Forbes Media has shared a number of interesting statistics this year in relation to its online content, from the number of Forbes posts being read on average each second – said to be 10,000 – to the fact that half the site's traffic each month is for older content, of 30 days upwards.
So how is the media outlet measuring all this data, and more? The answer is a combination of both standalone third-party tools, including Chartbeat and Adobe-owned Omniture, and its own handmade analytics platform, or "statistics engine", as Lewis D'Vorkin, chief product officer, called it when talking to Journalism.co.uk for a podcast about building the system.
It was three years ago when Forbes Media first introduced its own analytics platform. This powers, among other things, a public page view counter on online articles which is refreshed regularly, but at the beginning, this was not something everyone was happy about.
"When we started to do this, staffers would come into my office and ask me to stop counting the data in public because they were a bit embarrassed that maybe a post wasn't viewed that many times," D'Vorkin told Journalism.co.uk.
"And I said no, we're going to continue it, because you need to know how you're doing and it's transparent for the public as well. That's very important."
The unique challenge
The media outlet made the decision to build its own platform in an effort to support the production and publishing of content by its army of around 1,200 contributors, in addition to Forbes Media's own journalists.
"Every contributor should know, and every staffer should know, how am I doing? I publish a post, how's it doing? Where's the traffic coming from? Do people care? Are they interested? How does it compare with the other posts that I've done? So we really strongly believe that data is a feedback loop that helps power what we do.We really strongly believe that data is a feedback loop that helps power what we doLewis D'Vorkin, Forbes Media
"And I want to be really clear about this. The data informs what we do, it doesn't rule what we do."
The system makes it possible for the company to measure not only the performance on a contributor-by-contributor level, but also for the site as a whole – which again highlighted the need for Forbes to customise its own platform.
"If I make a post, and you visit me, and someone else publishes a post, and you visit them, that needs to count one time for you and one time for me, and not twice for the company," he explained. "It's one unique visitor that happened to go see two different people. So the ability to get that down cold for 1,000 folks is not easy".
But while there were reservations from some at the start, they soon became bitten by the data bug too. Six months in, when the increasing demand on the system led to it being re-constructed "almost from scratch", the response from those writers who had earlier been unsure about sharing the data with the audience performed an apparent U-turn in attitude.
"When we had trouble, it would show zero page views on the page," D'Vorkin said, "and when the statistics engine would break, all publishing stopped."
"The feedback loop had become so important to the people who were complaining about it, that they didn't want to publish unless they knew how they were doing.
"That's one of the reasons why we went through an intense effort to rebuild the statistics engine, because if people wouldn't publish, then [there were] no pages to put ads on, to make revenue, to keep the lights on. So the feedback loop is now integral to how people think about what they do.
"In other words, they love it."
Today, the system looks at a number of areas, recording data such as page views and unique visitors, as well as what D'Vorkin highlighted as a particularly valuable metric: repeat visitors.
"If someone comes to me once a month, that's a good thing," he said, "if someone comes to me twice a month that's a better thing, if someone visits me three times a month, that's even better, because that's a measure of loyalty. And we really track loyalty, that's very important to us."
The results are delivered to individual writers, who are provided with a personal "dashboard", he explained, as well as brought together as one to give the media outlet perspective across the site, or by specific sections. D'Vorkin reflected on the importance of the interfaces of these systems, and the need for data to be presented in a way that is "attractive, pretty and easy to consume".
"Engineers are not necessarily used to that, they build great engines, but the display of it is not their forte, so we actually have designers who work on how the data is displayed to our contributors, as well as how it's displayed to the editors who are digging even deeper into that data than the contributors can."
Linking performance with payment
The need to build their own platform to measure the consumption of content was also required in light of the payment system Forbes Media has established for its contributors, which is directly tied in to how well their content performs. And so the two systems – the 'statistics engine' and payment system – are linked together.
"That statistics engine will send that information about the size of your audience to a payments engine that knows the deal that you have with us, and that deal then calculates the money and then boom, the money will be paid to you," D'Vorkin explained.
"So it's an end-to-end system that we're building as well here."
And he is confident that with this system – paying contributors based on how much engagement their content invites – quality will not suffer.
"If we bring a contributor on to cover the computer hardware business, that's what they cover and that's all they can cover. So they have to figure out within that niche what is the content that people want..."
"If someone is covering the hardware business, and they're not being successful at that, they then can't write about Kim Kardashian to generate an audience, they can't do that. They have to figure out the specific vertical, niche that they're in. That's what prevents people from doing the things to generate traffic just to make money, because they must stay within their niche."
And for Forbes Media, the data is unlocking interesting insights about what content works best.
"The content that really really works is content that is very authentic, and that is very real for the person writing it, and thus real to the audience that's getting it. It's authentic.
"The content that works doesn't feel heavily processed, heavily edited, it feels like someone actually talking to the consumer, and that's very clear from all the statistics that we gather as well as the social sharing of things."
Was the most shared content on LinkedIn
D'Vorkin highlighted, that Forbes had the most shared content on LinkedIn which was based on a study by NewsWhip, a platform that measures the social traction of stories.
"That goes to the authentic quality of it and the data just reinforces all that," D'Vorkin added.
Free daily newsletter
- Tip: Use up-to-date and reliable covid-19 data
- Transphobia Project uses data visualisation to zoom in on outlets that spread biased transgender content
- IDA, Microsoft’s new AI-powered tool, helps journalists collaborate on large datasets
- How to find under-reported topics that readers actually pay attention to
- What skills will journalists need in 2020?