This guest post is written by Marcelo Calbucci, the founder and CTO of Sampa — a personal homepage creator that will be shutting down next month. He’s writing a series of posts about the lessons learned from the venture at http://blog.calbucci.com. He’s also the publisher of Seattle 2.0, a web resource for tech entrepreneurs and startups in Seattle.
Consumer startups are tough. You have two basic choices: A paid offering or a free offering (or freemium). If you charge people a penny, you’ll turn off the bulk of your visitors. If you offer free services, you might grow to be the next YouTube, WordPress or Facebook. Most entrepreneurs are not risk-averse and the dream of being big is just too appealing and the majority of us take the “free-route”.
Once you offer something for free, all shades of people will try to benefit from your service. You’d think a service like Sampa with a strong family and baby branding would just repel small business, teenagers, criminals, etc. but that’s not the case at all. And I suspect most blogging services; photo-sharing or web-site building solutions face the exact same issue we did.
Most entrepreneurs and investors will look at data analysis and talk about averages or totals: Averages number of blog posts per user per week, average number of sign-ins per user per month, viral coefficient, total number of active users, etc. Entrepreneurs who are more sophisticated will split their “averages” and “totals” in two or three groups. For example, fixing one of the dimensions into users that sign-in 30 or more times per month (very engaged), between 10 and 29 times per month (engaged), and between 0-9 times per month (on the brink of leaving) and then run the averages and totals for the different groups (e.g. “very engaged users upload 25 pictures/month, engaged users upload 7 pictures/month, etc.”)
Very few startups actually look at demographic and psychographic data as a way to group their users. Primarily, because it’s hard to get gender, age, income, interests and intentions without asking the user, and once you ask them you might just scare them way or get the wrong information.
One time we went to pitch Sampa to a VC in Seattle, and out of the blue he mentions this other startup growing amazingly fast – had nothing to do with our business. After the meeting I went to check the startup website. Their Compete and Alexa growth was just amazing. Their website contained profiles of all users since it was a public social network. So I clicked on the profile of the 20 people featured on their homepage (“most recent users to join”). Of those, about 75% were girls between the age of 9 and 13 – likely the worst demographic to make any revenue from.
Did the startup know about this? Oh, yeah. Did that VC that was looking at investing on them? Likely not.
In the middle of 2008 we decide to do a qualitative analysis of our user base. People of all kinds were creating sites on Sampa. There wasn’t an automated way to know if it was a baby site, a family site, a small business, a technology blog, etc. We looked at more than 300 sites, randomly selected and created a spreadsheet with the category, the demographic of the author (if we could figure out) and we plugged that into our own analytic system to split our averages and totals for each site category. The results sucked!
Just 20% of our users were on the target audience. That meant 80% were not building any kind of family or baby site. Ok, maybe we can live with that. But it turned out that more than 25% were by pre-teens. There are two problems with that: First, It’s actually illegal in the US and most countries to allow a younger than 13-year-old to sign up to your service without parental consent. Second, pre-teens are not a great audience to build an advertising-based business model.
However the data showed an even worse picture. Pre-teens were a quick burning flame. They would come, upload lots of pictures, write lots of blog posts, “bling” their site, invite 20+ friends and they would be completely gone in a month. That behavior skewed our data enough that once we looked at our growth, viral rates, and everything else, our business didn’t look so great.
Being Proactive Can Backfire
Can you force uses to comply with your Terms-Of-Service and still be successful on a UGC service? Yes, you can. Facebook manage to be very aggressive on the enforcement of their TOS, and so did Flickr. However, if you look at most Web 2.0 startups, they are not doing that at all. The most prominent case is YouTube, which allowed copyright infringement on their website and can plot a $1.6B exit based on their “turn a blind eye” strategy.
We didn’t do that at Sampa, and I’m sure we could have seen 2 or 3 times more growth if we had used the same strategy. We proactively removed pre-teens websites. They weren’t easy to find, but every time we found one, we would remove the website and notify the owner she was 12-years-old. They would be mad at us and tell that “Jamie, Emily and Sally also have a website on Sampa”, and we would say thank you and delete all their friends websites too.
We would also proactively delete porn websites. There is nothing wrong with porn. It’s not illegal or immoral in my view, but it didn’t go well with our family-oriented business proposition. Also, most UGC porn sites are infringing in someone else copyright and we just didn’t want to deal with DMCA or lawyers.
We also found criminal websites, from people trying to steal credit-card and passwords to the ugly side of online pedophilia. We had the FBI come over twice to collect evidence.
And let’s not forget link-farms. Although we had CAPTCHA and email confirmation for new websites, every once in a while someone managed to create dozens of websites in a single day all full of links to some bank, real estate agent, mortgage broker, auto dealer, etc. I’m sure the business that were benefiting from it didn’t know they hired a “black-hat” SEO.
Pretty much every Social Network-builder, website builder or content sharing site deals with the same issues we dealt with. A good number of entrepreneurs (and most investors) will be oblivious to those facts and just think that everything is going great and the growth is sustainable and proof they are creating great value and soon will be able to turn a huge profit or to sell for hundreds of millions of dollars, until someone takes the time to figure out what people are using their service for and finds out it’s really not what they thought it was.
Crunch Network: CrunchBoard because it’s time for you to find a new Job2.0
Facebook’s legal woes just won’t stop coming. Last November a company called Leader Technologies, which makes business communcation tools, filed suit against Facebook alleging that the social network had infringed on a patent that “relates to a method and system for the management and storage of electronic information.” The case is still ongoing, and it sounds like it has some legs — earlier this week Facebook was ordered by a Magistrate Judge from the District of Delaware’s District Court to give Leader Technologies access to its entire source code. As reported at Law360, Facebook has until the end of this week to hand over a hierarchical map of the source, and has until August 21st to share its entire codebase with the company.
Vudu had a great run. The company made boxes you set up near your TV that could download stream HD video without glitches, a sort of