Unmana wrote about how we decided to get started with Markitty: here is my account of the product development so far.
I had done a bit of coding more than 10 years back, but since then my exposure to the technical side was limited to managing software projects, data analysis, and occasional experiments with SQL or Excel macros. Building a SaaS product of our own was a task I had never imagined taking up.
Outsource or build in-house
We didn’t have a lot of money to hire a great outsourcing vendor or to hire a senior technical lead. We did try looking for someone who could be a technical co-founder but that was not going to be easy.
Unmana and I had both worked in geographically distributed teams for a long time and understood the communication overheads and leakages that need to be dealt with in such a setup. So we were very clear from the beginning that we wanted to have a local team and dedicated developers who can work closely with us. Apart from the cost, culture was the biggest driver behind this decision.
We wanted the kind of environment where we could walk over to a co-worker and ask a question, or quickly get into a meeting together, or sit next to each other to explain that the table needs a little more cell spacing! And more than anything else, celebrate small victories as a team.
Deciding on a technology stack
Markitty needed to be a lightweight web-based application that a user can start using within a few minutes, and that will integrate with other marketing channels like Facebook and Google Analytics (for websites).
Based on my personal experiences and reading what other product start-ups were using, we had narrowed down to two options — Python and Ruby. PHP was my last option, and only because there are more PHP developers in Pune and finding one might be easier.
Finding the hacker
Unfortunately, colleges and institutes in India (Pune, specifically) don’t teach Python or RoR as part of their regular curriculum. So there are only a few developers who know these technologies well.
After posting on various job forums, talking to recruiters, attending local developer events, and advertising on LinkedIn, I had gone through 300+ resumes and hadn’t liked a single candidate. Unmana and I were already working full-time on defining product requirements and doing research (apart from consulting) and we hadn’t found a developer yet — time was running out.
In50Hrs had announced its first hackathon in Pune and I thought that could be a good way to find some enthusiastic developers. I looked at the Facebook Group for the event to figure out how many Python or Ruby developers were going to the event. I only saw a few developers there, but there was this one guy posting daily updates about some new app he built each day.
I asked him on Twitter if he would be interested in doing a project with us. Within the next two days, we met him and worked out the details. This guy, of course, is Virendra.
When we met, I was surprised he was so young and was still a student. I never looked at his resume or asked him to do an aptitude test: his work spoke for itself.
Plans are made to be changed
We had documented product requirements very early on but a lot of things had changed by the time we got to the implementation. Also, Virendra hadn’t played much with data and he would be working along with his studies. So we cut down the requirements to the bare minimum and defined what was required for a working prototype — something we could use to demonstrate the concept to potential customers and investors. We also wanted to use this to see how well we would work as a team.
We thought we’d find another developer (maybe someone with more experience) in the meantime and have her/him build the actual product — which never really happened. During the implementation we learned that some of our assumptions about Facebook and Twitter APIs were wrong, and some things had just changed since we’d done our research. I was hoping to learn coding while working with Virendra, but my old Windows programming mind couldn’t grasp much of web programming.
We had planned to complete the prototype in one month and it took us a little over two months. But finally it was done and we were more than happy with the output. Initial reactions from friends were positive (obviously, since we have nice friends) and we had found our technical co-founder in Virendra.
Project management: walking the fine line
I have been a business analyst, consultant, and a project manager, but the role I needed to play here was no less challenging than any I had done earlier. When the product manager is also your life partner and you only have one developer to do all the work — it’s not easy to plan or adjust priorities.
Initially I was afraid it will be difficult to decide when take whose side if there was a conflict about including or excluding certain features. But as it turned out, Unmana and Virendra were usually on the same side and I had to argue with both of them, refusing to include more features or functionality because it would add too much complexity or take a lot of time.
I am proud of our small team and what we have achieved so far. Our dream is to make marketing easy for small businesses, and I get the feeling we are on the right track!
Markitty is an online marketing tool for small businesses that puts together stats from Facebook, Twitter, and Google Analytics, and suggests actions based on what you are doing.