No doubt about data at NYT and VG as big tech hovers

Jul 01, 2021 at 11:02 am by admin


Confronted with best-in-class cases from global media companies, Pradyuman Maheshwari simply wanted to know the recipe for the “secret sauce”.

There had been a few obvious answers to similar questions of sponsor speakers such as those of Piano and Google, but it was the altruism of two New York Times speakers that really sent delegates off on a high note on the final day of WAN-Ifra’s Indian Media Leaders eSummit.

Maheshwari – founder and editor-in-chief MxMIndia – had already asked Ola Stenberg of Norway’s VG about the “secret sauce”; from the New York Times’ Mariam Melikadze and Pranay Prabhat, he simply wanted to know what of its custom-built cookie solutions was for sale.

The answer of course, is that – unlike Jeff Bezos’ Washington Post or Canada’s Globe & Mail – theirs is not, but as for sharing… “that’s why we’re here”.

The familiar problem of relating the experience of developed-country publishers such as Verdens Gang and the NYT, with that of Indian media – where print, at least for the moment, remains strong – was also eased by the fact that Prabhat still has family in Delhi. “I’ll ask them,” he said of one question.

 

Chief product officer of Norway’s largest news site Ola Stenberg explained to moderator Maheshwari and delegates, the publisher’s process-driven approach to product thinking and the mindset that puts audience first in shaping digital innovation.

Founded in 1945, VG has gone from top newspaper to top news website, now with three million unique users, 2.2 million of whom are on mobile, figures that put it in contention with Facebook and Google.

A big challenge however, is demography, Stenberg admits, and with the age of the average user having moved from 30 in 2006, to 47, “we’re getting worse”.

VG asks itself, “how good are we at this,” he says, outlining the constant questions, defining problems and hypotheses, and focus on putting the user first.

“We talk about outcome, not output – it’s easy to order features, but formulating it as a problem is harder for some reason,” he says, adding that the “holy trinity” must also include feasibility and viability.

Mission teams – with an ‘owner’, two or three developers, one each from UX and data, and others from editorial, research, CX and business – have to answer why the problem is critical and how solving it will contribute… and if there’s no good answers in three weeks, there’ll be one more of “detox” before proceeding.

Innovations don’t always follow the same model: A COVID-19 app was a fast job, while a storytelling project was more relaxed. And “even though we allow this kind of chaos, we have to agree on some things, with the most important being mindset”.

Chair Pradyuman Maheshwari and Ola Stenberg of VG

 

Sharing what they learned – “because learning is a team sport” – involves everyone from the chief executive down in a weekly quarter-of-an-hour of shared time, ‘10.45’ which used to bring together “pretty much everyone who’s not at another meeting”, perhaps 50-100 in the office, and now continues virtually.

How do you compare that with the situation and needs in India? Stenberg says it doesn’t matter: “It’s for everyone, wherever in the world, putting the user first and understanding the job to be done for the user, wherever in the world. Logistics may make some big differences, but really, you’ve got to try that everywhere.”

 

The same atmosphere of denial followed New York Times’ data insights director Mariam Melikadze and engineering director Pranay Prabhat: Would delegates face the same problems in India; what about Tier-2 and Tier-3 cities where print remains strong? “I think it’s a matter of time,” says Prabhat, though admitting they had been pleasantly surprised at print’s resilience. “There was an anticipation that it would fall, and it didn’t fall. We do have readers who want to hold the paper.”

On the face of it, the NYT duo were to have talked about how they planned to cope with the now-postponed “cookie apocalypse” threatened by Facebook. And though circumstances may have changed, the story hadn’t.

Prabhat – who still calls India “my home country” – gave delegates a look under the hood of the moves made to deal with the shut-down of third-party cookies, explaining how cross-context identifiers had become fundamental to behavioural targeting, retargeting, frequency capping, conversion and attribution.

New York Times’ Pranay Prabhat and an expressive Mariam Melikadze

 

The New York Times had started building its own audience data platform two years ago, and now uses only its own first-party data with behavioural and survey data fed into multiple ML models.

Users are put into first-party segments and activated into the ad server using inhouse-built systems and best-in-class tools for audience segmentation. As the “ad tech storm approaches”, he sees multiple solutions emerging including hashed emails, while some plan to banish user IDs, with transactions through cohorts as “the new currency”.

“We need to understand our strengths, and how to build a better engagement,” he said.

Mariam Melikadze told how the Times had grown its subscriber base over ten years, collecting “all this data” but had always been very privacy aware. Then a project in 2018 went “back to basics”, using an ML model to identify how an article made a user feel – informed, inspired, hopeful, adventurous perhaps, or even neutral emotions such as ‘not-fearful’. With more than 100 “strong clusters of topics”, reactions were classified into 18 emotions and seven motivations, giving a breadth of ways an advertiser could plan their campaign.

“We started to think a lot about the audience, and now have 50 demographic segments and a lot of interest ones,” she says.

A cross-organisational taskforce includes more than 20 stakeholders and experts, covering areas such as data science, analytics, audience insights, legal and data governance. Launching the product – with “lookalikes and click-alikes” has shown the first-party data performed “at least as well” as third-party cookies, 0.33 against 0.30.

Delegates wanted to know how the NYT was “grappling with the fact all the money is being taken by Google and Facebook” with Melikadze explaining that everything was about direct relationships. “It’s getting harder by the day, but that’s why we build these products.”

And back to the question of what Indian publishers should be doing while digital is “not getting the same kind of revenues” and money continues to go to the legacy business.

“It’s a classic problem,” she said, “but digital wins, and those who can’t pivot to it will perish. It’s just a question of how fast, which will vary across countries. It’s really important to figure it out.”

She forecast that there would be “drastically different forms of digital that don’t look like anything before” – different kinds of storytelling, data that “won’t look like anything we’ve seen” – but a huge opportunity that people shouldn’t shy away from.

I found theirs the most constructive element of the three-day conference, and with Google – whose south APAC head of news and publishing partnerships Shilpa Jhunjhunwala had a platform earlier to talk “the role tech companies play to enable the process and support quality journalism” – hovering above their markets, those still wondering about the importance of digital should have been left in no doubt.

Peter Coleman

Tags: cookies data IMLeS NYT product VG
Sections: Digital business