4 ms·
5 years. (2011-2016) While working on my first startup (Sauce Labs), I made a robot that plays Angry Birds as a side/art project in the autumn of 2011. The ser
by hugs 10y ago
5 years. (2011-2016) While working on my first startup (Sauce Labs), I made a robot that plays Angry Birds as a side/art project in the autumn of 2011. The serious part of the project is that you can use the robot for testing mobile apps and devices. I started to sell the robots (still as a side project) on Tindie when that first launched in the summer of 2012. I made a couple of thousand dollars the first few years, but basically broke even profit-wise. I got my first "big" enterprise order of robots in late 2014 and that made me start wondering if I should ramp things up. I formally incorporated (http://www.tapster.io http://www.tapster.io) in May 2015 and took some seed funding (Indie.vc) to sell the robots full-time. Since incorporating, it's now veeeery close (this summer) to ramen profitable while paying a few salaries.
- brianpgordon 10y agoIt's interesting to me that your approach to automated testing includes a real, physical device. How well do you think your robot offering is positioned versus fast and free virtual automated testing within an Android/iOS simulator? I assume there must be some material advantage to testing on a real device, but I don't know a lot about the domain. What happens if the layout of the app changes slightly- are your testing programs purely based on physical coordinates?
- hugs 10y agoExample #1: How do you automate a test for what happens to your streaming music app when a phone call comes in? (This was a real scenario a client couldn't figure out how to automate any other way.) Example #2: You're a car company and have a mobile app that controls parking your car. You have a business (and possibly legal) requirement to test the app with the same unmodified hardware your users will use. (So you can't only use simulators.) Example #3: You make wearables that measures steps taken. How do you test the wearable to make sure it's logging steps properly? (You shake it with a robot.) Every company that makes a device (phone, car, wearable, anything with a touch screen, etc) has a secret robot testing lab. I didn't know all of these examples were real business problems (with the associated big enterprise budgets) when I started the project. But after enough of these conversations, I realized there's something there.
- benliong78 10y agoAt the Previous Company I was in (Bindo Labs) I was coding the NFC Payment Integration with the Point of Sales System (POS). One of the requirements for the integration was that every transaction has to happen less than 0.5 seconds including the hardware card-reading part. So I ended up having to wave the card around the reader all day testing it. I thought of building a automated testing robot but tight deadline prevented me from diverging, and it didn't end up happening. I'd be curious to know how big a business it is for you now that it's 5 years past. Sounds like a very interesting case-study.
- hugs 10y agoWe're now at about $150k in bookings halfway through the year. Last year was $30k revenue for the entire year. I hope the numbers keep going up or this will have to go back to side-project status!
- coverband 10y agoIn that case, my I humbly suggest turning the tapster.io landing page to a more informative experience, rather than expecting visitors to click over to the Tindie project site?
- pjc50 10y agoI've been involved in automated testing with a robot for point-of-sale systems. The pinpads are secure devices, so there's no way of injecting commands, and there's no adequate bug-compatible simulator. It's cute watching it carefully insert a card, type in the PIN, remove the card, etc.
- tmat 10y agoi just tried using sauce labs real device testing and I must say its pretty unusable for any real testing. Browserstack has this running pretty amazingly these days as I found out after saucelabs failed for me :(
- usloth_wandows 10y agoThat's hilarious! I totally expected some sort of automated testing suite for emulators/simulators, not a physical robot with a stylus! Awesome work!
- King-Aaron 10y agoThis is extremely cool, I need one of these little guys on my desk already!
- mavelikara 10y agoLong time Selenium and Sauce Labs user here. Thanks for all the good work Jason!
- michaelbuckbee 10y agoThis isn't meant to be snarky, but I seriously suspect that your business will really take off when the bot + ad click farming communities clue into what they could do with such a system.
- asimuvPR 10y agoI remember when it started selling in tindie. It was really interesting but I wrongly thought that no one would use that. Im glad I was wrong. It definitely opens yo my eyes for opportunities. Best of luck!
- scosman 10y agoThis is the first time I wished a company had a giant "jumbotron" video on their marketing page. You sell robots, add a video!
- hugs 10y agoYeah, the website needs some love. I didn't even have a separate website for the project until this year.
- takatin 10y agoI made this: https://www.behance.net/gallery/23269525/IOJS-logo-concept https://www.behance.net/gallery/23269525/IOJS-logo-concept I can build you a website. Email in profile.
- imron 10y agoYour website needs a video of the robot in action.
- crobertsbmw 10y agoI had this same idea about a year ago and started working on it with my cousin. I don't think he was as excited about the idea as I was, and we never got past the LEGO prototype. I'm glad you have done this. It convinces me that it's at least a <i>decent</i> idea.