Author Archive

We’ve been hearing lots of positive things about one of our newest Chartbeat Publishing features, Heads Up Display, which lays data you need over your own webpage and article content. Heads Up Display shares info like the concurrent visitors on the page, top articles on the page through simple rank pins, clicks per minute on each article, and a whole bunch of other stuff. And while it’s amazing to know that your top positioned article is the #1 article on the page – there was one question we (and lots of you) wanted to know: Is that article performing as well as other articles have done in that same position in the past, or is it just doing well because it’s in the top spot?

So we’re testing exactly this question with a new feature in beta right now that we’re calling Link Performance– all to get your feedback, thoughts, high fives, and no ways. With this guy now incorporated in your Heads Up Display, you can answer questions like:

  • Is an article getting a ton of clicks because it’s new or in a prime position or because the content is actually compelling?
  • Where are the hidden gems on my page?  Which articles are actually doing well for their position, but are maybe getting lost in the clutter?
  • Which headlines look like they need some love?
  • Which articles need to be removed?
  • Has my article reached its peak and need to be swapped out?

Here’s how it works: Link Performance measures certain effects like where the link is on your page and how long the link has been in that particular position. This measurement also takes into account the real-time speed (what we call click rate) that readers are clicking on a link to an article as a percentage the overall pool of people who could actually click now. We’ve built models for every position on your homepage – yours specifically, not some general homepage template – and these models take into account the amount of time each article has been in its current location. By knowing how long an article has been in that spot, we can tell how it’s done performance-wise since you posted it – has it reached its peak yet or still gaining speed? That way you know when to swap it out. We’ve added a visual tracker that shows just this – how long a link has been in this position and how it’s performing against other links that have been in that same position. The Heads Up Display uses these models to determine how each article is performing relative to those three factors: article position, how long the article has been in that position, and the number of potential readers. Simple, color-coded signals show you how an article is doing in real-time:

Overperforming/Green = Good, the article is doing better than average

Performing on average or baseline/Yellow = Potentially good or time to swap it out. Depends on the content and your call, as the decision maker, if it should be changed up.

Underperforming/Red = Not so great. Time to get another story in here that’ll do better in this position than this current story.

If you’re seeing gray pins, that could be for a number of different reasons, most likely because it’s a brand new story/placement and we just need another minute to gather and share the data.

Based on all this info, we hope you’re better armed to take action to shift things around, rewrite some headlines, layer on more content, whatever a particular piece needs to get a through-the-roof performance out of that link in that position (or another).

Again, this is something we’re testing to see how you’re using it (or not), what you like (or don’t), and what thoughts (or questions) you have as you play around with it. With your feedback, I, along with my fellow Chartbeat data scientists and our Product and Engineering teams, can iterate and build on this feature to make sure it’s as useful as possible.  So let us know what you think.

Note: The three above images are displaying mock data (we don’t share your data without your permission)

Real-time analytics like Newsbeat enables you to understand what’s happening on your site right now, but it can still be hard to understand whether an article is performing better or worse than it should be. Today, we’re trying to make that easier by introducing benchmarking for articles. Our algorithms model the typical path of a story published at that time on your site over the first 12 hours of its life cycle and give you an instant understanding of whether your article is performing above or below that typical path. Simply go to the performance tab and click on an article (our algorithms automatically ignore landing pages) and you’ll get a sense of how that article is doing.

In the example above,  the red line showing actual performance is just under what would be typical performance for a story on that site published at that time; it is underperforming against the benchmark. In the example below, the red line showing actual article performance is beating the benchmark of typical article performance; it’s a great story!

When looking at your top 10 stories you should expect them to beat typical performance by quite a bit; that’s why they are your top ten stories! However, as you delve into stories below the top rank you’ll see their performance converge with typical and instantly be able to see when an unexpected event like a link from Drudge or spike from social moves you away from that typical performance. Once you’ve published a story, come back to our dashboard and see if it is out performing your standard articles. If it’s falling short, make sure you’re taking every opportunity to promote it through your site and social media. When a story is exceeding expectations, then it might be time to give it some homepage love, or understand how you can apply this success to all your other content.

We’re constantly working on improving the models, incorporating more factors, and providing more context around the data to ensure that you can extract as much value from our product as possible. Please email us anytime at:, we’d love to hear from you.


One of the biggest values of a real-time data product is its ability to highlight unusual activity on your site – activity that directly translates into actionable information. However, in order to identify what’s “unusual,” you must have a good sense as to what is “usual.”

Over the course of our research, we’ve found that the number of active visits on an article, as a function of the time since it was posted, generally follows a few typical patterns. Armed with those patterns, we designed a newsbeat “Spike Alert” system to identify and alert our users when an article’s traffic suddenly and unexpectedly deviates too far above its expected levels.

For every article, we evaluate the incoming traffic on a source by source (i.e. Social, Search, Links) basis to identify whether there is a spike in traffic from that source.  (The number of active visits in the Spike Alert email is specifically from that source, not for the article as a whole.)

We determine that a spike is occurring by taking into account a number of factors, including:
  • Predicted increase in traffic. This is to avoid sending alerts for expected, and therefore, uninteresting, increases. For example, an article posted at 2 a.m. that starts to pick up traffic at 9 a.m. is not that interesting.
  • Rate of increase in traffic
  • Absolute increase in traffic
Newsbeat Spike Alerts are configurable at a site, section, or author level and can be received by email, SMS, or iPhone push notification.

As an initial means of throttling the number of alerts we send, we created a default thresholding system based on the site’s 30-day maximum number of active visits. Sites with large amounts of traffic will have a higher threshold than sites with smaller audiences. Going forward, we’re exploring a number of different options to allow for more customized Spike Alerts.

It’s incredibly important that we hear your feedback, so we can continuously improve our product. Please email us at with questions or feedback on Spike Alerts, or any other newsbeat feature!

Recovery Update

April 29th, 2011 by Isaac

Update: 1:17 PM

We have completed our DNS migration to Dynect ( They are the industrial strength DNS service powering Twitter and, and have been incredibly responsive. We’d also like to thank Akamai ( who were also extremely helpful throughout.

Once again, we’re sorry for the problems earlier today and we’re doing everything we can to make sure we’re using the most reliable services available moving forward.

We’ll be posting a detailed explanation of the DNS issues later this afternoon. In the meantime, please reach out to us at with any questions or concerns.


Update: 12:10 PM

We’ve completed our migration to the new DNS provider. Traffic levels in the dashboard will begin to get better soon but, because of caching, it will take a while for a complete recovery.

We’re sincerely sorry for this problem and we’re here to answer any questions you may have about your site and chartbeat. Please send us an email at anytime.


Update: 11:13 AM

We’re in the process of moving to a new DNS provider. As the correct IP address of the nameserver begins to propagate, you will see traffic numbers on your dashboard recover.

We will continue to update as we see improvements on our end.


Update: 6:39 PM

We’ve completed the DNS migration and the changes are propagating across the internet. Numbers should return to normal soon. A full post-mortem of the situation is posted here.

Thank you for your patience.

Last evening, we began to see domain name resolution issues with our DNS provider. As a result, a subset of chartbeat customers are resolving to an invalid IP address for the subdomain and are unable to load the chartbeat ping javascript.

This doesn’t have any effect on the users visiting your site, but it means that some dashboards will only show a sample of visitors for a little while.

We’re working to migrate our DNS records to a new service provider. We’ll send another update shortly.

Thank you for your continued patience. We’re really sorry and we’re working as hard as we can to get everyone back to normal.