Automotive aftermarket,Repair and Maintenance Services
How we helped Swiss startup tame big data and change car maintenance experience
When it comes to the automotive industry not everything is about brand-new cars and vehicles. Secondary automotive market for replacement parts, accessories, repair and maintenance services might not be as glamorous as car industry, but it is expected to reach USD 486.36 billion globally by 2025, according to 2018report by Grand View Research
In 2015 Swiss startup, Carhelper, set out to transform reservation process for car repairs and services. They wanted to create a more transparent and faster system which will help users save time and money, by totally reinventing car service booking procedure.
To accomplish that goal, Carhelper was looking to collect and aggregate data from multiple sources such. A major challenge was to connect and consolidate all repair shops data and prices, especially since repair shops aren't exactly eager adopters of new information technologies.
They aimed for a platform that would process complex calculations in the background, in the shortest time possible. By automating all reservation tasks, it was meant to show up-to-date availability and immediate price quotations as well as handle real-time communication between customer and repair shop. On the other hand, optimal user experience meant minimizing the amount of user inputs.
As a full technical support team (CTO company), we approached all those issues as one big technical challenge, and then looked into segmenting it.
In order to instantly deliver precise prices, numerous calculations and rapid data aggregation had to be performed in the shortest time possible. We had to fetch data from various sources, analyze and fit them into a usable structure: car plate and location information as well as repair shop prices and schedules. Anticipation of all possible behaviors including missing information or output errors was required, as data providers offer inconsistent and unreliable service.
Our chief task was to create the final product which can handle more than 200+ API request per user query with acceptable time under 2 sec.
Furthermore, user experience was critical - tool had to be user-friendly and easy-to-use, both for car owners and repair shop workers.
Our complex and innovative solution is composed of several different segments:
To set a solid foundation, we opted for cloud-based service, served by Cloudfare and Exoscale. This simplifies monitoring, enables automatic adjustments of our capacities and eases initialisation of new instances.
Automated testing provides production ready code, flexibility for daily A/B testing which in turn allows for frequent changes and informed decisions on the business side. Subsequent Carhelper's growth and scaling were greatly facilitated by this initial choice.
In order to acquire and retain users clean and modern product design was imperative. The goal of user experience design process was to distill complex user journey into simple and intuitive one. User behaviour research, removing complexities, and iterative real user testing helped us create product that empowers users to get the job done quicker. User provides minimum input data (e.g. only car plate number) and is able to instantly book their car service.
Data manipulation / core system:
Data aggregation and processing is at the core of our system. In order to provide user with accurate and relevant information (garages available, correct prices, distances and user reviews) we had to gather and process data from several different sources: multiple APIs for car manufacturers, models, spare parts, than services and work time databases from official manufacturers, Google Places API, Swiss car plate database and others.Since platform’s functionality depends on so many data providers, to bypass the problems occurring due to downtime and unavailability of data vendors we established methods for auto-healing missing data and set up alternative user journeys.
“An example of rewarding solution we came up with was Distance calculation. Initially, to get accurate distances between user and repair shops, we used Google Places API. As the number of requests to server raised over a million, paid Google service was no longer financially justifiable. So, we created ML system which learns from Google API, user inputs and our own data, so that now, we don’t require Google API anymore to calculate precise distances. This solution alone brought about great savings for Carhelper.”
CTO of Abstract
Agile approach and development using Tech Stack:
Technologies we chose for this project proved to be stable and reliable but flexible and dynamic enough to allow for daily iterations, testing and scaling:
- PHP (Laravel)
- NodeJS (Express.js)
- MySQL master-master replication mode - Redis master slave replication
- MongoDB (logs storage)
Providing instant and precise price quotations was well known problem in car service industry. We were not the only one who tried to solve it, but we were the first to succeed.
Today, Carhepler is the only instant price service on the Swiss market with 500+ garages as partners.
It is used not only by thousands of car owners to book maintenance, but it was quickly adopted by partners and repair shop owners when calculating their own prices.
Furthermore, Carhelper’s scalability and flexibility made expansion to other markets, including Italy and Germany quick and effective.
“We have been working with Abstract since 2017 and are extremely pleased with this partnership. Zoran Dobrosavljevic and his team work with the highest quality standards and surprise us with their fast and solution-oriented working method over and over again.
Thanks to Abstract's support, we were able to further develop our business case Carhelper.ch from a simple MVP to a sophisticated e-commerce platform. We are very grateful for the excellent and reliable cooperation.”
CEO of Carhelper AG
Want to team up with Abstract?