The Departure
On the 23rd of March 2012, I left my job as a division Technical Director at a defense company with the goal of building products and being my own boss. I had made the proper arrangements by saving up a good amount of money (7 months of bills), having an idea of the path forward, and actually tendering my resignation. My goal, after my final day at the office, was to take one week off and relax, and then get back to work. I never guessed how hard this was actually going to be.
Even though I told myself I was taking a week’s vacation before getting back to work (for myself), I still worked. I couldn’t help it. I was excited for the opportunity, and I thought the idea was fantastic… from an engineer’s standpoint.
The product was named “The Happens” and the idea was to programmatically log everything a business thought was necessary as an event. The problem that I was addressing was simply stated, but hardly realized: “When asking a business owner what is the current state of their business, responses will vary, but some examples might include “growing”, “shrinking”, “revenue has plateaued over the last quarter”, or “we just hired our 100th employee”. Follow-up their reply by asking “Why?” and most answers would appear to derive from guesses or belief as opposed to fact.”
Some people, like my father, understood the problem immediately. He should given that he was giving me a kitchen table business education ever since I could do arithmetic. But others, well let’s just say other’s really didn’t quite get it. They didn’t know, what they didn’t know, to paraphrase. They didn’t even get it enough to ask me how I was going to help businesses move forward with their perceived problem. To me it was quite simple. Provide and API to log an arbitrary JSON string to a series of events, user goes to site, and the site provides them with charts and graphs of those event series. The key, I was assuming, was the ability to combined graphs from different series to see if there was possible connections. All in all, I felt that with all of the blank faces and confusion, “The Happens” was not going to be something that started making money before my savings ran out.
The Pivots
I was deeply irritated and depressed for about a day, but it’s amazing what you can get yourself out of when you need to (as I work on borrowed time). I need to pivot, but I wasn’t quite sure where to pivot to. When working on “The Happens” I had decided that, being how abstract the concept was, I would need to blog about ways in which I was using the product. One of those ways was the relationship between the sentiment of tweets and the sign-ups for my product. I knew I had to pivot towards a smaller niche, so I decided that doing sentiment analysis could be a pretty interesting niche. And an even nicer thought came into view: not a single one of 117 products doing sentiment analysis took the companies tweets into consideration. This amazed me because it only made sense that if a company acted negatively, the crowd would respond negatively.
So I started thinking about a MVP and decided that the user would sign up for my product, provide me with one or more twitter usernames, and I would provide them with a weekly report that looked like the following:
The line graph depicts the relationship between the sentiment of the @username tweets, and the crowds mentions, replies, retweets, etc. The X axis was to depict the days of the week, and the Y axis would be split in half, separating positive from negative sentiment.
The two donut graphs depicted the sentiment rations for the @username and twitter crowd. All three graphs were created with RaphaelJS.
Next I moved onto the interface to display the weekly reports. I was exceptionally happy with how this page turned out being how lack confidence in all things HTML and CSS related. I owe much to Twitter Bootstrap for making my life easier.
The big idea here was that I wanted to help companies recognize specific moments in time where their output (press releases, blog posts, tweets, etc) would affect the very people that were listening to them.
The Journey
I briefed what I was calling “Weekly Crowd” to a local company that helps connect entrepreneurs with resources and an interesting mention of location came up in conversation. The mention of location caused my mind to spiral towards thoughts of epidemiology, and word of mouth, and Malcom Gladwell’s “Connectors”. I can only speculate that they didn’t find much value in the two screen shots (and, in turn, the product) but were fascinated with the idea of sentiment analysis in social media. And to be completely honest (in retrospect) that although the graphs were providing information that was helpful, it was also providing information that was pretty obvious. This decreased its value proposition.
So I changed the question, this time the pivot was a much quicker and faster. The question became even more focused on something that is without a question valuable in a world of high signal/noise ratio in social media: “Which individuals will direct marketing have the greatest network effect?”
This week I took my first steps to realizing the solution to the above problem, so I thought that I would conclude it with this blog post. Future posts will chronicle the development of my solution now called Influmetrics, just as this post paraphrased the past. I hope to make these posts a weekly occurrence — stay tuned.