Archive for March, 2013

Chartteam Spotlight: Wes Chow, Backend Engineer

March 13th, 2013 by Cat

In my line of work, I could talk about culture and hiring until I’m blue in the face, but that’s pretty much only entertaining for me, less you. I mean, yes, we have an amazing culture and the smartest people I’ve ever worked with but that doesn’t give you an idea of what it’s really like to eat (snacks!), live (beanbag nap chairs!), and breathe (fresh Union Square air!) Chartbeat. So we’re taking it back to Kindergarten with some show and tell right from the source, our favorite Chartteam members. wes portraitFirst up, is the inside scoop from Wes, one of our crazy-experienced Backend Engineers who joined Chartbeat last year to help us tackle the sheer scale of data challenges we face every day. What he brings to the table: Wes spent eight years building technical infrastructure for high frequency trading, before falling in love with us. And before that, he spent five as a founder and Director of Technology of Athena Capital Research. As is often the case, he grew weary of the finance world over time and decided to join the startup community by forming S7 Labs, which specialized in developing tech for startups. At S7, they built Songza Radio and Storybox, a notable Seedcamp NY 2012 finalist. So what cool things are you working on now? We’re obviously working most often in the real-time space and on an average day, we’re dealing with an incredible amount of data -- peaking at around 130,000 requests per SECOND (which is crazy!) The challenge that comes up is storing and processing it over long historical periods at a very granular scale. I’m working on projects that will help that. Soon we’ll be rolling out features that provide Chartbeat-style action-centric context to those real-time decisions with additional weeks or months of data. So, we now need to have systems for efficiently writing and reading a decent chunk of traffic for offline analysis, and I’ll be making that happen. What challenges are you facing as we continue to grow? My job will only become harder in a good way as we take on more and more international clients (hello, China!), and as our clients grow their audiences with more and more visitors to their sites and apps. We'd like to push the limits of what's technically possible. Chartbeat did groundbreaking work to provide real-time analytics at scale; now we’re lucky enough to see that it's virtually a given that publishers have to have Chartbeat. We’re applying that kind of boldness to all of our systems, and aim to build the most insightful, intuitive product out there. What have you learned working here? It's rare that a job demands such a breadth of knowledge. I like to say that funny things happen at scale; you need to be aware of everything that's happening from the context switches your OS is doing to big abstractions like work queues and databases. Chartbeat also has a very open culture and a rigorous peer review system, which results in high standards and low defect rates. Sure we break things -- everybody does -- but we don't take that as a tradeoff for moving fast. DSC01028What’s the weirdest Chartbeat moment you’ve had so far? Getting introduced to the Chartbeat Ornithology chatroom is by far one of the oddest/most interesting things I’ve experienced so far. It sounds weird but, really, it’s a place where we talk about... well, we talk about birds. Naturally the conversations flow around everything from the Black Oystercatcher to the African Crowned Eagle - both notable winners of the famed “Bird of the Day” contest. Naturally. If you could stump a new hire with an out-of-this world question at our weekly Pulse Check company meeting, what would you ask? So I’ll literally take this question “out-of-this-world”. Tony butchered this question at our last meeting, but I would ask: Suppose you were selected for the first mission to Mars ever. There’s a catch (of course) which is that you can't have any contact with civilization for N number of years. What is the largest N you'd tolerate to be the first person to Mars? -- Clearly, we have some awesome real-world problems that come up every day that only the brightest (although slightly quirky) minds out there can solve. If you’re a data-loving nerd at heart and curious about why backend development at Chartbeat is interesting and challenging, get in touch with Wes directly or apply to one of our openings. We’d love to meet you.  

Next Week, We Invade Berlin and London. Watch Out.

March 7th, 2013 by Dustin

I just booked my flights and am preparing to leave chilly NYC for....chilly Europe. All to spend some quality time with our favorite clients across the pond, and to share with Europe what we've learned here in the US - and vice versa. So what have we learned about the differences in the industry across different countries so far? Well, at Chartbeat, we've always talked about the rise of the social web giving birth to the necessity of real time data. When search was king it was predictive and frankly not all that exciting, making optimization consistent and manageable. You knew which levers to pull to get which results. I'm of course simplifying a bit, but this was the general practice. But now with the social web completely taking over, we all live in an environment of complex and unpredictable traffic patterns and audience behaviors that require newsrooms and sites of all kinds and sizes to be set up to adapt to anything that's thrown their way.
Some of the most innovative newsrooms that are adapting the best to mobile and social unpredictabilities are outside the US.
We've been working with major publishers and media sites in over 36 countries for the last 3 years, so we're pretty used to seeing who does what well and who doesn't. We understand your challenges, your needs, your hopes and your dreams. And the best part of our day is helping you move beyond data to actionable, adaptable metrics, information, and insights to overcome, solve, fulfill and realize these. Speaking with many of our current European publishers, we've learned a few things:
  1. The homepage is king, so you need help making your homepage as adaptable to a returning loyal audience as possible
  2. Mobile comprises a large (and growing larger every day) portion of your traffic, so you rely on the difference between mobile and desktop consumption -- not just from a push-to-mobile standpoint, but learning the differences in editorial and content choices on each device, each experience.
  3. Social media as a major traffic source is a future goal but not a right-now necessity, so you want to maximize your current social impact to learn from it before you truly need to live and die by it.
This is very different from many of the the side-ways traffic growing, mobile-learning, social-media-is-life US publishers. Clearly all media needs are not the same and that's why we have lots to share from each of you to each of you: from the growing importance of internal optimization to maximizing sideways traffic, from moving away from the page view as the metric for success to understanding your audience's true online experience (not just your headline clicks) through Engaged Time, from looking at social audiences as another traffic source to treating them like you treat mobile - as a completely different needs-and-wants-driven group. 
The most untapped metrics across the globe
The one thing that newsrooms across the globe need to be using to make decisions is the qualitative measure of their audience's consumption of their content, their Engaged Time. This isn't something you measure in page views, since they just tell you if a link was clicked, or time on site, since it just gives you an estimate of how long your audience may have been on the page. Engaged Time , is a way of capturing if your audience is actually reading, writing, interacting with your content and for how long. It is a metric that allows you to make decisions based off the quality of your content and it's how and when you make decisions based off quality that should matter most. The value of measuring Engaged Time, is not only that it measures the quality of your content and helps uncover the actions that you can take to provide that quality to a larger audience, but also in what happens when you take those actions. Let me explain. Think about someone who reads an article that they like. What happens? They spend more time reading it. They might even comment/like/tweet it. Whatever they do/action they take, because they are engaged, they are also more likely to come back for that same experience tomorrow and the next day and the next day after that. (My good buddy Josh, one of our data scientists, will be giving you all the data work behind that formula later this week.)
That formula is the holy grail for creating a loyal audience.
Now, imagine being able to surface and promote these articles repeatedly to that right audience, the ones that are most likely to come back if they like what they read today. Now, think about what that means if you have a paywall. No matter where you are across the globe, page views just don't tell you enough. The media industry can't get what it needs based on clicks alone. You need to know what's happening between the clicks to actually grow your audience. That's what I'll be bringing with me to Berlin and London next week. Get your questions ready.

Chartbeat API + Google Spreadsheets = Historical Data Cheat

March 6th, 2013 by Ben

You know you want it. The elusive white unicorn that is historical real-time Chartbeat data. We constantly challenge our products to give you better historical context in creative ways (like replay and typical performance), but we totally get that you may just want to log snapshots of your dashboard.

Well, you're not alone. Many of our client-partners have built their own suites of cron jobs calling the Chartbeat API's and storing in a database at some frequency to use historical information to give more context to their real-time decisions.

But, for the average pretengineer or those of us who don't even have enough hours in the day to make time to eat breakfast, going the extra step to get more out of your Chartbeat data isn't something you can make time for. Adding databases and time-based job schedulers take a backseat to the rest of your life.

Enter Google Apps Scripting & Google Spreadsheets.

With a Google Apps Script & Google Spreadsheet concoction, you can easily leverage the Chartbeat API to store only the data you care about, longer than 30 days.

Google Apps Scripting is a JavaScript cloud scripting language that provides easy ways to automate tasks and most notably, write directly into Google Spreadsheets. Here's a simple function calling the Chartbeat Historical Traffic Series API with our default response settings (active visitors today, in 5 minute intervals):

getChartbeatTraffic

You can then write directly into a Google Spreadsheet of your choice  by converting the response into rows then inserting those rows on your favorite spreadsheet:

insertData

To take it a step further, you can schedule a frequency to run it the script behind the scenes. Maybe you want to log your daily max by traffic source every night, or store your average engaged time every-day at lunch. Adding triggers to automate your script is as easy as a couple drop-downs:

triggers

As always, if you have a question or maybe just want some code we've already built to copy into your Google Drive right away we're here to help. Drop us a line, shoot us an email, or visit support.chartbeat.com.

Who doesn’t love a free t-shirt?

March 4th, 2013 by Andrew

tshirt_graphics_3 Here at Chartbeat we love connecting with people through our design. We live to make it easy and fun to understand and act on your site’s activity. We recently had some time to help spread some love outside of your screen with some Chartbeat tees. Here’s a walkthrough of our design process.  
1. Brainstorm
tshirt_graphics_2 What would make a good Chartbeat tee?
2. Mock 'em up
tshirt_graphics_1 Pick the ideas are worth trying out and take a stab at what they could look like!
3. Refine the designs from "cool idea" to "wearable tee"
tshirt_graphics_4 After getting feedback on what people would actually happily wear, fine tune your work and ship it! If you you didn’t get a chance to get your hands on a shirt the first time or have ideas for the next round, get in touch!