Blog

Getting Things Done

Monthly Archives: January 2011

Alarm

Alarm

Get out of your inbox and onto a Teambox

The webinar was fun to do. 70 emails turned into 8 minutes of stream review.

What am I talking about? Watch for yourself. 2 minutes of intro, 30 minutes of answering questions and 8 minutes of getting things done. Easily 3 hours of work slimmed down.

The Method

Inbox on Teambox showing us the method

Facebook Statistics

Facebook statistics

Change is good, expected and… overwhelming

Even for a seasoned startup guy like me, the whirlwind of growth and progress never gets old or easy. When the gusts of change start blowing on your backside, it is hard to maintain the sprint that prevents the head from being underfoot. Since joining the amazing team at Teambox in April, here is where we have grown:

  1. Total users: 250%
  2. Paid users: 450%
  3. Monthly revenue: 1000%
  4. Annual purchases: 3000%
  5. Daily signups: 400%
  6. Visits a day: 300%
  7. Number of employees: 250%
  8. Cash on hand: 20,000%
  9. System speed: 400%

There are a few more metrics I could share but I think this gives you a sense of our dynamic and as my old boss Paul used to say, meteoric. These are the areas we will proudly boast about and continue to focus on working hard to maintain or improve. But, there are some other metrics that most companies won’t talk about:

  1. Customer service requests
  2. Attritions
  3. Product bugs
  4. Feature requests
  5. Email questions
  6. Phone calls

It is not that our numbers are poor, or any better or worse than another company’s, it is just that these are not the things to tout f you want to get people excited. I think they share a lot more in relationship to how hard you are working to succeed. So here goes:

  1. Customer service requests: Up 300% (Most are managed in 24 hours)
  2. Attritions: Up 300% (Only logical, as you gain more customers, you will lose more customers. This is a place for us to improve as we work to make all of our clients happy)
  3. Product bugs: Flat. We have done an amazing job at keep our bug list to a minimum.
  4. Feature requests: 1000% growth.
  5. Email questions: 500% growth
  6. Phone calls: I get 300% more a week

By being aware of these numbers we can start to manage the ongoing tornado of explosive growth. We will get bigger and if we are good, smarter. Overwhelming is good if you get to take a deep breath and ride the excitement.

Motivational windows

Motivational Windows

Shortcuts

Shortcuts

Real Translations

Real translations

About to release Teambox 3

For the past weeks we’ve been working on a major update to Teambox. Because it’s such a big update, we’re pushing a version on this one and calling it Teambox 3.

On one side, we have migrated to the latest version of Rails, the framework that powers our application. This means all the new developments in the Rails community will be available for the next releases of Teambox, including security updates.

On the other one, and more exciting, we’re rolling out a complete redesign to the navigation system. We love it, and personally can’t wait to use it in our day to day while working with Teambox.

Expect a short downtime today at 12pm CET, while we upgrade our servers. After that, all accounts will have access to the latest version of the collaboration software we all love!

You can follow the migration’s progress on Twitter: @teambox_app.

Todays service interruption

Today Teambox.com suffered a very high load of malintentioned requests trying to find typical vulnerabilities in old pieces of software. The aggressiveness of the bot was high enough to cause us load problems.

The biggest problem came when the requests tried to fake logins with invalid cookie tokens. This revealed a performance issue in our database, which we were able to resolve quickly after a couple of DB tweaks. The performance of our public site has doubled as a result.

Another issue was that we couldn’t see nor ban the malicious IPs from nginx, as most of the traffic came through https. The IP nginx saw was HAProxy’s and it didn’t inject the X-Forwarded-For header with the original remote ip. After working closely with Engine Yard‘s support team, we managed to resolve this.

At the end, we only had service interruption for a few minutes, and after blocking all the illegitimate traffic from HAProxy performance returned to normal.