After attending several days’ worth of media conferences recently a dramatic story has lodged in my head.
Two siblings are let down in love. One decides that the other sex cannot be trusted and tries to live as independently as possible. The other thinks that attitude will lead to a limited life. Instead sibling two leaves the door open; a move the first sibling thinks is just blind hope. Both believe their sibling is on the road to ruin.
The opposite sex is advertising. The sibling seeking independence is a newsbrand who puts up a full paywall. The sibling with the open door is a newsbrand that maintains an open newsbrand website. It isn’t set on Albert Square but on what was Fleet Street.
In King’s Cross; speakers at the DMS14 conference identified these five reasons for the unreliability of advertising;
- The cost of online advertising has greatly declined because of the oversupply of inventory. Advertising cost per thousands, whilst high for press, are lowest (and very low) for digital advertising. Douglas McCabe of Enders pointed out that most national newsbrands are taking less than 10% of their total revenues (circulation plus advertising) from the digital side of the business (the exceptions being The FT on 45%, The Guardian on 32% and The Independent on 24%).
- Programmatic buying is commoditising advertising, driving down prices more. Stephane Pere of the Economist pointed to Magma Global’s prediction that only 17% of US display advertising would be non-programmatically bought by 2017.
- Traditional media is being forced to go head to head with global titans like facebook and Google who have no content costs and lots of data
- An overreliance on advertising in this new world means overexposure to the cruel whims of economic cycles. Some commentatorsalready warn that the British recovery might be built on sand.
- Advertisers needn’t rely on paid media so much because they can increase their use of earned media. Social media gives them a chance to get their message under the noses of consumers as long as they are creative enough to survive in the social stream. Meanwhile traditional newsbrands are caught between helping advertisers work this way through their channels (dubbed native advertising) and maintaining the “church and state” divide of commercial and editorial.
For these reasons Mike Darcey, the CEO of News UK, and Top Right’s CEO Duncan Painter (both spoke at DMS14) feel it is prudent to reduce their reliance on advertising by trying to make viable businesses out of paid content. All else, they feel, means the abandonment of the quality journalism mission in order to chase scale. Alternatively it means taking a misguided gamble on the return of advertising pounds. For Darcey, even the metered paywall falls short.
Have they been able to reduce their reliance on advertising? Painter has culled 300 Top Right events because they over-relied on advertisers rather than paying customers. News UK’s Katie Vanneck Smith (speaking at a YouGov/Press Club event) said Times Newspapers Ltd was making more money after just fifteen months of the dual revenue stream (advertising and digital subscriptions) than it was before the paywall. That happened despite the audience being a fraction of the size. She says the balance of consumer revenue to advertising revenue at TNL is now 70:30; less advertising heavy than the traditional quality newspaper ratio of 55:45.
Sibling Two disagrees. “The paywall horse has bolted”, said Andrew Miller, CEO of The Guardian (also speaking at DMS14), and “’Open’ is how The Internet works”. His deputy, David Pemsel (at Changing Media Summit 2014) said “we do believe that if we had not been open, if we did not have a brand that was not influential or indeed could garner response and create comment, people like Edward Snowden would not be coming to us, they would go somewhere else.” As a result of that story, described as “the biggest global story of the year”, The Guardian’s editor in chief was awarded “The Special Award” at the European Press awards.
What newsbrand CEOs generally agree on is that data is good. However, their different approaches (paywall vs Meter vs Open) mean they are approaching data collection differently.
Miller said he felt News UK was cutting themselves off from the window shoppers (including those outside the UK) who could be enticed to give their data. The Guardian, he said, was happy to entertain unknown audiences because they could be driven towards becoming ‘known’. They will launch a membership scheme later this year.
Darcey wants to know all of News UK’s digital customers. His example was that if somebody attended a Times+ event he knew they were very likely to renew their subscription. Katie Vanneck Smith’s example was that she knew exactly how many subscribers the columnist Caitlin Moran brought in. Similarly John Ridding, FT CEO, said his data team had identified the most likely FT pages that people looked at just before subscribing. These are actionable, behavioural insights derived from monitoring audiences at scale.
In these tales data is big and comprehensive. All knowing.
What is interesting is that, under these CEOs, the data directors within newsbrands now feel that they need to supplement what they know from (big) data. They want to understand why people do what they do. They want to know about cause and effect. Big Data begs specific questions. Do Times+ events convince people to re-subscribe or do they just attract those most likely to re-subscribe in the first place? If the FT knows the pages last seen before subscription, what are the earlier stages of the purchase funnel?
On the day before DMS14 there was a ‘pre’ day called “This is Product Development”. Tom Betts, Vice President Customer Analytics and Research at The FT and Anthony Sullivan, Director of Product at The Guardian both spoke about their increasing need to use conventional research methodologies like surveys and focus groups.
In Sullivan’s conclusions he made these three points:
- There should be understanding from multiple directions
- All methods should be respected and understood
- There is no substitute for talking to audiences (their investment in a sizeable lab for doing just that is the proof in that pudding).
They were wary of conventional research. Tom Betts spoke about discovering the usefulness of the type of research he always thought just “gathered dust on a shelf”. Anthony Sullivan showed a slide listing five approaches his team were taking. These included quantitative and qualitative research and both were accompanied by asterisks that warned that these methodologies were dangerous if you don’t know what you’re doing.
When I pointed out this wariness on Twitter Tom Betts replied saying he was a “former qual and quant cynic, now converted. Behavioural analytics only tells part of the story” and “just the latest evolution like when digital analytics met database marketing”.
Is this evolution happening because traditional research methods are the natural next port of call for Big Data practicioners – or because there are unexpected holes that need patching?
Peter Houston wrote an interesting piece for The Media Briefing on the gaps in big data. In it he pointed out that “publishers generally don’t have anything near what could be called big data. They have small data; multiple sets of small data”. He added, “What we’re really seeing is an end to technology fuelled ‘my data’s bigger than your data’ bravado and re-focussing of attention back on the value that data brings”.
Expectations of data are high in publishing businesses but their data teams are dealing with some limitations. A lot of impressive things will be able to be done because of big data (I watch with interest for news of the work of Chris Wiggins, the scientist hired as a Chief Data Scientist by The New York Times). Probably great things be done in directions not yet imagined whilst some current ambitions might have to be revised. One arena where much is expected quickly of data is the commercial floor where the lower CPTs of digital advertising are felt most keenly. Data has to perform for the business there.
Whilst ad directors might find the company data sets are less complete than they would like, those they would woo are questionning practicalities.
In GroupM’s Interaction 2014 report it says that, for advertisers big data promises “super-precision” but adds, “it is however, less clear that precision and efficiency are the same thing. This is likely to be the dominant theme of [their 2015 report]”. “It is for now at least possible that the dividend of reduced wastage is being matched or offset by by the costs of customization and personalization”.
Amidst all this we will probably see different news companies adopt different narratives about data. It is very possible they will reach different conclusions about how it is best used internally and how it is best used by commercial partners. Some might make claims that their rivals question. When they stand on conference stages will their approaches be so different that they conflict with each other? Will we be treated to more squabbling siblings? Will newsbrands send in the CEOs again? That’s what I’m expecting in 2015 and I’m booking my conference tickets already.
Disrupted industries at least make good theatre.
In the meantime, to keep an eye on the way the wind blows, I’m off to the Harnessing the Power of Big Data for Media conference run by the Reuters Institute in May. More on that after the event.