CT No. 22: Traffic metrics hurt your business. Measure this instead.

I'm not into the holiday spirit right now. It's just not one of those years.

You’re hella busy this week trying to get everything done before Friday. I get it. Feel free to put this in your pocket for later.

If you’re new here and haven’t yet, you can

Pageviews suck. Sessions too. Let’s not measure traffic anymore!

I've gathered several blog posts from various years that talk about how pageviews are straight-up not a good measurement of content success. Yet I still see pageviews as core metrics on marketing agency, publisher and multibillion dollar business content dashboards. Businesses make terrible decisions based on pageviews and basic traffic metrics. All the time. In 2019!

Because pageviews are easiest to pull from Google Analytics. No single metric is ever going to fully describe content impact on business performance, but no matter what, that metric is certainly not pageviews. The amount of times I’ve used the word “pageviews” in this paragraph? That's how tired I am of seeing pageviews as a metric in reports.

I mean, at least use unique pageviews if you’re gonna be that basic. Like, even 2009 Gawker figured that shit out.

Pageviews are popular because you can game them easily. Because you don’t need to think to understand “more.” Pageviews inspire business leaders to throw babies out with the bathwater, to think that content is interchangeable, to think that they can game the system.

You can rejigger content design so page numbers go up. And up and up. They go up with advertising and email blasts. They go up when you badly reorganize your content and your users need to click around to twenty pages to find what they’re looking for.

Our culture rewards only numbers that go up and to the right. Our culture rewards the simplicity of pageviews. It’s easy to sell the 20th century scourge that is impression-based advertising against pageviews.

Your content deserves better than pageviews.

The Google Analytics traffic metric Sessions improves on pageviews. Sessions measure individual visits to a website, so one user’s multiple actions on one website visit are just counted as one session.

But even sessions are kinda garbage, mostly because they only answer the question of “how much?” and the only answer is “more.” Sessions are a traffic metric that serves a mass communications model of reaching the most, as many as possible. More more more.

Now, you need a significant amount of traffic — let’s say 100 daily sessions — to make the next model work. But once you hit that significance, stop focusing on growing pure traffic because otherwise you will forget about the actual people reading your content. Focus instead on growing engagement.

I bet you think I’m going to tell you that you need conversion actions! Newsletter signups and assisted conversions and product purchases and paying subscriptions! Heck yeah, you need those! But if you go after those too aggressively you’ll alienate your users before you’ve even made a connection. (What’s up, every Conde Nast publication that serves me an email signup popup when I’ve literally clicked to your website directly from your email? Yeah I don’t visit your websites as frequently anymore. That’s right.)

We’re going to look at what you should measure between when a user arrives at your site and isn’t quite ready to convert yet. Measuring the slow burn of engagement matters for content marketing and publishers alike. So let’s get engaged!

How do you measure content based on how people read?

Any content creator worth their salt wants to know how many readers give a shit. How many readers said “Awww yeah, I'll read something like this again”? Or “I don't wanna close this tab.” Or “I actually learned something from that.”

Many have tried to measure this before. Scroll depth was one of those failed metrics that everyone implemented in the mid-2010s because the theory was that if you scrolled to the bottom of the page you were engaged. Now it just means you’ve scrolled past all the excessive ad units. Tracking scroll depth most often means that you fucked up your bounce rate benchmark for years because it counted any scrolled content as not-a-bounce. Bounce rate is a useful metric if you don’t fuck it up! (If your bounce rate is 8%, please change your settings so scrolling to the bottom of a page doesn't count as a second pageview. Love, your content analyst.)

You can try surveying your users to understand their engagement, but only a very small percentage of very engaged users will complete your survey. Looking at you, everyone who read the email but did not complete the most recent one-question Content Technologist survey.

You certainly don’t want to creep your users out by using personal data to track their every move. That's gauche; I’m not even going to talk about everything wrong with that.

You want to know: “Is this content viable?” So.

I wish I could tell you it’s just a 1:1 replacement for sessions and that you could easily find this measurement in Google Analytics. It’s not. (And Google Analytics is designed to help marketers sell widgets, not brain space, but we'll get to that in another issue.)

Remember that goldfish attention-span metric of 8 seconds that everyone quotes as a measure of our “distraction”? Digital readers are distracted, sure, but we're people, not goldfish. As much as you might say “people are stupid” if you get a result that you dislike, people are discerning.

We have filters. Most of us who read online have good digital filters, read in F-patterns, decide quickly whether your content is worth our time. We’re not going to read past the first subheading if we know the content is basic or offers zero original insight. Digital readers are going to pause on any kind of movement or complex visuals like gifs, data visualization or video.

We want to ascertain how many people skim your page but then decide to read at least some portion of that content. Did your content survive their filter?

Surviving the filter

The average reading time for an English-speaking US adult is 265 words per minute (wpm), with college students averaging 300 wpm. Speed readers read 700 wpm. So what's an average skim time? I bet I can find it somewhere in mass comm research but this newsletter needs to go out tomorrow, so I’m going to say: 600 words per minute.

Based on your word count, how long should a reader spend skimming on a page? I’d max out at 90 seconds for really long articles on this one, but for all intents and purposes let’s say: anything above one minute counts as an engagement. If you’re on the page after one minute, you give a shit.*

But that metric alone doesn’t count for anything. So I’m only going to look at number of users who have spent above one minute on any single text-only page.* Pages with complex images like infographics, gifs and videos will likely attract users for longer but not much.

So. Let’s calculate the number of users who deemed our content worthy of passing their filter. They’re the filter survivors.

But I’d go one further on the filter survival metric: look at returning users who spend more than one minute on any content-driven page (i.e., not your homepage unless you put significant content on your homepage). Those are people who have been to your publication before and still find your content valuable.* Likely they’ll be here again: once a month, once a week...

Even better than that, try calculating the percentage of users who remain on the page after filter survival, versus your total number of on-page users. That’s your filter survival rate.

Bonus points if they don’t bounce. Bonus points for returning users whose session duration is above five minutes. Bonus points if you've figured out a better metric and want to tell me about it.

Mostly the above has been an exercise. I’m sure far more fluent data scientists than I — not an actual data scientist, just a trained editor who looks at a lot of charts — have better thoughts on how to measure engaged content. But once you’ve spent time seriously considering how an engaged user might behave, you’ll have a better understanding of what to measure.

Look at growing your filter survivors. They give a shit. Look at maintaining your survival rates while growing your base number of users through organic search, referral traffic, word of mouth, etc.

Users who have survived the filter — win them over! Don’t enable dark-patterns or unnecessary pop-ups or spam them. Win them over, gradually, over time, with delicious and unstoppable content.

Even Freddie Quell likes delicious, unstoppable content. He’s a survivor. (J/k have you seen The Master? This is the one pleasant thing Freddie Quell ever does. Also, The Master is my favorite movie of the 2010s and this gif has been sitting on my desktop for a month, waiting to make its debut.)

*This discussion is Google Analytics-specific. GA’s an industry standard and free. However, you can also invest in measuring content a javascript heartbeat like Parsely offers.

Content measurement for committed publishers: Parsely review

Parsely rose out of publishers’ dissatisfaction with simple pageviews and some publisher-specific limitations in Google Analytics. It’s not as intense as Chartbeat, the more-traffic-now platform that terrified every working blogger I know. (Even a massive enterprise content marketing program I’ve consulted with said Chartbeat data was “too much.”) Parsely has always billed itself as the Thinking Publisher’s Content Measurement platform.

Parsely tries to distill the very real problem of Google Analytics confusion into a few basic metrics. Their sales presentations insist that more editors and writers will actually look at Parsely data because it’s easier to understand, and more editors looking at data is a good thing (very true)!

But is Parsely data actually better for content measurement?

Parsely at a glance

Parsely’s out-of-the-box setup measures two dimensions that require significant custom configuration in Google Analytics: engaged time on page and segmentation of most recent content published.

As discussed above, the tool’s engaged time on page is their signature metric. Especially in these days of tiny mobile screens, users move a lot when we consume content. We move our mouse around, scroll through content, put our fingers on the screen. Parsely’s tracker uses these tiny movements to indicate “heartbeat” or engaged time on page. It’s a great approach to content measurement, especially when trying to measure active readers who often have 20 tabs open at once and only spend 5 minutes actually reading them!

The other giant strength Parsely offers is segmentation of all of your most recent stories. In Google Analytics, unless you have dates in your URLs (gauche!), it’s super difficult to segment most recent content published and requires a custom setup with your CMS.

But in publishing, sometimes you just want recent and not evergreen content! You’ve published 50 articles this week and — even though that piece from a month ago is going strong — sometimes you really only want to look at the aggregate of content published this week. Parsely allows you to segment recent content metrics easily and provides explanation of those metrics in the interface. It’s rad.

Users who are concerned about real-time performance can also look at what’s happening in the past few seconds… but unless you’re breaking news, I don’t recommend obsessing over that metric. Even if you’re breaking news, I don’t recommend obsessing over that metric. It’s a distraction (and the obsession with immediate returns and growth is damaging democracy, but that’s another story)!

Can Parsely replace Google Analytics? No. I can understand way more about how users perceive content with GA. GA is far easier to filter and customize. You can teach content measurement with GA. GA provides all of the elements but forces you to understand why you’re using each one. It was clearly created by and for engineers.

Parsely simplifies content measurement so users can understand performance without knowing what/why is being measured. It helps content creators understand measurement easily so they can go about their days. (That said, editors and journalists are brainy people and are perfectly able to understand content measurement, even if business misunderstanding of content performance and metrics like pageviews has historically made them suspicious of all traffic measurements.)

Parsely setup requires a developer. The tool only recently started offering conversion tracking, and that’s not an easy setup. Parsely offers an API but in these days of easy connections, it doesn’t really play well with other tools without a developer’s support.

But can Parsely supplement Google Analytics? Absolutely. 100%. I would recommend it for any publisher or content program that is regularly publishing multiple pieces of good content and cares about reader engagement and editorial’s engagement with reader data.

Another plug: Parsely offers amazing resources, including a blog and their very popular Currents tool, which is free! And which will be reviewed in this newsletter sometime in the future.

Content tech news of the week

It’s all in moderation. Seriously, mostly I read about content moderation this week and how Facebook and YouTube outsource their moderation instead of taking responsibility. It’s a thing! Anyway

  • Here’s why Patreon is good at creating and enforcing user guidelines for content creators earning a living with their site. From the very amazing Nandini Jammi, who I just became aware of this week and am now worshipping.

  • The Terror Queue on The Verge is the latest in investigative pieces that examine how tech companies siphon off responsibility to giant consulting firms who hire poorly paid contractors to do the very hard work of moderation. So basically, Google and Facebook don’t pay people fairly to do some of the most emotionally taxing jobs at their company. But don’t pretend this is a new thing — publishers have been doing this for years by underpaying journalists and fact-checkers who are “paying their dues.” Oh, what a world!

  • And Facebook is still prioritizing scale over safety, from the cool kids over at Buzzfeed. In personal news, I finally quit Facebook this week, although I will probably log in one more time to download all of my friends’ birthdays. But 2019 was the year that Facebook quickly turned from an annoying but helpful tool into a right-wing garbage fire.

Did you like this issue? Feel free to press the Heart button and then


Visit The Content TechnologistAboutEthicsFeatures LegendPricing Legend.