Author Archive

An advantage of working at Chartbeat is that one has no shortage of data from which to derive insight. Our growth over the last year has meant we now track almost half a trillion events per month. This enables us to deeply understand patterns of user behaviour in 60 countries and across 50,000 websites. It allows a unique insight into where the world is going.

When we take a step back and look at what that information is telling us, we see three principal trends in the industry:

  1. Content teams often affect revenue per page more than ad ops teams
  2. The market for content has accelerated as traditional advertising suffers under the weight of the four horsemen: viewability, ad fraud, page speed and ad blocking
  3. Publishers have a deep desire to adapt, but there is considerable friction to them doing so

With that in mind, we’re announcing three new initiatives centered around what Chartbeat can do to adapt and increase the pace of change in the industry.

  • We are doubling down on our investment in content tools to meet the needs of our content creator customer base
  • To help publishers better make the case for the revenue impact of quality content, Chartbeat is opening up our Engaged Ad Refresh technology to all our Chartbeat Publishing clients and open-sourcing some of our newest ad measurement technology
  • We’re partnering with Moat to see if we can move the web faster through close collaboration.

Let me break that down.


Doubling down on content: Making great user experience means greater revenue

Content teams now often have more impact on per page revenue than ad ops teams. If a visitor spends fewer than 15 seconds on the page, average viewability is around 28%. If a visitor spends greater than 15 seconds on a page then average viewability increases to 60%. Publish clickbait and languish at 28%, publish content that actually captures attention, promote it with the right headline and it more than doubles.

Secondly, researchers at Microsoft Research found that a rigorous engaged refresh strategy delivers better per page performance for advertisers. When we assessed the potential impact over the Chartbeat network it also increases the total number of viewable impressions by 69%. Leveraging engaged refresh is probably the single easiest and most useful thing a publisher can do to survive in a viewable world. Both improving viewability and increasing inventory through engaged refresh have one common theme: revenue depends directly upon the user’s experience.

We’re going to be gradually rolling out our Engaged Ad Refresh technology to our Chartbeat Publishing clients. We want every publisher to be able to directly draw the connection between engaging content and greater revenue. To make that even easier, we’re also open-sourcing some key Chartbeat tech to allow publishers — our clients and otherwise —  who use DFP to better optimise for engagement rather than just clicks. Adapting for and building a business on attention should be determined by a business choice, not the technology you choose to partner with.


Collaboration beats Competition: Partnering with Moat

If we want the web to move towards a place in which user experience is paramount, then having every company sitting in their own silo trying to boil the ocean makes no sense.

The web was built with collaboration more than competition; each contributor an expert in their field contributing their skills to the global whole and building something magnificent as a result. If we want the web to get better faster we should do the same.

That means doubling down on what we do best and partnering with others to move the web forward faster. Moat has long been a voice in chorus with our own about the need to move away from clicks and towards attention. Moat has been the go-to for publishers on the ads side, just as Chartbeat has been when it comes to content.

As companies, we both believe that our ability to move the web forward will be advanced by working together. We’ve already seen customers derive tremendous advantage when they are able to stitch together our respective data sets, even when it hasn’t been easy. Enabling Chartbeat’s content tools and data to work in harmony with Moat’s ad tools and data can open up new ways for publishers to benefit without having to disrupt their workflows.

I’m excited to see what better collaboration can do for an attention-focused web.


The refrain is that when it comes to mobile the open web has had its moment and just can’t compete in a world of apps. While apps and the content within those apps (like Facebook’s Instant Articles) load blisteringly fast for a seamless user experience, the open mobile web is lethargic at best, weighted down with redundant code, crappy ads, and too many trackers. When we should be marveling at the instant access to the sum of the world’s knowledge from a device the size of a slice of bread, we spend our time staring frustrated at a loading bar and often just give up.

We have a choice.

We can choose to see the open web as a relic of its time and flee to the warm embrace of platforms and apps. Or we can say that the open web means something important to the world and if it’s broken it’s our job to fix it. I think we should fix it.

The open web has offered equality of access and information to the world. It’s meant that a generation’s unfiltered voice could make itself heard. It’s led to one of the greatest explosions of innovation the planet has ever seen. That wasn’t an app. That was all of us. And that’s important.

That’s why Chartbeat is — and I personally am — proud to partner with Google and many other media and tech companies to support and build the Accelerated Mobile Pages Project. It’s our attempt to fix the mobile experience and ensure the future of the open web.

What is AMP?

The AMP project is an open source initiative to deliver a significantly lighter-weight version of a webpage so that content on that page loads instantly, regardless of mobile, desktop, or any other format. It solves this problem at the page-level, not the platform-level. That means no matter if you discovered an article on Twitter, Globo or the Guardian; whether you open it with Chrome or Firefox; or whether you’re reading it on your iPhone or Droid, if the content was created in an Accelerated Mobile Page, you’ll have the same lightning-fast performance. It shouldn’t matter where you came from, you should get the same perfect user experience.

From the content-creator and publisher perspective, open really means open. Anything you publish using AMP HTML will look the same and load with the same speed no matter where or how your readers read it. It’s not a Google-specific link or a fix for only Twitter readers. It allows all of us to use the same open tools to create sophisticated reading experiences for the same open web.

How is Chartbeat involved?

Richard Gingras at Google reached out to us and asked us to be the founding analytics partner for the AMP project. I’ve always believed in the open web (my first startup was an ill-fated attempt to make OpenID popular) and we leapt at the chance to make the web better.

Our particular piece of the puzzle is working out how to make sure that analytics can have as close to zero effect on page load as possible.

Right now, when websites load, multiple analytics services load code that does largely the same thing. That means slower pages in exchange for little positive effect. It would be much better to have just one super-fast snippet of code load and report back on what’s happening on the page. That’s our job to figure out. We’ll be working to shape the universal metrics AMP will report on.

Our goal is to be able to enable the same level of insight for media companies while dramatically improving user experience.

What do we hope will come of this?

We hope the different media companies and tech services that support them can come together around this open-source standard and build a web we can be proud of. Just like the open web, anyone can create a page using AMP, whether you are a blogger in New Zealand or a mogul in New York. With the industry coming together around common standards it means that improving the page load of Vox also improves the page load of the New York Times. It becomes less about any one site becoming better and more about the entire web experience getting better.

At its core it’s a simple mission. Letting the web once again become a source of wonder rather than a source of frustration.

Last month, Chartbeat celebrated its sixth anniversary. Over those six years, we’ve worked with thousands of incredible partners; raised $15.5 million in venture capital; committed code to GitHub 72,000 times; watched the internet break over Obama’s election, the Egyptian uprising, and the color of a blue and black dress; and influenced how the media industry uses data.

In the spirit of if something was fun the first time you should do it again, I’d like to announce that we’ve taken a $15.5 million Series C investment, doubling the total amount invested in Chartbeat to $31 million. The round is led by Harmony Partners, joined by current investors DFJ, Index Ventures, Jason Calacanis and Jeff Clavier and new investor Digital Garage. We’re thankful for the faith our investors have shown in us and we will be celebrating internally.

However, new funds are more than an excuse for a celebration. It’s a promise we make to our investors and to our customers and partners that we won’t stop pushing forward, that we won’t stop building, and that we won’t stop trying to take this industry to a better place.

And this industry needs to get to a better place. We have a simple mission: We want to make sure that ten years from now, it’s possible to build a sustainable business around quality content. We want those in the business of content to succeed by appealing to their readers’ minds more than their index fingers, and to know that the quality of their content affects the value of the page.

With that in mind, today we’re launching two new tools for publishers’ editorial and advertising teams. In the short term, they help to solve the viewability problem for publishers. Over the long term, they help to prepare the industry for an economy based on attention rather than pageloads, an economy based on what matters.

We hope you like them.

Introducing Engaged Headline Testing

And a powerful new Heads Up Display

Chartbeat 2015_Heads Up Display_Headline Test Winner

One of the criticisms of data tools can be that they prioritize clickbait over quality. Tools like traditional click-focused A/B testing can aid and abet that issue, where the most provocative headline wins regardless of the content that supports it. Moreover, tools built for long product-focused tests can struggle when applied to the fast lifecycle of a news headline, where the half-life of the story has expired before the test has finished.

Our newest editorial tool — Engaged Headline Testing — is the first multivariate testing tool built specifically for media sites. For the first time, it optimizes not just for the click but for the behavior after the click. For a headline to win it must be both enticing enough to attract traffic, and also truthful enough to keep an audience. Not only does this optimize for audience engagement rather than pageloads, it also enables editorial teams to directly affect the overall ad viewability performance of their site.

In fact, we’ve found that by optimizing for clicks that result in more engaged time on a page — or Quality Clicks, as we call them in the new Heads Up Display — it means more ads are seen for longer.

A recent study shows when visitors are engaged for 15 seconds or more, viewability is 60% versus 28% for visitors engaged for fewer than 15 seconds.

Finally, Engaged Headline Testing works at the speed you need it to. It uses a propensity-modeling algorithm that assigns more traffic to the likely winner as the test is running, delivering better total performance than standard A/B testing.

For the first time, media companies get multivariate testing built for the goals they care about, at the pace they need — all designed to fit into their workflow in the completely-rebuilt-from-the-bottom-up Heads Up Display.

Introducing Engaged Ad Refresh

And automated line-item optimization

Chartbeat_2015 Line Item Tracker_Viewability Optimization

With Engaged Headline Testing, publishers are driving more people to more engaging, high-quality content. With our new Engaged Ad Refresh tool, they’re turning that engagement directly into revenue.

With Engaged Ad Refresh, publishers borrow an idea from TV and make money from the amount of engagement they can capture vs. the amount of pages that load. Chartbeat tracks the attention of each user as they interact with ads and content, and after an ad has received a set amount of Active Exposure Time, it identifies and serves the next best ad for that user. It’s making the web less like magazines and more like TV, where four 30-second commercials will run in a two minute break.

For the advertiser looking beyond viewability, an ad refreshed after 30 seconds of active exposure has received 30 times the attention required by the viewability standard.

For the publisher with engaging content, it significantly increases the amount of viewable inventory and makes the quality of the content directly affect the revenue opportunity of the page.

PGA TOUR has seen a greater than 10% increase in viewable impressions, meaning an overall increase in new viewable inventory by more than 10%.

For the visitor, it means sites will be built to hold attention rather than maximize the number of times visitors have to click on something.

In essence, we’re seeing the rise of the digital 30-second spot.

Making Engaged Ad Refresh tools work in a way that solved for both advertiser goals and publisher needs was one thing, but we wanted to build even more intelligence into how we made our ad choices.

Our line-item optimizer automatically serves under-performing line items to refreshed ad slots after a fixed amount of actively in-view time. This ad refresh technology is — in real time, in flight — optimizing for the campaign’s goal of hitting either viewability or Active Exposure Time targets.

It goes without saying this automation dramatically increases delivery efficiency for campaigns and reduces the need for padding upfront inventory or makegoods at the end of the campaign.

That means brands get the attention they pay for, not just what they’re promised.

That means publishers get paid for the attention the capture, not just the traffic they drive.

That means — most importantly — we all get a web that rewards and sees a whole lot more of the content that we actually want to read.

We’re rolling this stuff out in the coming weeks, so if you’re already using our editorial and advertising tools, just reach out to your account team and we’ll make sure you’re first in line. If you’re not using our Chartbeat Publishing editorial or ads tools yet, let’s change that.

Over the last couple of days, there has been a lot of talk about data privacy and specifically passing personal information (e.g., age, zip code, income) to third-party data sites. Chartbeat is paid by clients like to help content teams understand how people visiting the site engage with the content, so the folks behind these sites can create the best possible visitor experience.

With that in mind, I want to take a second to talk about what Chartbeat does and doesn’t do (since there are a lot of data providers out there and we all collect and measure different things), what happens when we inadvertently receive personal information in our data, and generally how anyone using data tools for their website can do so effectively while simultaneously taking care to protect their users’ information.

Chartbeat and the data we collect

First, there is a lot of legal stuff. I know, I know. But it’s actually all really important to know about us, what we care about as a company, and how we compare to other data providers out there.

  1. Chartbeat ONLY shares data about a specific client site or the visitors of that site with the owner of that specific client site. Under no circumstances do we share or sell that data to third parties, advertisers or otherwise. Specifically our Privacy Policy states: “Except as expressly provided otherwise herein, we will not sell, lease or exchange the personal information of our Customers or any end user (to the extent that we obtain such information) to third parties without first obtaining their express consent, unless required by law or to protect their status as a Customer.”
  2. Our JavaScript pings, which report information, only do so via HTTPS when client sites are HTTPS, to help prevent third parties from intercepting that data in transit.
  3. We do not ever intentionally collect personal information. As our Privacy Policy states: “Chartbeat does not collect any personally identifiable information from users of Customer Websites, provided that (i) Chartbeat does collect IP addresses from visitors to Customer Websites in order to show geolocation information, and (ii) the Customer configures the Chartbeat code on the Customer Website in accordance with the instructions and documentation provided by Chartbeat, so that URLs containing personally identifiable information of end users are not captured by the Service.”
  4. In our client contracts and Terms of Use, we specifically state that our clients need to scrub any personal information before passing it through to Chartbeat.

When personal information is passed through to Chartbeat

But the last point (point 4 above) doesn’t always happen. So then what?

If there is a time when we learn that personal information may have been passed to us from a website, we do the following:

  • Immediately get in touch with the client / owner of the site
  • Identify the location of the personal data (e.g., what’s the URL)
  • Advise the client on how to fix their code implementation issues in order to immediately stop the sending of personal data through to Chartbeat.
  • Determine the best way to purge our system database of this data and purge that data accordingly

It’s a quick, efficient, and effective reaction. But being reactive isn’t good enough. We’ve also got to be proactive.

In light of the concerns raised this week, we’re also performing ongoing audits of our entire network of thousands of client sites to see if we can identify instances of personal data to alert clients and ensure they update the data they pass to us immediately.

How we can all get better at taking care of our users’ data

The above is all specific to Chartbeat and we take the data we receive incredibly seriously. If I’ve gotten nothing else across at this point, I hope it’s that.

However, it’s important that we all, as website owners and data users, do our part to be better shepherds of data of all kinds—personal or not. A few ways to do so (and I’m positive you all have more suggestions, so please email me with them and I’ll update this post accordingly):

  • Never pass data in the URL itself. Because most analytics providers report at the URL level, the contents of the URL are likely to be stored by any analytics firm you work with.
  • If your web pages are served via HTTP then consider moving them to HTTPS. HTTP is insecure and data sent over HTTP could be read by a third party on the network.
  • Get to know your data partners. They are the experts in the data they collect, the way they collect it, how they store it, where they store it, and how it’s used. Ask them about it. Make sure you completely understand the terms of use, privacy policy, and any contractual language before putting their code on your site. And when you’re at the point of implementation, make sure you’ve checked with them to ensure you’ve done so properly. If you have any questions about Chartbeat implementation, our Chartcorps team is your go-to.

Chartbeat is in the business of building a better internet. As much as that means making sure the best content gets the most attention, it also means making sure we all, as users, fully understand the data that’s powering the web. We’ll do our very best to continue to be transparent about what we measure, how we measure it, and what that means for you.

Facebook announced today that is is making some changes to its News Feed algorithm to combat clickbait. Primarily, the social network will be looking at how much time people spend reading away from Facebook.

“If they click through to a link and then come straight back to Facebook, it suggests that they didn’t find something that they wanted. With this update we will start taking into account whether people tend to spend time away from Facebook after clicking a link, or whether they tend to come straight back to News Feed when we rank stories with links in them.”

Focusing on attention and time is nothing new for Facebook. On its last earnings call, Facebook specifically spoke about the size of their market opportunity in terms of the available time and attention they were able to accrue. On a more practical note, Facebook has been factoring how much time people spend away from Facebook after clicking on an ad into its pricing algorithm for some time now. In some ways, the news today is simply a wider application of that action.

Second, the decision to enable greater previewing of links, effectively giving the visitor more information to decide whether the content is interesting to them, potentially confirms a theory that Chartbeat’s data science team has held. On average, traffic from Facebook spends about 60% more time reading than traffic from Twitter. While there are likely a number of factors in this, the more sophisticated previewing in Facebook is a clear differentiator that we think affects this.

Take together these two actions confirm that Facebook is taking its users’ experience incredibly seriously and are leaning more and more on the fundamental concepts of the Attention Web to do so. That’s good news for quality publishers everywhere.

But what does this mean for great short-form content? The one potential challenge to this was raised by Matt Galligan of the excellent news service Circa:

It’s utterly logical to be concerned that content designed for brevity would suffer under this algorithm. However, I think this underestimates the comparative wealth of attention that even content designed to be brief gets. The depressing truth of the Internet is that short-form content hangs out on the same end of the distribution curve of the Internet as long form when it comes to attention.

As I’ve mentioned elsewhere, the majority of pageviews on the internet get fewer than 15 seconds of engagement. Facebook is looking for those incidences when people come ‘straight back’ to the feed, suggesting that the threshold they’ve set for clickbait may be rather low. If your content matches the intent of your headline (ie. you’re selling what you’re promising), then you’re highly likely to beat Facebook’s threshold even with short form.

Bottom line: Focus on creating quality content, match it with an accurate headline, and you’ll be fine.