The story of Cardmunch
Entrepreneur turned Investor, Founder of K9 Ventures, Team & Starting Up Expert
CardMunch had the right idea but the wrong pitch to investors.
CardMunch digitized business cards by applying parallel processing to a common issue.
A huge network can be more powerful than revenue.
Lesson: Founder Framework with Manu Kumar
Step #4 Parallel Process: The story of Cardmunch
CardMunch, I am going to give you a little more background on that, so Bowei and Sid came to me and essentially talked about they wanted to do something in the context space; talk more about relationship management type stuff. My comment to them was look, my problem is more on the input side. I go to an event, I end up with all these business cards, and I have no idea what to do with them. In fact, my entire desk on one side of my desk is nothing but a sea of business cards.
Unfortunately, the problem still has not been solved. The way the idea came around was I had the sea of business cards on the side of my desk. I am like, “Okay, I need to somehow digitize these, so I can then make them searchable.” I can hire an intern and given the number of cards I had sitting there, I figured it would probably take three months for somebody to sit there and actually type in all the cards. How the heck do I get a thousand people to be doing this in parallel? It's like how do you apply parallel processing to the problem?
That's how the idea of CardMunch came around where we take a picture of the card, and then you have all these people all over the world who can actually be typing in the information that's on the card. In fact, let me go a little deeper on that example because it will actually tie in to the revenue model side of it as well.
In CardMunch, we were having real humans transcribe the business cards. We had an actual cost associated with getting each business card transcribed. In fact, it was costing us almost up ten cents in some cases to transcribe a business card because we wouldn't have just one person do it, we would have typically two and sometimes up to five people to transcribe a card.
The reason is that humans are just as bad as machines and they make as many errors as machines do. The problem is that the machine is almost deterministic and it comes up with almost the same answer each time. We wanted to have multiple humans transcribe the same card, and then we would correlate to see how many of them actually matched and that's how we would come up with an accurate answer.
We had an added actual hard cost associated with it, and we figured that if we were to make CardMunch free, we would essentially be losing ten cents on every card and we would be bankrupt really, really quickly. What we did when we launched, we actually launched as a paid app where people would be paying to transcribe the cards and effectively it worked out to somewhere between 20 to 25 cents is what people would be paying to transcribe a card. That in turn had an impact on the distribution because now we have essentially put a price barrier into the product that limits the number of people who are willing to use that product. Once LinkedIn acquired the company, they made it free and the usage started increasing dramatically because now it is free and everybody wanted to use it. It by far gave the best results for a while.
The lesson that I learned from that was in that situation, we didn't have a lot of choices about what our revenue model and what our business model could be because if we were to make it for free, we would bankrupt the company, so the only way around that would be to either go out and raise a lot of money, and then be able to subsidize making it free based on having raised venture money.
We actually tried that approach but we tried it with the wrong pitch. Our pitch was around, “Oh, we are going to charge people for transcribing cards and they are going to get lots of distribution and this is how much money we will make.” That turned out to be the wrong pitch.
Our pitch should have been we are going to make this thing free, we are going to eat all the costs of transcribing the cards, we want everybody who is in a professional or business setting to be using this, and we are going to acquire a network that can grow faster than what LinkedIn can grow because we were essentially getting physical connections where people were meeting and exchanging business cards.
In hindsight, I can say that that's what our pitch should have been but we learned that by iterating over it. There's lots of interesting lessons learned through CardMunch as well.