Episode 2: March 28th 6am to April 4th
On March 28th at 4:20pm Emily Melton one of my favorite VCs (from Draper Fisher Jurvetson) emails me to say:
>>Munjal,
>>I have been following the recent press around Riya and I wanted to send a quick note of >>congratulations. It is often hard for a company to live up to pre-launch hype, but it appears you >>have exceeded the expectations. Great work to you and your team!
>>Hope you and your family are well.
>>Emily
Simon Tocker from the UK wrote me on March 29th to say:
>>This is the best thing I've seen on the web for ages, I've held back from using things like flickr as I >>didn't like the layout for others to view my photos, your system is in beta and its miles better.
>>I was expecting results for face recognition at 20% but you far exceeded my expectations its >>amazing.
>>....
>>Thanks Simon
Most of the conversations from people who tried Riya said something similar. I blogged about many of theses posts here .
Ed Baig from USAtoday wrote a very nice print article of the product after actually trying it on his photos (however this had little impact to our number in daily registrations - more to the point of why online is so much effective than offline).
Internationally I got this request:
>>Guten Tag, hello
>>I'm see your product in www.computerworld.
>>We are a fotofinisher-laboratory in Switzerland
>>So, we are interesting for your product: automaticaly face recognition and name-setting.
>>Please send me your documents and pricelist.
>>Best regards
We got an honorable mention in the Web2.0 awards here
bg @ microsoft.com signed up to try Riya. Rumor had it this is Bill Gates real address.... not sure if it is a fake or not. I suspect it is.
We got 834 emails to our support address in the previous two weeks. I read everyone. Many of the emails started out with "I'm really sorry to have to report this
We did a survey of the first users and they said the following: 1073 responded to the survey
a) 73% were satisfied or very satisfied with the accuracy of our face recognition, 23.2% were somewhat satisfied, 3.8% were unsatisfied.
b) 71% were satisfied or very satisfied with the service overall, 25.8% were somewhat satisfied, and 3.1% where unsatisfied.
c) 90.7% of people would recommend Riya to a friend.
I did most of this analysis and survey myself. Why? Well no one else volunteered..;-) and I wanted to hear first hand what people thought of our service. I wanted to read the intonation in their emails and the tone of their criticisim not just get a sanitized report from someone else.
I really credit Burak and his entire team for this achievement (of most people being satisfied). These guys are the geniuses behind Riya's algorithms. It was a huge huge huge accomplishment that people thought it worked at all and even more so that it was good enough. Face recognition technology is very brittle and frankly just doesn't work most of the time. Burak and his team had made this happen. Drago, Vincent, Lorenzo, Kuang-chih, Diem, Baris, Danny, and Burak deserve all of the credit. Just amazing work. I cannot stress this enough.
All of this sounds great right? Well inside the company we were getting this sinking feeling... Why if everyone loved us so much were the number of photos dropping? Between week 1 and week 3 number of registered users per week dropped 70% and the number of photos uploaded would dropped 65%. On top of that our sharing rates where very low.
Was this just the Techcrunch effect?
- There was definitely evidence for this in our referral graphs
Was this truly that people were blowing sunshine up our butt (e.g. I love your stuff in words but when
they voted with their mice they were not coming back?
- There was no evidence for this but no evidence that this wasn't true. Tara is well loved in the Web2.0 community and was that giving us
Was it that Tara had done such a good job with blogs that we had signed up almost everyone who read Techcrunch already?
- There was evidence for this. Our registration numbers were already coming close to the 50K people who read Techcrunch.
Was it that Riya had to be a client application for more people to use it?
- We had certainly heard this from a number of people
Was it that Riya needed more social and community features?
- Thomas Hawk and other web users had written just this or told us directly.
In truth we just had to wait and watch. Only time would tell us which of these hypothesis were correct or not. Everyday at midnight Ginto's daily metrics report email came out ... and everyday it showed the numbers dropping and dropping. This is what I meant by the cocaine effect of Techcrunch. It is possible that if the numbers could have started small and grew slowly week by week it would have been better to emotionally handle than seeing the numbers just drop and drop and drop because of the huge spike at the beginning.
We all starting to conjecture about what we needed to do next... if you have ever been in a situation like this you know it gets very chaotic. Everyone is just trying to make the numbers go up and no one has much real information... it is just a free for all. This exploration will consume the next weeks...
I have to get home to Deven, he is quite sick today so I will pick up on this more tomorrow.
Awesome story! Great reading for all of us would-be entrepreneurs. Keep it coming Munjal.
Posted by: Ali | May 30, 2006 at 11:29 PM
Well apparently you have been hit by the crowd of early adopters, would-be web entrepreneurs and techies that were around to 'review' your app. Not the kind of users that came to stay.
Did you consider waiting for the downhill to bottom? That would be your real initial users. Find who they are and build from there. (I guess that's probably what you did)
Posted by: Harry | May 31, 2006 at 05:49 AM
Munjal,
salam from Bombay.
This is a great startup insight blog. Thanks for sharing this with us. As an entreupreuner I am excited to read this. I look forward to the rest of the series.
Posted by: RYK | May 31, 2006 at 11:29 AM