We are doing something different here at PipelineDeals by sharing with you the tools we use to run our business. Traditionally we haven’t really shared much about what goes on behind the scenes so you might think of this post a first-time view behind the curtain. You can also follow the recently launched PipelineDeals Dev Blog for more of the hardcore tech details around the business. We were inspired to do this post by some of our peers in the SaaS world from the likes of this post from Gabriel Weinberg and this post from HelloFax. I don’t know if we will ever share as much as Balsalmiq, they are 100% open book and that works for them. So if you want to see what it takes to run a SaaS B2B application at scale, here we go:
Integrating all your quantitative IT data in one place is a tall order. Even while still in beta, DataDog lives up the challenge. Datadog is a social platform letting you not only slice and dice metrics from your applications, hosts and cloud providers but also share what you discover with other members of your team. Their metrics are still pretty basic and new features and third party integrations have been slow to evolve but its still early for a company we like where they are taking things. Did we mention their fanatical support? We are big fans of supporting our customers in anyway we can, glad to see someone else is too.
Employees of PipelineDeals are spread all over the US. It's important for us to have a robust communications platform. Over the years we've tried a multitude of ways to stay in touch, including Yammer, Skype, and Google Chat. All of these, we felt, were not cutting it. Only when one our devs proposed running our own IRC server did the a-ha moment strike, and we found the product we currently use and love, HipChat. Today we're all on Hipchat al the time. We have separated our "rooms" so that developers can talk developer talk, the support team can colloaborate on customer issues, and a general "Company talk" room for general news and announcements. Hipchat has achieved our goal of bringing our team closer together, even though we are all far apart.
Whether it is 1 PM or 2 AM, Pingdom makes sure everyone on the team knows our application is down. We've found it to be reliable and indispensable.
Amazon is the de facto standard for growing your cloud based infrastructure from 1 to 1,000,000 users. There is an art to surviving in the Elastic Compute Cloud (EC2) ecosystem and we believe we are getting better at it. Back in April we fell in love with the simplicity and power of their Elastic Load Balancing service, started avoiding EBS for mission-critical functions, and begun spreading ourselves out geographically. AWS may make growing your infrastructure easy but don’t be fooled, scaling in the cloud still requires a complex, coordinated dance of not just well-applied technology but measuring, testing, benchmarking and thinking hard.
If there is one thing every web developer hates dealing with, it's DNS. It’s a classic double edge sword, easy to get right and also really easy to get horribly wrong. You shouldn’t choose your managed DNS hosting provider for their beautiful user interface and DNSMadeEasy’s clunky, confusing retro-90s interface won’t be winning any awards soon but their great support, built-in monitoring, automatic failover and global traffic redirection will. Their uptime is phenomonal, and they are the best in the game.
Over the years we have tried quite a few bug tracking apps. The one that really stood out was Fogbugz. Their interface is beautifully functional, and we found that Fogbugz, of all the bug trackers, really suited our style of bug tracking well. In addition their support and level of service have been fantastic. The app is fast, customizable, and always up.
The other main selling point was for our customer support. Fogbugz integrates with Google Mail, so any time a customer emails us, a ticket gets automatically created. This allows the support team to ensure that we get back to all customer requests, and there is no toe-stepping (E.G., 2 replies to the same email).
A simple intuitive and now shareable list and to do app. As thin or thick as you want it to be. We love it and use it everyday.
Google Apps for Business (Email, Documents, Calendar, Chat)
We integrated PipelineDeals with Google apps for domains, and that has been a great move for us. Internally we use Google for Domains to run our email both for individuals and for our Customer Operations (but not without FogBugz). Google Docs are a no brainer - in fact one of us in Seattle and another in Houston co-wrote this email using Google Docs. Google Chat was the go-to standard for a couple of years but HipChat wins the award as the Google Chat killer for us internally.
We don’t send marketing emails as much as we should but when we do we use CampaignMonitor. Our marketing team appreciates it for its simple and straight forward reporting, a sleek design, and helpful (and responsive) customer support.
Google+, well primarily Hangout (ok, ok, we don’t pay for it but we really use it.)
Say what you will about Google+, Hangout is a win for us. Our team spans five different states here in the US. Hangout brings the whole team or subsets of the company together and let’s us get things done in real time. Written communication, even in HipChat, can be lacking and cause misunderstandings. Face to face, one on one, or in a group reaps all the benefits of in person communication and you get the added benefit of not having to deal with bad breath.
We love the Chef configuration management framework but at times we feel like it uses a battle axe to carve a pumpkin. This is especially true when it comes the components needed to run even a basic Chef server. Before you know it you are managing everything from Ruby and Java to Solr, CouchDB and RabbitMQ. We want Chef to work for us and not vice versa so we were not ashamed to pay Opscode a very reasonable fee to manage our Chef server for us.
What if your email service provider gave you detailed searchable logs so you could see what happens to every single one of your messages? What if they offered built-in analytics and API hooks to send events like opens, clicks, bounces, subscribes and unsubscribes to third party apps? Don’t be fooled by their small time image these four dudes really know their way around email and despite their confusing pricing it is very competitive if you do the math. So leave the hours of grokking postfix and spamassian to folks who know what they are doing.
Monitoring the performance of your web application is not an option, it’s a necessity. So right after you sign up for Pingdom (because who cares how fast your app is if it is not even up) head on over to NewRelic. Their signup and integration process is a relatively straightforward and painless process plus it’s thirty minutes you won’t regret. Wondering why your site slowed down all the sudden? It’s probably because a customer just imported a ton of data or maybe your database is about to keel over or even better your on the front page of Hacker News and your requests/min just shot up ten fold! You won’t know, but NewRelic will.
We decided to leverage RightScale when we planned our move to AWS a few years ago in order to not have to build everything ourselves. Without a dedicated systems administrator it was great for turning raw EC2 instances into usable load balancers, application servers and databases with little to no effort. Despite all these goodies RightScale has never invested as much in their support offering ($300/hour) as they have in their platform. Today it remains as expensive as ever and response times only marginally beat AWS.
MixPanel provides a level of visibility into our application with helpful metrics like funnels and cohorts that are not in Google Analytics. We know we need to spend more time with this data to improve ourselves. Right now, it is worth it to be capturing this data over time and know that we can dig into any specific event or data set at any time.
We are giving Squarespace a go after a few years on Typepad. Early results are good with better comment management and reporting. The basics are a little rough around the edges, for example, editing using the WYSIWYG editor (that we would expect to be flawless) can be prone to formatting errors. But other than that it is a step in the right direction.
We wanted something a step up from YouTube that delivered high quality video in a nice player with good metrics. We got that with Viddler. It was definitely an eye opener for us to see how much our customers use video to augment their PipelineDeals usage. And when we say use, we mean time spent watching, not just views. We did have some discussion about sticking solely with YouTube for driving traffic and exposure so we kept our YouTube channel going but all the videos on our marketing site and in app run off Viddler.