Mobile navigation


Man vs Machine

Machines are great at measuring things, but are the things they measure a reliable guide to good journalism? Treating content like widgets might benefit the bottom line, short-term, but might well end up damaging publishing, long-term. As Peter Preston points out, people win Pulitzers, not machines.

By Peter Preston

It’s a sea change in the whole nature of journalism: a great shift from man to machine. Once upon a pre-digital time, editors relied on hunches when front page priorities had to be chosen. They thought they knew (from experience, from proprietorial instruction, from fragmentary market research) the stories their readers would respond to most passionately, pushing pence across a newsagent’s counter. But it still remained a passably human process. You could take a flyer on an otherwise obscure political tale and see what the reaction was. You could equally choose a lead you knew wouldn’t sell – Northern Irish threats, marches and violence, for instance – because it was important, because readers ought to be informed whether they snapped to attention or not.

And now? Forget hunches, at least of necessity. Forget guesswork and mystic claims to have found your audience’s wavelength. The essential routine in any major paper’s office arrives with assertive information attached. You write your story, file it and see it posted. Then data clicks in. How many people are reading those pearly words? In Britain? In America? Anywhere in the world? How long are they taking to scan its paragraphs? Are they engaged or indifferent? And see, theoretically, how the advertisers come rushing to sell on the back of such seeming certainties.

The New York Times is particularly proud of its new analytics tool called Stela - a clear relative of the Guardian’s Ophan - that collects data from many different sources and makes them journalist-friendly. On the July day five police officers were shot down in Dallas, the newsroom changed its headlines to reflect new facts or new angles more than a dozen times. “With each corresponding headline change, Times reporters and editors could monitor traffic shifts and reader response across social channels and comments on the Times page, through its new custom-built internal dashboard”, according to Nieman Lab. In a choice between man and machine, the machine seems like a winner. Click once for obedience.

Keeping tabs

Which in turn, of course, sets the whole business of journalism in a prospectively different context. If what journalists produce or process is governed by the new world of algorithm existence, so the people sitting at their desks must face new superintendence too. Or, perhaps, you may say, people who weren’t sitting at their desks at all. A few weeks ago, BuzzFeed unearthed a new Telegraph memo to staff. “Over the weekend, we have installed a number of under-desk sensors across some areas in advertising, editorial, technology, production, newspaper sales and marketing. They will be in place for a duration of four weeks.

“These devices are part of our drive to make our floors in the building as energy efficient as possible and reduce the amount of power we consume for heating, lighting and cooling the building at times of low usage. Accordingly, they are designed to record occupancy across each 24-hour cycle for all seven days of the week to make sure we are making best use of our space in the building.”

There was, inevitably, a bit of a newsroom ruckus and the sensors promptly disappeared. But the incident showed not only what’s available to managements who want to monitor, but the variety of explanations for such action currently on offer. Anxious to know what staff are doing on their office smartphones? Here - in Times and Sun territory - is a recent message to staff announcing that “all your traffic will pass through Wandera, whether you use the mobile for business or personal uses… the Wandera system (from an eponymous data security company) will record every website address you visit … and the volume of traffic to and from that website. Usage reports will be provided to News UK.”

Of course, there are safeguards for some editorial staff. Of course, limiting smartphone costs is a natural management concern. But reassurances don’t come easy in this kind of environment. Consider what journalists say about police and security agencies who want to track their web and phone usage via Theresa May’s “Snoopers Charter”, then spread such suspicion liberally around newsroom workspace. (Think of the email mountain News UK found in the hacking debacle.) Tracking devices exist and may be perfectly valid. But if the Home Office shouldn’t be allowed to trawl through your records to identify sources, are office managers back at the ranch any more welcome?

Commoditisation of news

The simple fact is that the technology that enables so much of today’s journalism can also, too easily, be seen as a threat to its survival. We’re used, at one level, to “feature farms” churning out copy fodder to fill space cheaply. The Guardian’s Nick Davies embellished the concept of “churnalism’ when he revealed how refettling PR hand-outs had become a staple of ordinary editorial life. But, continually, on every front, the need for machines and machine-honed processes seems on the march. The day of the news factory – in literal terms – is upon us, not just in the way that big newspaper groups have turned local publishing into distant publishing via consolidated sub-editorial hubs many hundreds of miles from the newsrooms whose stories they handle. And maybe the factory analogy doesn’t fit so neatly any longer. Think of news supermarkets instead: the same monitoring at the tills, the same rigid computer controls, the same emphasis on stock replenishment and display: plus calculation of sell-by dates too.

If (as calculated for the analytics operation Parsely) your average post on Facebook lasts for only 2.6 days, then news and feature consumption can become ravenous going on impossible. Call for more formula lists that can be wheeled out time and again, suitably touched up to chime with events. Call for more archive excavations with added paint jobs. Call, Digiday recently reported, for sites in the style whose CEO Neil Vogel says that its health vertical, VeryWell, has about 50,000 pieces of content, with an average age of 285 days, the most popular of which get the most updating. “When we write something, it has to be good for probably six months to three years.”

Swimming against such tides can seem antediluvian, or worse. We all know that journalism is under huge financial strain. We deal daily with a litany of cuts. We reckon that the technology that has undermined the old verities of print existence must somehow be a motor to provide new means of survival. Men and women cost money: computer and video kit, by contrast, equal investment. And, of course, there is a good helping of shrewd truth to much of this. Journalism has always depended on ploughing the profits it makes in good times into the ideas that will sustain it through lean years.

Where real value lies

The difficulty, though, is constructing any sort of balance from such opposites. Journalism has, somehow, to pay its way. That means assured revenue flows, keen costings, efficiencies. But journalism can also be the stuff of inspiration and innovation. No computer programme can run an investigation such as the Panama Papers: the data journalism there has to be governed by human impulses and curiosities. No word programme can make the great jokes from Craig Brown that adorn the Daily Mail or the wonderfully witty political sketches of John Crace in the Guardian. No robot editor can turn a sheaf of expense forms into the devastating parliamentary rip-offs that gave the Daily Telegraph investigative credentials again. Drones may fly in the skies over Syria, but they were, tragically, no match for Marie Colvin on the ground.

All manner of questions inevitably follow. Does a sub-editor who polishes one memorable Freddy Starr hamster for the front page equal two more average performers who never write a headline that sings? Is an interviewer who takes three weeks to do an in-depth job as useful as one who turns round 2,000 words in a day? How does productivity rate in a world where the Pulitzers go to long range inquiries?

The answer, of course, is that productivity alone isn’t nearly enough. Just look at Britain’s most successful editors – say Dacre of the Mail – and note how, through thick and thin, he keeps staff levels reasonable to good, unlike rivals who cut to the bone and complain thereafter that the meat has gone missing. You can’t do away with thinking time, digging time, ideas time. A local paper that doesn’t know its locality is a walking ghost. A newsroom ruled by the clicks regime will never do something different.

Too pat, too pious? Perhaps. I’ve struggled in the valley of debts and looming disaster long enough to understand how their imperatives drive. But reporting the news is one amongst dozens of creative industries. There has to be creative room to breathe, and finding it is one of the big challenges as the digital revolution rolls on - and, suddenly, we discover that machine world isn’t so full of certainties after all. Data means targeting. Data means knowledge. But data on Facebook and Google won’t pay the rent and keep your staff ticking. I seem to remember how long ago - well, maybe around ten years back - the data was all about integration, about multi-tasking on great newsroom hubs. The Telegraph led that way. And today? The hubs are broken; the spare parts scattered; the thinking caps on again. In a world without settled answers, the real engagement is obvious enough: not engaged algorithms but engaged, infinitely flexible brains.