Ask HN: This is my MVP - would you pay $10/mo for my monitoring service? | Hacker News Reader

Ask HN: This is my MVP - would you pay $10/mo for my monitoring service?

11 pointsfamfam16 comments

Here is what I propose to offer. Would you be interested in this? It this space too crowded? Am I adding enough value/reducing enough cost?

For $10/mo

10 url checks - alarm on status codes, header regex, content regex, content size, response time (DNS* lookup, first byte, content download, or the whole thing), etc. I want this functionality to be the key differentiator. I will capture the full response (headers + html) on failed requests and store them for 90 days, so you can visualize them.

10 system checks - Any combination of Ping, DNS, SSL cert, TCP, UDP, SMTP, POP3, IMAP

All checks can be tested on the spot to ensure that they're set up correctly. (Will not be recorded as 'official' tests).

100 notification credits - will integrate with twilio to provide sms or voice alarms. 2 credits per mms notification and 5 credits per voice notification. Notifications via Twitter or Email are free/unlimited.

10 check executions per hour. (e.g. 6 minute interval)

System will only initially run out of a small number of US DC's, but if it takes on it will expand to worldwide testing nodes covering multiple backbones. (IOW, the MVP won't be very good at dealing with network reachability - something I've never found much value in with my own monitoring)

Interface will strive for extreme simplicity.

----------

Post-MVP:

Once I get some sense of my costs (how much hardware it takes to run say 1,000 customers or 10,000 urls on), I want to really kick up the offering on the url check. I want to move to full DOM simulation with click checks etc either via HTMLUnit, Selenium, WebKit, or something else. This will be a premium offering.

Android, iPhone, desktop integration.

Agent software so you can perform additional checks from a location of your choosing, on whatever interval you like. Likely be Java. Would it make sense to open source this? Would it be giving away too much? Maybe open sourced after we gain "significant" traction. Or maybe it would be developed openly, but not allowed to be used for commercial purposes or relicensed.

Two way API, custom checks.

This is a very, very crowded space. I suggest you have a good idea for marketing which cuts through the noise prior to trying to sell into it. Many of your competitors give away the 90% solution as a loss leader just to get a foot in the enterprises for the (much more expensive) last 10%.

Have you spoken to customers about this already? Who is saying "I already pay for monitoring but crimety I keep getting false negatives because while it checks status codes it doesn't let me fire regular expressions against the content."

Scrap the credit system. It adds mental overhead for customers. People who pay for monitoring systems will generally tend to run fairly reliable setups. If you get one idiot on the service who needs an SMS every hour, you can either educate him, rate-limit him, fire him as a customer, or just subsidize it from your customers who know what they are doing.

Charge more. If I care about uptime, I care a lot more than $10. You'll get pushback from poor college students. So what. They are terrible customers to have and you won't keep them, anyhow, because somebody can afford to offer this cheaper than you can.

See my larger response below. In my own experience the detailed monitoring was always what found our problems and outright 50x was very rare. But I worked at a very large etailer with a complicated SOA, so it might be a "head only" problem. I think my basically plan to was to try to dominate the low level offerings by brining in features that are only available to the high end offerings, even though for the most part they don't require a ton of extra work (browser simulation being the main exception).

Good point about scrapping the credit system. My gut says you're right. I guess the legalese would just need to be rejiggered to say "we reserve the right to put you on a leash."

I think you can differentiate your product from the sea of competitors in the same way Apple succeeded with the iPod when there were dozens of MP3 players already on the market. Rather than competing on features (GB, bitrate, etc), they appealed to the masses with "1000 songs in your pocket".

I'm a developer, so I know what ping, DNS, TCP, and UDP mean, but for the average Joe with a Wordpress blog, maybe a more compelling sell is:

Check the following: Is my site up? Are my ads showing? Has my site been hacked? Is my site slow? Is Google Analytics showing? Is my email up?

Marketing to this audience might run counter to the "charge more and eliminate needy customers" market, but you might find other ways to generate revenue. For example, you can run analytics on different hosting companies and see which ones are consistently slower than average, and market the fastest ones to your clients (at $100 a piece affiliate revenue).

These are great points. You may be onto something. But (and it seems like so many ideas ultimately come back to this!) how would you market to these customers that don't know that they want these things to be monitored? It seems more like something that a blog platform would make as an offering, anyway - an integrated feature. Because it's hard to imagine someone getting paid monitoring on their free blog.

Yeah that's the tough part. Since these people are the least likely to pay (why pay $10/mo to monitor a site that makes $20/mo), you may want to have a free option, subsidized by the affiliate revenue.

Then you can try to upsell them to your paid plan.

Thanks for the responses so far folks. It's interesting, because I raised this idea in a separate thread, and there was a lot of "i'd use that" kind of responses. Glad I'm double checking... :)

My initial thought (and maybe my MVP proposal does not reflect this) was to bring some of the features that things like Gomez offers (detail timing, screen caps, transactional walkthroughs, etc) down into the pricing space of the lower end offerings. I've worked for a large internet retailer before where we used Gomez - talking tens of thousands of dollars a year for just a few tests. Ridiculous. But having used it, at least on our site, I can say that it was definitely the fine grained monitoring that saved our asses -- things would NEVER outright 50x, just fail in really subtle/glitchy ways. Maybe it was because our system was for more complicated than Johnny PHP/Rails. IMHO 50x means the network is down, the IT guy unplugged something, you're getting DDOS'd or your system is just inherently unstable. It never meant "Apache died" - because Apache never did die, for years and years. Maybe I'm wrong to try to take "head" features and push them down to "tail" pricing.

Perhaps I could try to create a higher level offering that's still more affordable than Gomez. I don't know what that pricepoint would be though. Watchmouse has a transactional/clickthrough offering for $500/month. They look hard to compete with - their offering is quite robust. As for other offerings, I don't know. They don't tend to publicize a lot of their pricing, seem to have a not-so-self-service sales cycle.

I don't see the point of trying to create a "cheaper than cheap" option and competing with the simple-ping tail.

I guess I just though I saw an opportunity where some people may be getting greedy with their margins, and thought I could fill in that hole with an offering that was equivalent where it mattered, but at a much lower price point. Low enough where people in the tail would maybe even step up into it. Perhaps the world is perfectly happy in a bimodal setup - the "<$5/mo ping and L7 check against one URL" type of monitoring and the ">$x,000 or $x0,000 year total solution" type of monitoring.

As stated by others, this is a very crowded space. To be honest, I'm pretty happy with the free monitoring services I get from Monitis and Pingdom.

Your primary value differentiator (header/response codes against a regex) are more overhead that I have to create when I sign up. For me, just knowing if the system pings and responds on port 80 and 443 is sufficient.

Also especially agree with patio11 about the credit system. I'd only bring that in if certain customers are costing you too much to SMS.

For me, you need to be able to provide more sophisticated services. Just because you can ping a server is not a valid reason to think you can charge money for it. These are the additional services I need to see included in order to pay $10 /month:

1. Time the download of resources on each page, similar to what you get from HTTPWatch. Report on which ones are blocking and which ones are giving 404s

2. Provide a way to script the call, with user name and password, and the ability to store cookies and follow the redirect URL, up to N(5?) levels. This makes sure that not only the home page is up but a (test) transaction can also complete as expected.

3. This is more optional, but provide a sample page that users can host on their end or an API they can leverage to restart app or take similar actions, so that if the monitoring finds a problem, the answer is not calling the admin's phone.

#1 Totally makes sense. But what constitutes a resource to you? img, script, css? And from my perspective. The challenge for me to overcome is bandwidth. Also, define blocking. You mean long download times, or literally blocking the browser, script-wise?

#2 Any transactional test would obviously maintain cookies across each step. I would definitely want to bring transactional tests into the picture. V1 might be a URL sequence. V2 could be actual browser simulation.

#3 That's be easy, for sure.

As others have said, this is a crowded space...with a lot of opportunity for a clear leader to emerge (IMO).

$10/mo isn't a bad deal, but I really think you need to go way above what you're thinking here to dominate this.

Offer to keep a backup/cache of the site at a known point (this is basically a 1-line wget script). Offer some haproxy-bsed auto-failover auto-maintenance page stuff for when a site does eventually go down or get fubar'd.

From my personal perspective, it's good to know when something goes wrong, but it's fantastic to also be able to have some auto-fail or recovery option available (automatic, or user-initiated from an iPhone app) so that if I can't get online to rectify the situation promptly, I at least know users are being hit with 404's and SQL errors, etc.

The low-end of this market is not so hot anymore. Young, un-sung startups (the ones you are targeting at $10) have started switching to managed application environments such as Heroku and App Engine.

It's a valid consideration, but I'm sure you can't quantify what percentage the market has shrunk by in light of this (nor can I). My gut would say no more than 5%. Until you can run Zencart/Xcart/Wordpress etc on Appengine or Heroku, there are going to be a lot of people who need monitoring. :)

how would this be better than watchour.com ?

if i run out of notification credits, do i not find out when my web site goes down?

i hope this feedback helps.

More and richer checks per dollar. Better reporting. Better interface. Active development. Post-MVP will add a lot of features they don't have.

>Post-MVP will add a lot of features they don't have.

Sign up with me and post-MVP you'll get a free pony! ;0)>

Too many promise the Earth; if it's tied in a bow and waiting in your warehouse then forgive my cynicism.

(Edit:too)

Okay, except I'm explaining my plan to a peer, not pitching a client.

Reply on news.ycombinator.com