
- extracted and normalised key attributes like sample size and units according to ZAGENO’s data model.
- created a Golden Record on product, variation or price level.
“Onedot provided an artificial intelligence (AI)-powered data preparation within a short period. This fast service automatically integrates, transforms and enriches data from various sources of structured and semi-structured data, even exceeding human precision.”
Christian Böhm, VP Scientific Content at ZAGENO Inc.
Life science researchers will significantly benefit from the increase in data quality and categorisation accuracy. Locating products faster and more precisely will enable ZAGENO to take the next steps in customer centricity while delivering the best possible search results. Automatically adding and categorising up to 600’000 SKUs per batch, speeds up the onboarding process and the go-live of the new big vendor, drastically.
With a precision and recall of over 95%, the incoming vendor products are successfully categorised into over 140 categories. Plus, only a few iterations and minimal training data are required to train the categorisation classifier. Golden records are identified using fuzzy matching across predefined product attributes. Onedot shortens the time-to-market, improves conversion rate and reduces the operational costs.
Being an ambitious high-growth company with a focus on customer satisfaction, ZAGENO plans to extend their demanding biotechnology product range to 5+ million products within the next year. Hundreds of new vendors need to be automatically onboarded, millions of incoming products need categorisation to match ZAGENO’s growing biotechnology category portfolio, and database management must achieve a new level of quality assurance. All of these challenges will be manoeuvred by applying Onedot’s data preparation service.