|Introducing myself and why my startup is doing it wrong||Shaun Thambyah||7/17/14 8:24 PM|
By way of introduction, I’m Shaun Thambyah, I’ve been most things tech through my career. Mostly in corporate Australia, I spent a few years running the online operations for some big brand name insurers. While there, it became pretty apparent that something was missing in our monitoring capability. Nothing could tell us if our sites just looked right, so the idea for obvious.io was born.
We’ve spent a lot of time building a product. Everyone told us this was wrong, but we didn’t listen. I guess we're doing it wrong in lean startup terms. The truth is, we had to prove the technology behind the scenes would work, because sometimes having a crazy idea doesn’t mean you can make it work, and making something work doesn’t mean people will pay you for it.
So here we are, that crazy idea does work, we’ve sorted that out. But can it go from being a ’neat idea’ to a product?
We built Obvious.io as visual monitoring tool for websites, imagine if something like Pingdom could actually “see" your page, just as your users see it and alert you the moment something doesn’t look right. Obvious.io takes regular snapshots of your page and compares them with a baseline, we use some nifty computer vision techniques to identify what parts of your page have changed, you can choose ignore the dynamic bits so you won’t get woken at 3am because someone published a blog post or a news item changed.
We believe monitoring your user experience to make sure your customers can use, transact and navigate your website is simplest and highest value way of maintaining your brand’s reputation and ensuring downtime does’t cost you valuable conversions.
Hopefully we're not doing it completely wrong. We’d love your feedback.
|Re: [SiliconBeach] Introducing myself and why my startup is doing it wrong||ryancross||7/17/14 10:00 PM|
Interesting product. As someone who works on the web alot, I can definitely see the value. I’d be happy to meetup to give you some feedback if you want.
I’d specifically be interested in integrating development testing into the system. (like behat / selenium / etc) and possible ways of integrating it with the development workflow. Also, is it capable of testing visually across different devices and interaction modes?
|Re: [SiliconBeach] Introducing myself and why my startup is doing it wrong||simran||7/18/14 4:42 AM|
I absolutely love the service - i'm seeing it as a much more featured version of http://www.changedetection.com/ - which i absolutely love and have used for a long time as well.
In the spirit of "you build it for dogs, and then the cats come and take over" - i see immense use for it not as a testing service but as an alerting service for pages you want to generally keep an eye on.
As an example - the page http://airbrushventuri.com.au/course-timetable is fairly static, but changes occasionally when the timetable changes. I want to know when that happens to see if any times have been added that i can join in on. changedetection.com has numerous times alerted me and i have been able to check the new time.
What that's saved me:
* having to check the http://airbrushventuri.com.au/course-timetable page daily/weekly
* having to write a script to monitor that page and alert me
Obvious.io looks like i get much finer control - which is brilliant... i'd pay for that service :) in fact, i went looking for pricing... saw you are making a bundle on sms charges ;) but couldn't find pricing on regular checks otherwise (even though it says i have nine hundred and something free checks left).
My main feedback would be, please make it possible to check pages daily or even weekly... i don't particularly need some things checked hourly (which is your current max) - as that would end up costing me (when i'm a paying customer).
Huge opportunity here in "change alert" generally, and not just in testing i think...
Fantastic work :)
On Fri, Jul 18, 2014 at 1:24 PM, Shaun Thambyah <sh...@obvious.io> wrote:
|Re: [SiliconBeach] Introducing myself and why my startup is doing it wrong||Shaun Thambyah||7/18/14 5:04 AM|
Thanks very much! Sounds great, I've pinged you an email to organise a catch up.
The testing aspect has been on the radar for a little while, plugging obvious.io into a CI pipeline could add a nice visual check during a release or even as an acceptance test. The API is in progress, with some functionality already existing to allow you to visually compare two monitors (eg, dev/test, or blue/green etc). Ideally that would be pluggable into any jenkins, bamboo, TC, etc pipeline. I'll keep the group updated on when the API is released in beta.
The rendering aspect across multiple devices is an interesting one. There is the ability behind the scenes to change things like user agent and screen resolution, which would effectively fake that scenario. Very interested to know what you see the use case as being for that?
|Re: [SiliconBeach] Introducing myself and why my startup is doing it wrong||Shaun Thambyah||7/18/14 5:29 AM|
Thanks for the valuable and detailed feedback, very much appreciated.
I've never seen changedetection.com before but it looks like a very interesting service. I can see why people are hooked on it, it seems like a useful tool to hook into something like IFTTT.
Pricing is definitely something we have under test at the moment (you get 1000 free checks and 20 free SMS alerts on sign up). We're experimenting with an Amazon style pay-for-use model, billed monthly, currently we charge $0.002 per check, 1 check per hour equates to around 720 checks per month which would cost $1.44. The pricing is on our landing page, but I'm going to take that as feedback that its just not visible enough - we'll fix that up. I 'borrowed' Pingdom's pricing structure for the SMS alerts, I suspect their business model is built around consumption of SMS's rather than the checks themselves. I'd be keen for feedback on what a reasonable price for both would be. This is actually one of the hardest things I'm finding, setting a reasonable price point (one which everyone wins) so any thoughts you have would be most appreciated.
I take your point about the maximum check frequency, i'll do some experimentation with the UI to see how we can do this nicely. I also need to work out exactly what the maximum check frequency would be in order to be profitable. This leads a little back in to the price point question, would it be more appealing to purchase a 'pack' of checks, similar to the SMS packs and auto top up when the packs expire?
I like the "Change alert" concept, it seems that theres a "low frequency" group of users that would find this appealing. Thanks again for all the feedback.
|Re: Introducing myself and why my startup is doing it wrong||Matto Rochford||7/18/14 6:09 AM|
Shaun, sounds like a cool idea, and the website looks great, good luck with it.
The way the pricing is presented as less than a cent per check sounds like good value to me.
* Typo under the 50 million ticker thingy: " ... we have have your website covered" - the word "have" is repeated. Delete one.
* Don't use the Google website as your example site to be monitored. That confused me and made me think Google was part of the sign up process, part of "how does it work". (Furthermore, your example site under 'How does it work?" could show me something that's gone wrong - maybe a nice image in the before screen shot and missing image icon in the after screenshot - illustrating the point of your service.)
|Re: [SiliconBeach] Introducing myself and why my startup is doing it wrong||simran||7/18/14 6:27 AM|
I'm yet to find *anyone* (including Amazon employees and their billing department) who actually understand AWS billing :) it's harder them reading your mobile bill which are intentionally designed to be hard to understand! :)
Personally i like "packs" - that way there are no surprises... perhaps purchase X credits for $5 and i'll get alerted when that's about to run out and i can top up - i know for sure i will not be surprised with a $100 bill at the end of the month!
SMS on downtime is "critical" - you gotta get the site back up... SMS on a change... i suspect that will be "critical" in quite rare circumstances... so my guess is that not that many people will go for the SMS option regardless of price... so perhaps the SMS prices are okay, because in the "rare circumstances" that someone needs them, they will be willing to pay the premium.
I am possibly the farthest thing from a pricing expert... but i would think about it in the following ways:
* figure out your cost (bandwidth + cpu + overheads, etc)...
(bandwidth could be quite variable as you have go download each component of the page every time - and that could mean large images so it might be best to charge per gig rather than per request?)
* add a profit percentage (perhaps 20%)
* add "value" profit - which in your case i think can be "massive" - you can really do value based selling as it doesn't appear you have too much competition at the moment, you are pioneering functionality in this space and you seem to have a good quality, well executed product) and that might give you a price to start testing with?
For low frequently users like me, i'd pay $5 per 1000 request easily... to me that's 3 years worth of checks on one page (daily checks)... come to think of it, i'd probably pay $5 per 365 requests :) (but if you force me to check hourly, then it's way to expensive for me at least...)
With regards to me not noticing the price... you've probably sacrificed some usability for beauty... the front page is beautiful, but i just noticed that the top nav (which clearly has a link for pricing) doesn't come up unless i scroll... and i didn't scroll last time, i went straight to signup, then login... and after logging in, i could not see the price anywhere - i didn't think of logging out (or going back to the home page to see more menu items).
I suspect you are storing the "frequency" and "last_check" in a database and then querying the DB every second to see what needs to get run... if something like that is the case, why not let me set my check frequency myself... if i want a check every 22days 10hours and 15seconds, then let me enter that (in an advanced mode if need be)... :)
As perl's philosophy goes... make simple things simple, but make complex things possible :)
Overall i have to say... congrats on an absolutely BRILLIANT product... so refreshing to see more than just a landing page and vapourware :) Lean is "overrated" and more used like a cliché now... there's a lot to be said for building and releasing something you believe in :)
|Re: Introducing myself and why my startup is doing it wrong||Shaun Thambyah||7/19/14 5:55 AM|
Thanks for checking it out, appreciate the feedback on pricing.
Sorted the typo, thanks for the pick up.
Thanks for the feedback on the "how does it work" section, I like the idea of, before/after screenshots showing what we do. Right now it seems confused between a "how does it work" and "what does it do".
|Re: [SiliconBeach] Introducing myself and why my startup is doing it wrong||Shaun Thambyah||7/19/14 6:10 AM|
Wow, thank you sincerely for taking the time to sit down and write some really concise and constructive feedback. There's a lot to be said the Sydney Startup community. Genuinely humbled.
I'll sit down next week and take a solid look at pricing and packs. I took a similar approach to your suggestion below to come up with a baseline cost and worked from there. The 'value' component is a tough one to judge. I'm taking the approach at the moment that the market will help me set the price. However as you mention, the check frequency means market players could actually be excluded because of a small config item. I like the idea of allowing frequency to be entered in an advanced mode.
I've been looking over the code base and perhaps packs are a good idea. I'm thinking offering both would open the product to more customers. Generally those that either aren't comfortable being billed monthly or want a solid idea exactly how much they can spend.
I really appreciate all the feedback, most importantly it's helping me understand the the different user demographics that exist - something that I would completely struggle with on sign ups alone.