Waseem Jan Syed Waseem Jan

What I learned from my failed startup

September 15, 2018 Waseem Jan 9 min read # chefgalaxy

In 2014 I was hired for a small contract on Upwork for building a module related to profiles in python flask. I completed the task, my work was liked and I was given another small contract. After finishing multiple small contracts and getting more and more involved in what was being developed, I was offered an opportunity to become a partner and co-founder. I instantly accepted the offer.

It was a startup related to the culinary industry named Chef Galaxy. The basic idea of Chef Galaxy was to connect customers with chefs for their events like birthdays, weddings, etc, an Upwork for chefs.

Chef Galaxy landing page

After spending 4 years on it, we finally shut down the server on 23rd August 2018. It FAILED, but I still learned a lot from it. Here is what I learned from it.

Validate

The first and most important thing is to validate your idea. In our case, as Chef Galaxy was not originally my idea, I am not really sure if my partner had actually validated it. I joined it when MVP was already being developed and at that time I didn’t know much about idea validation, so I just kept my focus on development. Beware that MVP is not needed for idea validation, it can be done even with a simple landing page offering a service that is fully manual behind the scenes. Validating the idea gives you the idea about how big is a market for your product or whether the customer exists at all. Additionally, startups are not about how cool your product is, it is about whether the product is actually solving a problem customers are facing. We should not be building our product for ourselves, but for the customers who are ultimately going to pay for our product. Validating the idea saves a lot of time and money.

Speed

This is one of the most important reasons for Chef Galaxy’s failure. I started building Chef Galaxy’s MVP in early 2014 and finally launched it on production in late 2016, after almost 3 years of development. An MVP should never ever take this long. It was started keeping in mind MVP but during the course of its development, we slipped from our trajectory and kept adding features, convincing ourselves that a fully completed product is much better than a not so shiny broken product. The thing is, even after launching in 2016 we felt like it was not ready. So we should always jump in before we are ready, fail early, learn from it and iterate over it. Small failures are easily affordable compared to one huge failure where the energy to take more risks is already drained.

Where exactly did we waste so much time?

  1. Adding more and more features like admin panels, dispute resolutions, and real-time notifications. None of this was necessary for our core functionality.
  2. Spent way too much money and time on the look and feel of the product. All icons and graphics were recreated several times and the layout of the pages was redesigned multiple times.
  3. We also wanted to be a Facebook and Quora for Chefs so a whole new social network around culinary services was developed which should have been a product of its own. Moreover, that is not an innovation, and the chances for success are tiny for clones. I felt like we were doing too many things and actually tried to resist but maybe I didn’t resist enough. I think my resistance was dismissed because I was an engineer and not a business person.

Money

Save as much money as possible and spend money only on development. This is something I feel we followed as we didn’t have any office, bills, or expensive subscriptions but I feel there was a lot of room for improvement. Like

  1. If we didn’t waste time on the look and feel of the website we would have saved a lot of money we were paying our graphics designer from Ukraine and Layout designer from China. Additionally, we also had a CSS guy from Pakistan so every time there was a redesign of a page, we paid our CSS guy too for a rework. All of that was a waste.
  2. I feel the amount of money we spent on servers was also too much. We had two environments, Test and Production. The Test was a simple EC2 instance using an RDS instance with everything installed on the EC2 like Elasticsearch, Celery, and Redis but Production was designed to be highly scalable and reliable from day one where Auto Scaling Group and Load Balancer were configured along with services like VPC, S3, Elasticache, and OpenSearch. All of this was charging us around $180 a month and I think it was unnecessary. We could have done all of that on a single EC2 instance and then used another tiny one just for testing.
  3. One more thing that was not needed was registering the company. One can always do it later when the product is profitable. Similarly, if you are too tight on a budget then even creating a private Github repo is a waste, one can always use other alternatives like Bitbucket or Gitlab.

In the end, it’s always easier to save money than it is to make money.

Sell

After losing speed during MVP development, the second most important reason for Chef Galaxy’s failure was not being able to sell the product. We always thought we just have to build and launch it and customers would start signing up. That’s not how it works, it didn’t matter how cool our product was, only building something is not enough. In fact building a product for engineers like me is most of the time the easiest part, selling it is the difficult part. During the development, we always had discussions on sales, marketing, and partnerships and I always thought there were some guys (Founder’s wife and wife’s friend whom I never met as we all were working remotely and my partner was managing them) working on it. It was only once we launched the product I realized we didn’t have any solid plan. I tried to convince my partner on spending some energy on SEO but he never agreed. Relying upon a product to go viral amongst customers if it’s B2C is very difficult.

One thing that my partner realized later and actually wanted to work on but never got a chance because he got some family issues was content creation for educating our target market on our product. I think good content would definitely have helped us in our SEO too.

Partner

I was chosen as a partner by our founder. Now I am not a graduate of a top university in the US. I was also not from our founder’s friend circle or someone he knew before. But I was someone who shared the same vision and had the same passion for the product as the founder. I was also someone who had skills (software engineer) that the founder didn’t have (Finance Manager) and I think these are the properties one should look for in a partner and co-founder.

I think we did a pretty good job in finding good partners for Chef Galaxy. Moreover, during the course of development, we let go of 1 person, a database administrator from the UK.

DB administrator was let go because we simply didn’t need him. We were already using AWS RDS in addition to other AWS services like VPC, security groups, and network firewalls which in combination took care of a lot of things like backups, performance, security, and integrity of the data. Moreover, these things were not an issue for us because we simply didn’t have much data and load on our database. I am a full-stack engineer and I was already taking care of the database through Flask Sqlalchemy models, automated management commands to alter the data, and using Alembic for data migrations.

Other things to take into consideration for building a team are trust, honesty, integrity, and likeability because when money starts coming in, these attributes can save you a lot of stress. Chef Galaxy never reached the scale where money could create issues.

Ready for failure

When the product was completed and launched, I moved to another city and got a full-time job there as I felt I can do the rest of the work for Chef Galaxy quite easily in my part-time unless, of course, it becomes viral in that case I could always have resigned from my fulltime job. Now I feel like moving to another city and taking a full-time job long before I shut down the server for Chef Galaxy was really a good decision. I think if I was still in the same place working on Chef Galaxy it would have been really hard for me to let it go and it would definitely have taken an emotional and financial toll on me.

In the end, I would say despite getting failed, Chef Galaxy has given me so much. Even today in my full-time job I sometimes apply the technical lessons I learned from Chef Galaxy. Something even more important from the technical stuff is the friendship I made with our founder, it exists to this day. During the course of our journey, we used to have long calls and those calls were never always about work. Our founder was in UK and I was in Pakistan so we had a lot of topics to discuss from cars, homes, weddings to current affairs. We also used to discuss our family issues and how to handle them. I would never trade those memories for anything.