A trade journal of a still-emerging field, written by Adam Tinworth.

Posts tagged metrics


I’ve just finished* running a workshop on analytics for journalists at news:rewired this afternoon.

Here’s a selection of links I promised the attendees to allow them to explore some of the issues contained in the presentation in more detail:

And let this fine five minute rant from The Guardian‘s Chris Moran be a lesson to you all:

The Presentation

*Actually, a lie – I wrote this at midnight the night before, and scheduled the post…

Martin Belam on the reasons for his latest post:

And having just sat through an event where one of the questions was a worry that knowing something about SEO or writing for social risks “losing the craft” of journalism, I thought it was worth drawing attention to Richard’s words.

That idea of “losing the craft” of journalism is just fascinating to me, because it’s something I encounter so often. The root of it is about the complete absence of any detailed feedback on the reaction to a single story published in print. We never truly knew who read what (and believe me, wether they mean to or not, people lie and lie and lie in market research). Thus, we’ve come to rely on our instincts for our sense of news.

The problem with that is journalism is not art for art’s sake – it’s a public service. It exists to inform people about… stuff.

(I wanted to use a more precise word there, but given how BIG journalism is, and how much it covers, “stuff” seems more appropriate.)

We now have a feedback mechanism to explore how our work is being received, thanks to the sort of analytics we can get online. And that means we can improve that work in pretty close to real time so it finds a larger audience. That’s not losing the craft of journalism, that’s improving it.

Why be a crafter when you can be a master crafter?

Journalism requires an audience. If you have a good story, using the sort of techniques Richard Beech talks about help it find an audience. And if your story isn’t worth finding an audience for – why the bloody hell are you writing it?

The content we talk about least right now is content that gets a lot of reads, but no social media shares, Likes or retweets. Too many systems are set up to monitor these external marks of quality – from Facebook’s newsfeed algorithm, to Google’s search algorithm.

How do we identify and surface this sort of work, when people, for whatever reason, don’t want to share it or link to it, but do want to read it?

Moon rising

This is not news. The fact that people think it is is news:

The Guardian’s website is being swamped by unidentifiable “dark traffic”, and executives at the company cannot figure out where it is coming from.

“Dark traffic” reflects views on a website being driven from unknown sources, that can’t be picked up clearly from analytics packages or referrer logs. It’s pretty much a direct result of the shift to mobile consumption of news – and the dominance of apps in social sharing on mobile.

The Atlantic first identified “dark social” traffic back in 2012 to describe traffic coming messaging apps that had been stripped of referrer data because messaging and email use a less visible system than the that used by web pages.

Essentially, if you knew about the dual shift to mobile and apps, you knew this was coming. If you didn’t…

The core things to bear in mind:

  • It’s only going to get harder to tell where people are coming from
  • None of this stops you understanding where people go, and where they go after that

Image by ramson, and used under a Creative Commons licence

Another reminder of the dangers of building your business on one company’s platform:

Gawker web stats

Those are Gawker’s figures – and why did that valley happen?

Read and his co-workers think that this is the result of some algorithm changes Facebook made in May. But just like everyone else on the Web who doesn’t work at 1 Hacker Way, they can only guess at what those changes are. Even when Facebook announces it is making changes, it doesn’t do a whole lot to explain what it’s really doing and why.

Previously: Surviving the Facebook clickbait algorithm change

NYT homepage traffic

Talking of homepage traffic, Zachary Seward has actually looked into the NYT homepage traffic for Quartz:

Traffic to the New York Times homepage fell by half in the last two years, according to the newspaper’s internal review of its digital strategy.

Is that indicative of an overall fall?

Overall traffic to the Times isn’t falling; it’s just coming in through the “side door” more often.

So, no, then.

As the report itself says:

Traffic to the home page has been declining, month after month, for years. Traffic to section fronts is negligible.

Yet, how much time and effort is going into producing that homepage and those section fronts? An amount massively out of proportion to the traffic, I’d suggest.

This is absolutely symptomatic of the problems with traditional publishers moving online. They bring the habits of print with them, they don’t use metrics intelligently to understand how traffic actually moves across and around the site, and they end up wasting time and effort in pointless pursuits.

We cannot afford this moronic head-in-the-sand attitude to how our readers actually use our sites.

It’s that time of the year when some of the social services you use start sending you stats. The one I got from Pocket – a “save and read it later” service – was actually quite eye-opening…

Pocket consumption stats852,713 words! That’s a fair few books right there – and that’s not counting the material I consumed staright away in my feed reader and never got as far as Pocket-ing. And, indeed, all the unread items that are still in my Pocket…

You can see the stats in full, if you wish, but there’s not a lot more.

One of the reasons I’ve switched to Pocket from Instapaper is this stats element. They have a good stats service for publishers, for example, that gives you a sense of how often people are consuming your articles in their service. Here’s how it’s looking for me right now:

Pocket Publisher StatsAs people’s reading tools diversify, I suspect being able to move easily through a range of analytics tools to get a complete picture of how and where your content is being consumed will become a core skill for publishers and editors. 

A hotel bedroom desk in Berlin

There are many rewards to blogging, but it’s easy to get distracted by the easy one. Big numbers are a big distraction. My post on citizen journalism earlier in the week got big numbers. That obviously makes me happy. 

My post on Bruce Sterling and the video of his speech at NEXT Berlin? That didn’t get big numbers. In fact, calling the numbers “small” seems, well, generous. Yet, it provoked this post in response – well, not response exactly, but it acted as a catalyst for a thought process that became that post. 

That’s actually a bigger reward to me – I published something which someone I think is pretty smart used to write something pretty smart. I’m bringing insightful things to the notice of insightful people. That’s worthwhile.

I often preach the “hits stands for How Idiots Track Success” mantra – the interactive journalism MA students at City University can vouch for that – so I should remember to practice it, shouldn’t I?

After all, if I wanted the big numbers, this would be called something like “The Digital Journalism Expert” and have headlines that ran along the lines of 10 things digital journalism can learn from the Tumblr /Yahoo deal. But it’s not, and they don’t. You just have the meandering thoughts of one man on his blog, and just because my reputation built through this blog is the source of my income and my family’s financial security doesn’t mean I should let big numbers blind me to the truth: influence can be small, subtle and complicated, and often driven by small acts of generosity.

Find something cool. Pass it along. Change a couple of people’s thought processes. Who knows what might happen? 


Katy Howell, CEO, Immediate Future

How do you concentrate your content? It’s pretty clear we need to think about social content with purpose. Social media is a business channel – but there are some big players in this. 80% of business buyers network online for work, 91% research for work – and 70% are promoting themselves. We don’t talk about brands, we talk through them… 

Decision maker sand senior decisions makers are rarely lurkers – they’re an active group. 67.9% of B2B content marketing is targeted at lead generation. There are some challenges – the deluge of crap, it’s resource intensive and how do you know you’re capturing the quality of people you want?

If you’re  a planner, you can’t help but put things in boxes. It helps you get your head around them. How can you make sure your content works hard for clients? Well, you target. Everyone talks about being personal. How many people present are creating content for an individual? Very few. 

She’s so excited by Google planning tools. People don’t buy in a linear way, and B2B purchasing is not all digital yet (we need chips in people’s head (is that what Google Glass is?)) Thee tools don’t give you your journey – but they give you a benchmark. It helps you understand the journey. 

IBM makes millions from “snaffling it” – just looking for requests on social media and having conversations on the fly. They don’t just sell – they send them things like a best practice guide for writing RFPs – content marketing. It jumps outside the funnel. On the other hand 66% of buyers use LinkedIn for identifying buyers, but 55% used Twitter for identifying the final supplier. (Again, your customers might be different).

Think about behaviours – but also think what they do before and after. Where do they go from LinkedIn? How do you follow up a capture form? Should you integrate this with your wide campaigns? Search support is important – it’s not all about search, but you can’t miss a trick. And – in lead generation – you have to work with the people around you: sales, social media, the board, your content resources – you have to know the structure. You have worked hard, and paid for great content. Make sure the flow fits, because you can’t measure unless you get to the end goal. 

Bloody Strong Content Plans

Have a bloody strong content plan. It took us six weeks top plan customer magazines – it should be the same for social content. Think about publication date, search tersm, calls to action, trackable links, distribution and monitoring/KPIs.

BUT – content for lead generation has a different purpose. You need killer calls to action. There are calls to action on social media that you wouldn’t get away with on social media. You need to be disruptive to get noticed. How do you deal with conservative clients? Stealth and lying. You have to be a little on the edge – she doesn’t mind running around in her bra and knickers if it gets the job done. You can often find a change agent client-side. Euan Semple calls them trojan mice projects. Make sure the in-house person gets the glory. We live and breathe social media – our clients often don’t. 

It’s hard to find content which has a clear purpose like lead generation right now. It’s about small tweaks – knowing behaviours, clearer calls to action. 

Katy Howell at #b2bhuddle

There’s repercussions for poor content. We are seeing drops in the number of people downloading whitepapers, ebooks and webinars. 4.8% drop for whitepapers, 5.1% for ebooks and 21.6% for webinars. If a whitepaper is just an extended blog post, people get annoyed quickly. Spammy approaches to content have consequences – search engines won’t put up with it forever.

So, make things easy and clean, and use social proof. If it’s relevant, you’ll capture more of the people you want. Ensure links have your brand in them in some way – makes them more trusted. Keep capture forms short. Use all relevant advertising tools: Facebook Exchange (jury still out), SlideShare capture (part of Pro), Twitter lead gen (14% rate), LinkedIn ads (not great) InMails (amazing). 

People don’t sharpen their tools well. They don’t test and learn. Start with benchmarks. 46% of buyers find whitepapers useful for defining what they actually need. 44% find videos, webinars and podcasts for identifying suppliers. 51% use supplier e-mail for final selection. Document everything you do – short links, campaign codes and the like allow you to track effect.

You want to know performance by platform, but also by subsection of the platform. Not just LinkedIn, but groups, or news feeds, or company pages.  Details. Track the actions, like Likes, clicks and comments. How is each content toe performing. What times of day work best? What topics work best? How about headline experiments – one of their clients saw a 40% uplink in traffics by putting what they would get through the link in brackets. Bundle all of this into a performance framework. 

Test and dump. Rest things that aren’t working. Do A/B tests. Know what your hottest lead source is. 

Feed all of this information into the purchase loop. Know what content contributes best at what point in the purchase cycle. It’s hard. Katy is not there yet. But she’s working on it. 

Question & Answer

Are you doing analysis on costs per lead on social? Yes. Two clients have been doing cost per lead – it’s lower than DM, but higher than e-mail. And it wavers all over the place, because the content isn’t right yet. Month on month things change. The industry’s in churn, so it’s difficult to think about pay per lead.