Ask HN: Here is my current batch of ideas. In exchange, I ask for your feedback.

With regards to #6: The bank I used to work at used Gomez and we were (well they probably still are) paying out the ying yang for a handful of tests each hour against a few core webapps.

So I had this idea to create a distributed monitoring service that would use nodes (similar to the genome@home project) with actual residential internet connections to give websites a true idea of how fast their site was from different locations around the world. Node owners would just be regular people, and they would get a small percentage of the revenue from each test they ran, and the infrastructure would be designed such that a failing or slow test result would initiate verification tests from other nodes to insure that it wasn't just an issue due to the node's internet connection slowing down or failing. Gomez runs different test from their own servers around the country and sometimes we could tell what the issue was based on which nodes were down (i.e. the west coast server is the only one that's failing). So we'd have all these nodes running all over the world and we could tell people "well here's how quickly this page and all it's resources load from New York City, LA, and Russia..." and give them historical and real time metrics and notifications based on those response times.

I got as far as creating a client that worked relatively well and was ready to upload results for each test ran to a master server, plus it made requests to the master server to see what the next test it needed to run was. With all the other availability testing services out there I kind of moved on, but I've still got the source if someone is interested.

Yeah, I've been a Gomez user before and the pricing was utterly shocking, to say the least. Additionally, I think our pricing plan only even allowed tests from 10 nodes or so. I was wondering if I could write a php endpoint that acts as a testing node and just buy cheap hosted php boxes all over the world on as many backbones as I'd need to cover. Obviously testing through real users is better, but 9 times out of 10 the interesting stuff in Gomez is a function of problems in your backend, not ISP problems... and anyway, what are you going to do about those? We got all kinds of false positives when the Gomez nodes would glitch out. Almost no signal and all noise.

2 - yeah, I think there are multiple challenges - #1 sales cycles and custom requirements #2 how to bootstrap testers #3 the infamous problem of "no one googles your solution when they don't know the concept exists" -- do companies think "oh i should go look for a company that provides beta testers?" don't think so. maybe it just means that paid search won't work. maybe it just needs a lot of CPM marketing on sites for developers and good old word of mouth.

4 - not sure, never tried vox.

5 - no one seems to be directly attacking the vertical very well. i think "families" is a vertical, right? :) some things you can repurpose to get what you want, but the average person would never think to use it that way. the ones that are specifically built for families -- well they're either just awful, or seem like some kind of re-skinned groupware.

6 - you can do that in google analytics. it's a hack, but you can definitely post up metrics from the server side, if you just build up the urls that it needs. but anyway, this is definitely something where i was just thinking more about availability and performance monitoring.

I like your feedback, since I am a pessimist and can see the truth in all your criticism :)

1 - totally agreed it's a Twitter feature waiting to be vacuumed up. but it's SO easy to implement. and I know how to make it go way viral. i think? :)

2 - why does anyone beta test anything? i think people just like feeling special, having special access, being early adopters, being influencers, etc. plus, i imagine the site as totally having badges/rewards built into it by design. and version.future - framework to allow companies to reward their testers (name in credits, direct payments, tip jar, whatever)

4 - agreed. maybe i could use #3 to solve this problem :)

5 - i know there's a market for this one, just from being a parent and knowing lots of parents. i don't think facebook is acceptable since there are weird issues with friending your parents etc, and grandmom is probably not even on the site. blog-wise, there are verrry few blogging platforms that have the right security controls.

6 - agreed. it would be easy enough to try something simple: 10 urls, 10 sampling nodes, 10 tests an hour, $10 a month - and just see what kind of interest is out there.

There is something I really like about #4 the guided daily diary. I wonder if you could turn this idea into a dating site -- users can submit a guide and the top ranked ones get answered the next day. Users of the site can browse everyone's answers for that day.