Parallel sessions were the order of Publish Asia's second day, but one everyone wanted to hear was that about changes at the South China Morning Post

Having worked at the Hong Kong-based English-language publisher since 2011, chief operating officer Elsie Cheung was well placed to explain the changes since it was acquired by Alibaba owner Jack Ma. And to redress the balance after the conference's first day in which all of the speakers had been men.

"We want to become a global media and lead the conversation on China," she said. As part of the new focus, attention goes beyond Hong Kong to Asia and the US, with new products being created for these audiences.

A new workplace culture was being driven by new offices, laptops "for everyone", and limited desk space in an activity-based workplace. Varied seating options included quiet collaborative spaces, with a community initiative called 'SCMP Serve' and an 'SCMP Fun' social club. "People envy us and you can feel the energy," she said.

On the digital front, a big change was in the newsroom, and a 30-strong a dedicated print team had also been set up. "Most editorial staff have to think about repackaging for different audiences," she said There were also product teams focussed on areas such as user experience and audience behaviour. More on the SCMP transformation to follow.

Earlier, ICFJ Knight fellow Nasr ul Hadi had reported that tech experts "remain sparse" in newsrooms in a recent survey - with only five per cent having tech degrees - and hybrid and digital-only newsrooms outpacing traditional ones. Newsrooms relied on limited digital skills and less than a third of journalists in south Asia received on the job training. And he said the training they wanted was "not what management think they need".

The survey had also found that while journalists used social media to discover and share news, only 11 per cent used its verification tools.

From Taiwan's United Daily News, data development general manager Anson Mok told of the publisher's work on priorities of user experience and predictive modelling. "Understanding audience needs should be fundamental," he said.

Machine learning algorithms tracked 900 audience attributes, with advertising products targeted to specific audiences. Using the same greater precision as on Facebook, it was achieving an open rate of 25 per cent with a smaller number of newsletters sent.

The podcast aspirations of Malaysia's BFM Radio were outlined by Umapagan Ampikaipakan, while Jang Group managing director Samad Ali told how QR codes had been used to give new life to classified advertising and beat off a competitor. The Pakistani publisher had also capitalised on its 70 years of publication with a "virtual gallery" of front pages.

"Everyone says newspapers are dead, but the key challenge it to recreate the habit through innovation and then reconnect offline and online," he said

In a parallel session, delegates had learned how a Malaysian publisher had done exactly that, delivering a special issue with both audio and video components. Star Media technical services manager Anandan Thangasamy told of the Pepsi promotion in which either a sound chip or embedded LCD video player had been included in copies, a project which took months of preparation and planning to overcome a range of challenges.

Other special editions had included a resistance band in a milk promotion, and a carrier bag which readers were encouraged to fill for charity. "Newspapers need to fight for attention," he said.

An uplifting finale to the conference came from the UK, Financial Times tech lead Lindsay Nicol, and Ben Whitelaw, head of audience and digital development at The Times and Sunday Times who turned to the use of AI and machine learning to optimise their websites and build engagement.

Whitelaw explained the strategies behind the 233-year-old newspaper's efforts to convert registered readers - who were allowed two free articles a week - into paying subscribers... and to get readers to "use what they've paid for".

Also of interest was an editorial policy which saw just three updates of the website a day with only rare breaking news. Readers "informed by data" were served special offers and offered the services a "digital butler", Ask James, to tailor communications to their interests and "get them to the point where they will give us their credit card details".

Nicol told how the FT had sought to encourage more women readers to its site, setting up an ML-based BOT to scan images and detect gender and emotion. "It's not perfect," Nicol admitted - showing a parliamentary group photograph in which "Janetbot" had thought prime minister Theresa May's chin was a "sad male" - "but it's a start". More in future posts and our upcoming print edition.

Peter Coleman

Pictured: Lindsay Nicol