Vidmad - Rival to Youtube's Search Algorithm [2014 - 2016]

When we found out that Youtube's search algorithm and results are biased and inappropriate. We thought of doing something about it on our own. We developed an automated solution to fetch all youtube videos on a daily basis and rank them on the basis of merit score. We kept updating the merit score algorithm to improve the rankings.

We initially just wanted to improve search rankings but soon realized there is a potential for other important areas e.g. to even predict which specific video will go viral tomorrow basis the trends. So we started providing multiple solutions through our automated system. We scanned metadata of over 40 crore videos and were hitting google API daily limits in less than 8 hours consistently.

But we started facing roadblocks as we expanded e.g. We were unable to get country wise details from google to improve suggestions or there was no ability to even pay google to increase our API Request limits. Apart from this google marked us as 'Pure Spam' which meant we were erased from search results entirely. Overall such issues led to the decline of this idea. But it was a great initiative in the sense that I learnt a lot including things like managing and processing huge data by optimizing queries, learning new database languages etc. We were offered an intial funding offer valuing the idea at more than 2.5 crores but at that time we had rejected the same for various reasons.

Share-Blogs - Network of Blogs for everyone [2008-2010]

When we realized that each individual has different opinion on different issues and it is an innate desire to express the same in some form which can be through direct communication, social media channels or even blogs.

Blogs has always been a different channel since authors could write in great detail on a specific topic compared to other mediums where only summary is preferred. We observed that the overall blog lifecycle started with a person trying to write a blog on some topic. Then once the effort was put in they would try to publicize it on social media or other networks to catch attention. Since it is the 1st blog they wrote, people would prefer to read it and some may compliment the user. The user getting somewhat success would write another article and the blog would slowly get lesser viewers. Now after 2-3 articles, considering the value-effort is in negative for the writer the author would stop blogging. This was a common cycle but we wanted to stop the same.

The existing network of blogger was good for just writing the blog but it did not solve the problem of getting viewers of similar interest to a blog. To solve this problem, we started working on a blog platform where users would write blogs or share existing blogs. These blogs would in turn come as recommendations to other blog users bases on previous interests and exisitng blog topic. So our journey began and we crossed more than 30 own articles. The amazing part of the same was the learning of SEO. We realized that to improve visibility we needed an earlier search rankings push and for the same we optimized every bit of our code including performing various speed tests. We also learned that google rewards patience and does not show immediate results. After few months we got our breakthrough when one of our Housefull 2 review article started getting rank 1 in search results for multiple keywords. Those who do SEO related work would understand how tough is this to achieve without putting your money in Ads.

All was going well until the time an inappropriate article was posted on the network which led the project to a standstill. Months after when we decided to resume the project the incentive started to reduce as we found there were multiple sites starting to provide similar solutions eg. Medium by Twitter. But the learnings of this startup helped me a lot in future work and will always be useful in times to come