braddwyer

Tournament username: roboflow.com
Karma: 359
Recent posts
Shared by braddwyer · 275d ago · 0 comments
Shared by braddwyer · 362d ago · 3 comments
Shared by braddwyer · 603d ago · 28 comments
#hackathon
Recent comments
braddwyer · 63d ago

If there's something different you want to be working on in a year, why wait?

FWIW, I tried doing the "4 hour workweek" thing and traveling the world; I got bored. My advice: think about what you'd do if you had a bunch of money/freedom and "retired" then try to figure out how to just do that now.

braddwyer · 64d ago

Why would you want to exit?

braddwyer · 104d ago

Of course, but you’ll have to wait until the memoir comes out a decade from now :p

braddwyer · 107d ago

At my last company (social and mobile games) we innovated on 4 distinct viral loops that each led to over 1 million new users:

1) We were on of the first to use Facebook "profile boxes" (essentially a piece of real estate apps could put directly on users' Facebook profile pages back in 2008) in a novel way. Our core gameplay was to hide Easter eggs there for users to collect; it incentivized them to put the box in a prominent location at the top of their profiles so they could find the eggs more easily. Friends visiting your profile would also see these eggs, wonder "what the heck is this," and click on an egg, which would bring them to an install screen prompting them to join and add a patch to their own Facebook profile as well. Then the process would repeat with their friends (and then the friends of their friends, and so on). This worked great for years, until FB removed profile boxes as part of a major redesign.

2) Incentivized gifting: we would give 1 unit of in-game currency for each egg you sent to a friend (this worked great until every other app also discovered how well it worked and FB banned incentivized invitations due to users getting invite fatigue).

3) This one wasn't really a viral loop so much as an arbitrage play: in the early days of Facebook ads, we discovered we could acquire a user for about $0.15 and the average user would return ~$1 in ad revenue (and later virtual currency spend) in about the first month... and that wasn't even counting the viral externalities of those users begetting other users! This was essentially a money printing machine. At one point we were spending thousands of dollars per day on ads and the only limit on our growth was that we couldn't keep the servers up and running under the incredible load.

4) Aggregated news feed posts. Someone at Facebook posted a breakdown of how the news feed algorithm worked at the time. The "graph rank" was calculated by a number of variables for a piece of content summed over all of your friends who had shared that particular thing. This meant that if many of your friends shared the same "object" it would be very likely to appear at the top of your news feed. We designed an entire game around this insight: the first level was built to be super easy and encourage you to share that you had completed it via "Open Graph" actions. This meant that we could seed it with a cluster of people in a city and very rapidly it would snowball into the highest ranked post in most peoples' news feeds. We picked up 5 million users in a single month with this growth hack (primarily in South America and Eastern Europe where FB games were still novel at the time) and were the #1 free iOS game in Argentina and Uruguay (amongst other countries) that Christmas.

braddwyer · 111d ago

My two favorites are “Zero to One” and “Founders at Work”.

braddwyer · 119d ago

Social gaming during the heyday of Facebook games.

It took a little over two years to cross $1M (primarily selling $10/mo subscriptions and virtual currency) but it was ramen profitable from day 1 (via Adsense ads).

braddwyer · 122d ago

I dropped out of school the first year my company made $1M in revenue. Probably should have been more aggressive in retrospect.

braddwyer · 123d ago

The book "Monetizing Innovation" has a great distillation of how to think about bundling to segment users and maximize revenue.

braddwyer · 128d ago

Biggest "competitors" honestly are people building things internally or deciding not to use computer vision in their projects at all.

The big cloud providers & some other startups have overlapping offerings but most of our customers aren't evaluating us head to head with any of them. No need to "crush" them; oftentimes our tools can work well together with theirs. Eg while we have our own 1-click training tool we're also the easiest way to use Amazon's, Microsoft's, and Google's: https://blog.roboflow.com/automl-vs-rekognition-vs-custom-vision/