Client: Dr. Motor, car parts manufacturer and long-term i4B client · Delivery: 4 two-week Scrum sprints · Model: multi-tenant SaaS · Industry: automotive aftermarket · Delivered by i4B
Dr. Motor's existing parts catalogue was neither intuitive nor up to date, and dealers, wholesalers and mechanics struggled to find the right product. In four two-week sprints we built a multi-tenant SaaS web catalogue with fast Elasticsearch search and similar-product suggestions, while managing the product data and its export to TecDoc. The catalogue holds 1,755 products and is open to users who have no TecDoc access.
Project context
Dr. Motor, a manufacturer of car parts and a long-term i4B client, asked us for a better way to present its current range to the people who buy it: dealers looking for parts for end customers, wholesalers and mechanics. The catalogue they used was not intuitive, and its data was incomplete.
Our task had two sides. We handled data analysis on an ongoing basis, managing the product data and exporting it to the TecDoc catalogue, and we built a company-wide car parts catalogue on top of an up-to-date product database. We also took on implementation and post-implementation support.
The challenge
Data quality was critical, because it affects among other things how a supplier is positioned in the TecDoc directory. The catalogue also had to run as SaaS, which meant building on a multi-tenant architecture so that more organisations can be added later.
Four steps to a working catalogue
Analysis of the existing catalogue
The earlier solution had incomplete data and inconvenient search.
A review of the catalogue's structure, strengths and weaknesses, followed by interviews about the client's needs. Together they gave us the requirements and the input for implementation.
Scheduling the work
Implementation and testing had to fit both our team's availability and the client's.
A plan for the implementation and testing stages on both sides, based on team availability.
Implementation in Scrum
Deliver a fast, intuitive search and suggestions of similar products on a stable SaaS platform.
Work in two-week sprints with regular meetings to discuss each functionality as it was built.
Testing and feedback
User feedback showed that people still had trouble searching.
The client tested on a dedicated environment and we turned the feedback into improvements. Search now works by parts group or vehicle category and recognises whether a query is about a vehicle or a product group.
Technologies
Outcome
A freely accessible parts catalogue that reaches buyers without TecDoc access.
We met Dr. Motor's expectations for data validity and search speed. The intuitive search was possible thanks to Elasticsearch and to our domain knowledge of how people actually look for car parts. Presenting the range as an open web catalogue extends Dr. Motor's target group to users who do not have access to the TecDoc catalogue.
Key terms
- TecDoc
- The aftermarket catalogue standard maintained by TecAlliance. Data quality affects how a supplier is positioned in it, which is why we manage Dr. Motor's export.
- Multi-tenancy
- An architecture where one application serves several organisations, each with its own data. It lets the catalogue add more organisations in the future.
- Elasticsearch
- The search engine behind the catalogue, used for fast search by parts group, vehicle category or free text.
Frequently asked questions
How long did it take to build Dr. Motor's catalogue?
Four two-week sprints in Scrum, preceded by an analysis of the old catalogue and a schedule agreed with the client. The team had two developers and four analysts, and the client tested each increment on a dedicated test environment.
Why build a web catalogue if the products are already in TecDoc?
Not every buyer has TecDoc access. A freely accessible catalogue lets dealers, wholesalers and mechanics outside TecDoc find Dr. Motor's parts, which extends the client's target group.
What makes the search faster to use?
It runs on Elasticsearch and searches by parts group or vehicle category, and it recognises whether a query concerns a vehicle or a product group. The behaviour was refined after client testing showed that users struggled with search.
