This is our first post in nearly a year. We’ve been busy behind the scenes, but didn’t have anything that had a visible impact on our users. If you do not care about the last years of events skip down to the last couple paragraphs to find out what changes we are making to notify.me service.
A bit of history.
From May of 2009 we put notify.me on autopilot and started working on extending the service into a b2b product, code named firengin. firengin hooks our notification engine into intranet monitoring products (think nagios and the likes) and be able to intelligently route/escalate notifications to operations teams. We added bi-directional communication so a team of people can effectively chat about a series of monitoring alerts with an integrated event stream. Everything is tagged, archived and indexed for simple search. We added features to automate post-mortem reports, markup/tag portions of the conversation to build a knowledge base, and import into the wiki for reference later. We had discussions with a couple high profile companies and established that there was a market.
After progressing to the 60-70% mark we decided that it would be best to scuttle firengin and shut down the venture. Here is why:
- I spent 50% of my available time traveling to SF/SJ (we’re based in San Diego) talking/working with venture and angels. Raising money is a pain. Unless your have an extremely hot web property or a proven VC track record its difficult to break in. Spending a large amount of time on things that do not directly involve your product is distracting, harmful and counter productive. I’m 100% sure that my next venture will be structured to get to critical mass before requiring external funding.
- We changed focus to the B2B model a year into starting the company. The entire team had already been working for quite a while without a paycheck. As long you continue to make visible progress it’s easy to motivate a team — Favorable customer/tech blog reviews are great to keep the spirits and productivity high. Switching focus to a product that needs 6-9 months of development before it can be publicly announced had devastating effects on motivation.
- Selling into IT is long sales cycle. While the overall US yearly budget for IT spending is quite large ~200-250 billion it has been decreasing for the past 5 years. Especially in a recession, selling new tech into the IT space was hard work. The problem is not going from 1000 sales to 10,000 sales, its getting the first 100 customers. Those first 100 customers need to take a big leap of faith by making the investment to integrate the new service into their technology while paying a premium for service. It is a tough, long sell.
Why did we decide not to just concentrate on notify.me?
It only took a couple months to see that baseline notify.me would be a niche service. There is nothing wrong with the niche server unless your revenue plan dictates that you need an extensive user base. Niche services need to deliver a large value proposition.
Fast forwarding to today, notify.me has seen continued and steady growth. Here are a couple stats I gather up yesterday.
Notifications delivered per day:
- IM 502k
- Email 68K
- SMS 32k
Amount notifications clicks per day: 61k
notify.me has been supported mainly by myself http://twitter.com/mp3tricord and sdether http://twitter.com/sdether. We’ve used a couple partners to help support the service which brings me to the crux of this posting. A long time ago I made a promise to myself I would never manage another qmail/sendmail server. So we used an outside company to deliver our 100k email/sms messages. We ended up choosing fastmail.fm because they one of the best email services on the net. Rob Mueller runs a stellar operation and has helped notify.me tremendously. However fastmail’s email service was never meant to be used as a SMTP gateway and after the sale to Opera (congrats guys on that) it looks like they are politely requesting us to stop by the end of the month. So we are planning to do just that.
So what does mean for our users?
The email situation left me with three choices.
- Switch to another email company that specialized in SMTP delivery. There are a couple http://aweber.com http://postmarkapp.com/ http://www.socketlabs.com/ http://sendgrid.com/
- Run my own SMTP mail server (maybe postfix this time around)
- Stop sending emails/sms
To send about 3 million emails a month through the cheapest company listed in #1 is about $2k/m, too much for my budget.
Running my own SMTP server is breaking a long standing promise I made to myself. Life is just too short to manage your own mail service.
Which leads us to discontinuing email and sms delivery. When I started notify.me the idea was to deliver all notification via Instant Messenger services. It was the only way to guarantee a timely delivery. Our IM service uses Xmpp/Jabber because its federated nature meant the extensive base of Xmpp accounts out there could plug straight in.
Unfortunately, even people that have Xmpp accounts often do not even know they have one. If you have a gmail or facebook account, you have an Xmpp account. In addition every user that signs up gets a <username>@notify.me account that can be used via many clients, such as Pidgin, Adium, Trillian, etc. There are also Xmpp clients for most smartphones, which unlike SMS does not incur a service charge.
Our plan is to stop email and sms delivery of notifications on July 27th. After that the only way to receive notifications will be over IM. We do apologize to all the people that rely on email/sms delivery. If you don’t want to use your IM account one other possible solution is to move your email notifications over to feedburner.
Email was meant to get users unfamiliar with IM to use the product and then migrate them over. Unfortunately we did a poor job communicating how to make the transition and we apologize for the rather abrupt 1 week transition that email and sms are now left with.