The Dev World – Sergio Lema

Some years ago, I worked on a big project to manage car repairs from a website.

The idea was fine. The client had a lot of money to invest.

But the project lasted 2 years with absolutely no profits.

Why didn’t it work with so much investment? With a so big team?

Let me give more context.

The Vision

I won’t name the client for privacy concerns. Let’s say the client is named GearGlide Autos.

It was an existing international company with already a lot of business with garages and professional vehicles.

They wanted to create a website to estimate the cost of repairing the customer’s vehicles.

That’s a great idea because when I go to the garage, I never know if changing the tyres will cost me 200€, or if they will see another problem while changing the wheels and finally it’s 5.000 €.

The workflow was:

• the customer needs to repair his car;
• in the website, it describes the problem (tyres, glass, motor or other);
• the website gives an estimate of the price;
• and it gives 3 garages near the customer to book the repair.

The concept was fine. And GearGlide Autos has already a big garage network.

So, let’s put a lot of money into it. Think big.

The Request for a Big Team

GearGlide Autos came to us asking to build the website. They asked to build big.

Ok.

We started with a big team of 20 people: developers, product owners, QAs and more.

We started developing the website as requested. Think big.

GearGlide Autos had a lot of features in mind for the website.

They were very dynamic. Proposing new features, and new concepts for every Agile Sprint.

But all those ideas were before the application is publicly released.

The Big Project

We build a microservices architecture. Starting with 6 microservices, but reaching 15 microservices as the project had more features.

Java backend, React frontend and deployed on AWS.

We’ve added all the side-project tools for a good-quality project. Sonar, ElasticSearch, Nexus and more.

It quickly reached tens of thousands of lines of code. A lot of design patterns to handle many discount methods and repair workflows.

15 developers working on it. PR merged every day with a complex CI/CD workflow to deploy the individual microservice in an adequate environment. All this maintaining a high availability in AWS.

The Reality

10 users per day on the website.

Launched in 3 countries. With thousands of garage partners.

About one assignment per day.

But the benefit was only if the customer paid. Some garages asked the customer to cancel the assignment. This way, the price will be lower as he pays directly to the garage.

Most of the time, the customer asked for revisions as something doesn’t work correctly on the vehicle. You can’t estimate a price for that. So, all the negotiation is done directly with the garage. The website is useless here.

The End

GearGlide Autos saw no benefits with the application. They started to build an internal team of developers (to be cheaper).

But they saw all the tools to migrate: Sonar, Nexus, Jira, Confluence, AWS, ElasticSearch.

They saw all the knowledge needed to maintain this ecosystem.

Even internally, the project was not profitable.

6 months later, I asked one of the internal developers how were things going.

They cancelled the project. Everybody was reassigned to other projects.

The website was not even available on the internet.

Conclusion

GearGlide Autos had a big vision. They had all the garage network necessary.

But they were specialized in B2B (Business-To-Business) models. Here, you go to big business and sell your idea. With only one client, it’s a lot of money.

When going to B2C (Business-To-Client) models, you need to build trust with the customer. You need to create an image. You need to be known to a lot of people. You need a lot of clients to make money.

In my opinion, they spent too much on a strong website instead of marketing.

First, test your project (a small project) with a small audience. Try to grow the audience. If it works, grow your application.