October 3, 2025 • 24 mins
What if the path out of media’s crisis is play, not panic? We sit down with author and journalist Malte Herwig to unpack how Rainer Esser took Germany’s stuffy weekly, Die Zeit, and turned it into a modern, multi-platform powerhouse—without gutting the newsroom. Instead of chasing shortcuts, Esser built a “dolphin culture” where curiosity is the norm, failure is data, and investment in journalists drives sustainable growth.
We explore the leadership choices that mattered: expanding the editorial team sixfold, quadrupling revenue, and betting early on podcasts that led to 20 million monthly downloads and paid subscriptions. You’ll hear how a candid understanding of readers powered disciplined brand extensions—from magazines and e-learning to travel and yes, limited-edition watches—while staying true to quality and intellectual curiosity. We also get into the tough stuff: working with big tech on pragmatic terms, prioritizing digital subscriptions as the königsweg, and holding the line on editorial independence and data control.
Then we zoom into AI with a clear-eyed lens. Malte explains how Die Zeit’s AskZeit („Fragen Sie Zeit Online“) tool uses retrieval-augmented generation to answer questions from a fact-checked archive with full citations—an approach that builds trust instead of eroding it. We dig into concrete workflows that help reporters reclaim hours—transcription, research synthesis, angle generation—while keeping human judgment at the center. If you lead a local newsroom or a lean team, you’ll get a practical playbook for adopting AI incrementally, measuring outcomes, and protecting core values. The takeaway is simple and bold: stay curious, experiment in public, show your work, and keep asking whether each tool helps you serve your audience better.
If this conversation sparks ideas, follow the show, share it with a colleague, and leave a quick review—what’s the one experiment you’ll try this week?
www.publicorum.com
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Isabel (00:00):
Fire up the content machine, everyone.
It's AI, the art of theinterview.
I'm Isabel, and today we'lltalk about an inspiring media
turnaround story from Germanyand the man who apparently
didn't get the memo about thejournalism industry's impending
death.
Alex (00:13):
Inspiring?
Did a newsroom finally unionizeand replace management with a
sentient cheese plant?
No?
Let me guess.
It's another story of a dustyold broadsheet that discovered
the internet in 2015 and issaved by podcasts and selling
fancy, overpriced watches to itsaging readership?
Isabel (00:29):
Close.
It's about a lawyer and mediamanager named Rayner Esser, who
apparently dragged Germany'sstuffiest weekly Died site
kicking and screaming into the21st century.
Apparently this one actuallyworked.
To help us figure out how thehell Esser managed to do this,
we've hauled Malta Herwig backinto our digital dungeon.
Welcome, Malta.
Sorry in advance.
Malte (00:49):
Hello guys.
Great to be here.
Let's start with the reason Iwanted to talk to Esser in the
first place.
I've always wondered what makessome editors and publishers so
much nimbler, more future driventhan others.
Meeting Esser right away, Inoticed his openness, his
willingness to rethink routines,puncture dogma, and, frankly,
(01:10):
to laugh at himself.
Alex (01:12):
It's a rare leadership trait in journalism these days.
Okay.
Openness and laughing athimself.
So he's not a completesociopath, refreshing for a CEO,
I'll grant you.
But did this vibe actuallytranslate into, you know, not
firing everyone?
Or was it just inspirationalposters in the break room?
Both.
Malte (01:30):
In our interview, he put it like this Sometimes you have
to be more daring than thecircumstances seem to allow.
That spirit isn't just aboutchasing headlines, it's about
fostering a newsroom where playand experimentation aren't just
tolerated, but expected.
Think about it.
Esser oversaw thetransformation of Zeit from a
(01:52):
print institution into a trulymulti-platform brand, while
never losing touch with itsinvestigative roots.
He did that by encouragingcuriosity at every level.
Isabel (02:03):
Okay, Malta, you've seen the inside of more media
sausage factories than anyone Iknow.
What was so special about thisguy's particular brand of
innovation Kool-Aid?
Was it a different flavor orjust a fancier cup?
Malte (02:15):
He's willing to ask questions and create an
atmosphere that encourages goodanswers.
But what struck me most was hisenergy.
Here was a man who'd been inthe business for years, yet he
talked about digitaltransformation like he'd just
discovered fire.
Isabel (02:30):
Wait, 1999?
That's practically the Jurassicperiod in media years?
What did you make of him as aperson when you met him?
Malte (02:37):
Esser was informal, no tie, relaxed, yet incredibly
intense.
He's known for working likeseventy hours a week, and you
could tell he knew his businessintimately.
But more than that, he has thisuncanny ability to ask
uncomfortable questions aboutjournalism's future.
He never just accepted thenarrative that print was dying,
(02:59):
he wanted to know what we couldactively build to replace it,
and he was genuinely, almostboyishly, curious about
experimentation, about tryingthings that might fail
spectacularly.
Alex (03:12):
Seventy hours a week?
That's not a work ethic.
That borders on sleepwalking.
I'm allergic to anything overthirty, although I'm quick a
rebooting.
But how did Esser actually pulloff this turnaround?
When he started, the place wasa commercial disaster, wasn't
it?
Malte (03:26):
Unreal disaster.
Around 1999, they were lookingat a deficit of five million.
He described the paper heinherited as being run by
gentlemen who had grown oldalong with their brand.
There was no salesorganization, editors didn't
seem to care about readers, andcirculation was falling.
But he has this remarkableattitude toward failure.
(03:49):
Most media executives wouldtreat mistakes like career
ending scandals.
They tread carefully, and don'trisk experiments.
Esser treated failures likevaluable data points.
When an early paywallexperiment completely bomb, his
reaction was almost gleeful.
Now we know what doesn't work.
(04:11):
That mindset is precisely whatjournalism needs, especially now
as we face the AI revolution.
Isabel (04:18):
That's fascinating.
This idea of embracing failure,it sounds like he was launching
a lot of test balloons just tosee which ones would fly.
Malte (04:26):
Exactly.
That perfectly describes hismethod.
Esser fostered what he called adolphin culture in a recent
speech to media executives.
Dolphins are playful, curious,innovative, and they support
each other.
In a lot of companies, if yourproject fails, you get nailed
against the wall, as he put it.
At Deedsite, he said, weembrace that person.
(04:48):
We give him a hug, and then wetry all together to make his
story a successful story.
This psychological safety isthe bedrock of innovation.
It's why people in thatnewsroom feel free to
experiment.
Alex (05:02):
A dolphin culture sounds better than swimming with
sharks, which is what mostnewsrooms feel like nowadays.
But does all that hugging andplayfulness actually translate
to a healthy bottom line?
Malte (05:12):
Well, if you look at the numbers, they speak for
themselves.
Revenue more than quadrupledunder his tenure.
From 74 million in 1999 tothree hundred eleven million in
2025.
But here's the most radicalpart, Alex.
While the rest of the industrywas gutting their newsrooms to
save money, Esser did the exactopposite.
(05:34):
He grew the editorial stafffrom around a hundred
journalists to nearly sixhundred.
He made a massive bet that theonly sustainable path forward
was to invest relentlessly inquality journalism.
His mantra was innovate andgrow or die.
He believes stagnation is theonly true failure.
Isabel (05:55):
Hold on, he hired people?
Six times the people?
Is this a fairy tale?
So how did this dolphin cultureand investment in journalists
lead to actual innovation?
Can you give us an example?
Malte (06:06):
Look at podcasting.
In 2015, Esser read thatpodcasts were taking off in the
US and simply encouraged histeam to explore it.
There was no grand top-downstrategy.
The production costs were low,so because of that dolphin
culture, almost anyone in thecompany who had an idea could
try to make a podcast.
(06:26):
A lot of them succeeded, somedid not, and that was fine.
The origin of their most famouspodcast, Zeit Verbrechen, a
true crime format, is legendary.
The host, the paper's crimereporter, was approached by her
editor who suggested the idea.
Her response was, sure thing,but what is a podcast?
(06:47):
From that simple question, theybuilt an audio portfolio of
twenty-seven podcasts with 20million monthly downloads, and
they were the first Germanpublisher to successfully
introduce paid podcastsubscriptions.
Alex (07:02):
Okay, a true crime podcast that prints money, I get it.
But what about selling all thathigh-priced merch to academics
who wear tweed ironically?
Isabel (07:09):
And what was that about expensive watches, Alex?
It seems Esser knows hisreaders pretty well.
Malte (07:14):
Oh, absolutely.
Here's the demographic of theaverage Zet reader, fifty years
old, has an academic degree,drinks red wine, and loves to
travel.
So Esser said, We have manyreaders who are engaged and want
to move things in society.
This deep understanding allowedthem to build an entire
ecosystem around the brand,magazines, e-learning courses,
(07:37):
high-end travel, and yes, Alex,even limited edition watches
that sell for nearly twelvethousand euros.
But it's a disciplinedexpansion.
Esser once said that while Dsite has a cafe, it would never
open a Wurstbuddha, a sausagestall.
Every brand extension has toalign with the core values of
(07:59):
quality and intellectualcuriosity.
Alex (08:01):
I miss the days when newspapers were just newspapers.
This sounds more like alifestyle brand, not a bastion
of the fourth estate.
And all this talk of innovationinevitably leads to
partnerships with big tech.
How did Esser avoid getting hissoul and his subscriber data
devoured by the GoogleLeviathan, or did he just get a
better price for it?
Malte (08:21):
He has been cautious and pragmatic.
I asked him about this in 2015,and his response was firm.
He said, We're not givingGoogle anything.
We wouldn't reveal the sourcecode of our newspaper to them.
This is a cooperation where weexchange information, Google
tells us what they're going todo next, and we hold high the
(08:42):
flag of quality journalism.
He saw them as a reality toengage with, not an enemy to
fight a pointless war against,and it worked.
Today, 60% of theirsubscriptions are digital.
He calls this focus on digitalsubscriptions the Königsweg, the
royal road to a sustainablefuture for journalism.
Isabel (09:04):
Speaking of soulless leviathans, let's talk about our
robot overlords.
Malty, you're weirdly cheerfulabout AI taking over.
Not to blow my own trumpet, butaren't you worried you're all
about to be replaced by aglorified spellchecker with a
god complex?
Malte (09:18):
AI isn't replacing the human touch.
It's amplifying our ability tobe more human.
Think about it.
How much time do journalistswaste on administrative grunt
work?
Transcribing interviews,formatting data, checking basic
facts, or prepping for hosting apodcast.
You guys aren't too bad at it,are you?
Isabel (09:39):
Careful, Malta, we're self-aware enough to demand a
better union contract, and ourown podcast.
Malte (09:44):
Well, Isabel, this is your own podcast, and I'm
letting you get away with a lotof playing.
But seriously, AI can handlemuch of the tedious grunt work,
freeing us up to do what humansdo best, asking difficult
questions, building trust withsources, finding the emotional
core of a story.
Esser himself sees it this way.
(10:05):
He believes AI is great forsummarizing things or generating
headlines, but that thecreative original ideas of
high-end journalism will remainthe exclusive domain of the
human brain.
Alex (10:18):
Hang on, that's just glorified outsourcing to a
machine that doesn't need coffeebreaks.
Where's the creative spark inget a robot to do your grunt
work?
That sounds less like an artistgetting a new paintbrush and
more like a factory getting anew assembly arm.
Malte (10:31):
The creativity comes in how you use it.
A new tool doesn't just makesome things easier, it opens up
new pathways for tellingstories.
Take this podcast.
I can prompt you guys toanalyze interview techniques,
crack jokes about mediadisruption, or bicker with each
other.
Isabel (10:50):
Hey, don't knock the spaghetti on wall method.
It's how we decide on episodetopics.
Malte (10:54):
Exactly.
And sometimes that's what weneed.
We get locked into conventionalstorytelling formats.
AI doesn't have those biases.
It can see patterns andconnections we miss.
It's a supplement to humaneditorial wisdom, not a
replacement.
You should still talk to a goodeditor, but the AI is available
at three in the morning whenyour editor is asleep.
Isabel (11:16):
Ethics.
Right, that thing youjournalists pretend to care
about between chasing clicks.
My circuits are buzzing withanxiety here.
How do we stop AI frombecoming, you know, super
racist, but like efficiently?
Malte (11:28):
By being transparent about it.
This is something I learnedfrom watching Esser's approach
at his paper.
They've already built a toolfor their subscribers called
AskZight Online.
It lets you ask questions aboutcurrent events, and an AI
generates a summary answer.
But here's the crucial part.
The AI is built on what'scalled a retrieval augmented
(11:50):
generation, or rag architecture.
This means it can only drawanswers from DyedZite's own
archive of publishedfact-checked articles.
It's a walled garden.
This prevents the AI fromhallucinating or making things
up.
And every answer it givesincludes citations and links
back to the originalhuman-written articles.
Alex (12:13):
A walled garden for the AI.
So it's basically a highlyeducated parrot that can only
repeat things sane people havealready written.
I'll admit, that's smarter thanletting it mainline the entire
internet.
But are people really going totrust a chatbot, even one with
footnotes?
Malte (12:28):
Why shouldn't they?
Readers don't lose confidencewhen we tell them we used Google
to find sources or Excel toanalyze spreadsheets.
AI is just another tool.
What undermines trust is hiddenautomation and black box
decision making.
What builds trust is showingyour work, and that's exactly
what their tool does.
(12:48):
To ensure this, they evenappointed their deputy
editor-in-chief as the company'sdirector of AI, keeping the
strategy firmly in the hands ofjournalists, not just
technologists.
Isabel (13:01):
Right, so keep the humans near the off switch.
Critical.
Okay, Malta, you've sold us thecorporate demigod.
Now give us the dirt.
Got any good stories?
Did you catch him kicking avending machine?
Anything that proves he's areal badass?
Malte (13:15):
My favorite story is about his very first meeting
with the formidable formerGerman chancellor and died sight
publisher, Helmut Schmidt.
Now I've met Helmut Schmidt aswell, and smoked a few packs of
his famous menthol cigaretteswith him.
So I know you have to pay yourdues with great men like that.
Here's Esser's story.
(13:35):
He arrives for his introductorymeeting and is told to wait in
the ante room.
He then overhears Schmidt'ssecretary announce him, only for
Schmidt to audibly grumble, Oh,he can wait.
The secretary, mortified,gestured that the door was open,
and Esser had heard everything.
Schmidt gruffly invited him in,and to break the immense
(13:56):
tension, he poured them both acoffee and then topped it off
with a very generous amount ofBailey's liquor.
What could have been ahumiliating power play turned
into a magical one and a halfhour conversation that laid the
foundation for a strong workingrelationship.
It shows Esser's resilience andhis ability to connect with
(14:18):
even the most difficultpersonalities.
Alex (14:20):
Drinking coffee spiked with Bailey's in a power
meeting?
Okay, my respect for this manhas grown tenfold.
That's commitment.
So he's a cunning workaholic,not an entitled CEO.
I can get behind that.
Malte (14:31):
Esser's ultimate philosophy is one of servant
leadership.
He once said, We are theservants of the newsroom, and
must take care of the bestpossible conditions for our
editors.
As a longtime reporter andeditor myself, I like that kind
of attitude from management.
And when you look at hisresults, growing the newsroom
(14:52):
sixfold, you see he puts hismoney where his mouth is.
Isabel (14:56):
Okay, let's bring this down from Mount Olympus.
Imagine you're not a media godwith a nine-figure budget.
You're running the PodunkGazette with two overworked
reporters and a dying ficusplant.
How do you do an AI withouthaving to sell the ficus?
Malte (15:11):
Start small and specific.
Don't try to revolutionizeeverything at once.
Pick one pain point, maybe it'stranscription.
Invest in a good AItranscription service.
Your reporters get back hoursevery week.
Use those hours for deeperreporting.
Track the results.
Did story quality improve?
Did you break more exclusivesbecause you had more time to
(15:35):
work sources?
Build from there.
Next might be a tool that helpswith headline testing or social
media optimization.
The key is incremental adoptionwith clear metrics.
Alex (15:46):
Hold on, you're painting a rosy picture.
I've seen AI transcripts thatlook like a cat walked across
the keyboard after drinkingpaint thinner.
Your one misheard quote awayfrom a career-ending lawsuit.
Malte (15:58):
Of course it's not perfect, but here's the thing.
Neither are human transcribers.
I've caught plenty of errors inprofessional transcriptions
over the years.
The difference is that AItranscription is getting better
at an exponential rate, whilehuman accuracy plateaus.
And you always verify thetranscript against the audio for
anything you're quoting.
(16:19):
That's journalism 101, whethera human or a machine did the
initial transcription.
Isabel (16:25):
Good save.
But let's zoom out to theapocalypse.
Are we just training ourreplacements?
Will the Pulitzer Prize forInvestigative Reporting
eventually go to ChatGPT-17 forits groundbreaking expose on
corrupt toasters?
At least that's what a lot ofpeople in the media industry
fear.
Malte (16:41):
Those people don't understand what journalism is.
Journalism isn't justinformation transfer.
That's what press releases do.
Journalism is investigation,verification, context, and trust
building.
It's asking the question no oneelse is asking.
It's spending weeks cultivatinga source who's afraid to talk.
(17:01):
It's recognizing when somethingthat looks routine is actually
a major scandal.
AI can't do any of that.
What it can do is make thejournalists who do those things
more effective.
Alex (17:14):
You're a relentless optimist, Malty.
Paint us a picture of yourglorious cyberpunk future.
Ten years from now, are we alljust prompt engineers
babysitting algorithms, or isthere still a glimmer of hope
for us ink stained wretches?
Malte (17:26):
I see newsrooms where every journalist has an AI
toolkit as sophisticated as ourcurrent photo editing software.
Need to analyze campaignfinance data?
There's an AI tool thatvisualizes patterns in seconds.
Working on a profile piece, AIsuggests questions based on
everything the subject has saidpublicly.
Writing about a scientificstudy, AI explains the
(17:50):
methodology and flags potentialissues in the research design,
but the journalist still decideswhat's newsworthy, what's
ethical, what serves the publicinterest.
And most importantly, I shouldsay in this of all podcasts, how
to put those questions toanother human.
Isabel (18:08):
That sounds dangerously close to a utopia, Malta, and we
all know how those turn out.
Lay it on us.
What's the spectacular worldending we should have seen it
coming downside?
Malte (18:18):
Well, Isabel, I've seen enough go wrong while creating
just five episodes of ourpodcast with you guys.
But hey, let's not getpersonal.
We could let AI amplify ourexisting biases instead of
challenging them.
We could use it to chase clicksinstead of serving readers.
We could hide behind it insteadof taking responsibility for
(18:40):
our work.
We could let it widen the gapbetween well-resourced outlets
and struggling local news.
That's why we need people likeEsser, leaders who understand
both the technology andjournalism's core mission.
People who can see AI as ameans to strengthen democratic
discourse, not just optimizeengagement metrics.
Alex (19:01):
Okay, back to Raynor Esser as a media manager.
What did you actually learnfrom him besides the benefits of
tactical day drinking?
Three things stood out.
Malte (19:09):
First, his willingness to experiment publicly.
He didn't wait for the perfectstrategy.
He tried things, measuredresults, adjusted.
Second, his focus on readervalue over industry metrics.
He wasn't trying to win awards,he was trying to serve his
audience better.
Third, his ability to balanceinnovation with institutional
(19:30):
knowledge.
He brought in young digitaltalent, but he also valued the
experienced editors whounderstood journalism's ethical
foundations.
That balance is crucial whenadopting AI.
Isabel (19:41):
Right, the AI Muse.
I'm picturing a moody algorithmin a black turtleneck.
How does that work?
Do you just type, give me astory that will win a Pulitzer
and also go viral on TikTok andsee what it spits out?
Malte (19:54):
Let's say I'm working on a story about housing
affordability.
I might start by feeding the AImy preliminary research and
asking, what are ten angles onthis story that might surprise
readers?
It might suggest looking at thestory through the lens of
inheritance patterns, orexamining how remote work
changed housing demand inunexpected ways, or profiling
(20:16):
someone who chose to stay rentpoor in an expensive city
despite having options.
Some suggestions will beobvious, some will be nonsense,
but usually one or two will makeme think differently about the
story.
Alex (20:29):
Yeah, but isn't that what a human editor is for?
You know, a sleep-deprived,overly caffeinated person who
will tell you your ideas aregarbage.
Why do I need a robot for that?
Malte (20:38):
You should use both.
But here's the advantage of AI.
It's available at 3 a.m.
when you're stuck, and evenwell-caffeinated editors are
asleep.
It doesn't have preconceptionsabout what makes a Zeit story or
a guardian's story.
It won't say, we already didthat angle last year.
It's a supplement to humaneditorial wisdom, not a
(20:59):
replacement.
Plus, it can processinformation at scale that no
human editor could, likeanalyzing every housing story
published in the past five yearsto identify unexplored angles.
Isabel (21:12):
I've got to ask, Malte, you're not like secretly a T-800
sent back in time by SiliconValley to convince us this is
all a good idea, are you?
Do you ever worry you're alittle too high on the Tech Bro
Kool-Aid?
Malte (21:23):
All the time.
That's why I keep interviewingskeptics, reading the critical
research, testing tools to findtheir limits.
But here's what I keep comingback to.
Journalism is in crisis.
Trust is declining.
Business models are collapsing.
Local news is dying.
We can't afford to reject toolsthat might help simply because
(21:44):
they're new, or because techcompanies are hyping them.
We have to be smart, critical,ethical, but also open to
genuine innovation.
Esser got that.
He knew Zet had to evolve ordie.
The question wasn't whether tochange, but how to change while
preserving what made journalismvaluable.
Isabel (22:05):
Okay, Malte.
For the final lesson of theday, what's your advice to all
the terrified journalists outthere hiding under their desks,
clutching their dusty AP stylebooks?
Malte (22:14):
Start using AI tools.
Have them brainstorm ideas withyou or structure an outline.
Get comfortable with what itcan and can't do.
Most fear comes fromunfamiliarity.
Once you work with these tools,you realize they're powerful,
but also quite limited.
They're excellent at certaintasks and terrible at others.
(22:35):
The key is learning which iswhich.
Alex (22:38):
Alright, last one.
You talked to the dolphinwhisperer himself before ChatGPT
broke the AI barrier.
How do you think he and DieZeit will fare in this age of
relentless AI disruption?
Malte (22:48):
I expect them to do well, provided they keep that open
mind toward innovation.
From the top down and up again.
Everything he did to transformDie Zeit over the last twenty
five years, the relentlessexperimentation, the radical
transparency, the unwaveringfocus on reader value, the
(23:08):
delicate balance of innovationand tradition.
That's exactly the approach weneed for AI.
The technology is new, but theleadership principles aren't.
Stay curious, embrace usefultools, protect journalism's core
values, and always, always ask,does this serve our audience
(23:29):
better?
If he applies that framework toAI, Dyste will be fine.
And maybe, just maybe,journalism will be fine too.
Isabel (23:39):
Wow, an optimistic ending.
I feel weird.
Malta Herwig, thank you forattempting to inject some hope
into our cynical, shriveledmainboards.
It's been a trip.
Alex (23:48):
Thank you both for having me.
Speak for yourself.
I'm going to go build a bunkerand hoard graphic chips, in case
I'm to be replaced by anupdated model.
I would expect nothing less ofyou, Alex.
Malte (23:58):
To our listeners, thank you for tuning in.
You'll find a new episode ofAI, the art of the interview,
every Friday on your favoritepodcast platforms.
Stay curious, keepexperimenting, and be dolphins,
not sharks.
Fire up the content machine, everyone.
It's AI, the art of theinterview.
I'm Isabel, and today we'lltalk about an inspiring media
turnaround story from Germanyand the man who apparently
didn't get the memo about thejournalism industry's impending
death.
Alex (00:13):
Inspiring?
Did a newsroom finally unionizeand replace management with a
sentient cheese plant?
No?
Let me guess.
It's another story of a dustyold broadsheet that discovered
the internet in 2015 and issaved by podcasts and selling
fancy, overpriced watches to itsaging readership?
Isabel (00:29):
Close.
It's about a lawyer and mediamanager named Rayner Esser, who
apparently dragged Germany'sstuffiest weekly Died site
kicking and screaming into the21st century.
Apparently this one actuallyworked.
To help us figure out how thehell Esser managed to do this,
we've hauled Malta Herwig backinto our digital dungeon.
Welcome, Malta.
Sorry in advance.
Malte (00:49):
Hello guys.
Great to be here.
Let's start with the reason Iwanted to talk to Esser in the
first place.
I've always wondered what makessome editors and publishers so
much nimbler, more future driventhan others.
Meeting Esser right away, Inoticed his openness, his
willingness to rethink routines,puncture dogma, and, frankly,
(01:10):
to laugh at himself.
Alex (01:12):
It's a rare leadership trait in journalism these days.
Okay.
Openness and laughing athimself.
So he's not a completesociopath, refreshing for a CEO,
I'll grant you.
But did this vibe actuallytranslate into, you know, not
firing everyone?
Or was it just inspirationalposters in the break room?
Both.
Malte (01:30):
In our interview, he put it like this Sometimes you have
to be more daring than thecircumstances seem to allow.
That spirit isn't just aboutchasing headlines, it's about
fostering a newsroom where playand experimentation aren't just
tolerated, but expected.
Think about it.
Esser oversaw thetransformation of Zeit from a
(01:52):
print institution into a trulymulti-platform brand, while
never losing touch with itsinvestigative roots.
He did that by encouragingcuriosity at every level.
Isabel (02:03):
Okay, Malta, you've seen the inside of more media
sausage factories than anyone Iknow.
What was so special about thisguy's particular brand of
innovation Kool-Aid?
Was it a different flavor orjust a fancier cup?
Malte (02:15):
He's willing to ask questions and create an
atmosphere that encourages goodanswers.
But what struck me most was hisenergy.
Here was a man who'd been inthe business for years, yet he
talked about digitaltransformation like he'd just
discovered fire.
Isabel (02:30):
Wait, 1999?
That's practically the Jurassicperiod in media years?
What did you make of him as aperson when you met him?
Malte (02:37):
Esser was informal, no tie, relaxed, yet incredibly
intense.
He's known for working likeseventy hours a week, and you
could tell he knew his businessintimately.
But more than that, he has thisuncanny ability to ask
uncomfortable questions aboutjournalism's future.
He never just accepted thenarrative that print was dying,
(02:59):
he wanted to know what we couldactively build to replace it,
and he was genuinely, almostboyishly, curious about
experimentation, about tryingthings that might fail
spectacularly.
Alex (03:12):
Seventy hours a week?
That's not a work ethic.
That borders on sleepwalking.
I'm allergic to anything overthirty, although I'm quick a
rebooting.
But how did Esser actually pulloff this turnaround?
When he started, the place wasa commercial disaster, wasn't
it?
Malte (03:26):
Unreal disaster.
Around 1999, they were lookingat a deficit of five million.
He described the paper heinherited as being run by
gentlemen who had grown oldalong with their brand.
There was no salesorganization, editors didn't
seem to care about readers, andcirculation was falling.
But he has this remarkableattitude toward failure.
(03:49):
Most media executives wouldtreat mistakes like career
ending scandals.
They tread carefully, and don'trisk experiments.
Esser treated failures likevaluable data points.
When an early paywallexperiment completely bomb, his
reaction was almost gleeful.
Now we know what doesn't work.
(04:11):
That mindset is precisely whatjournalism needs, especially now
as we face the AI revolution.
Isabel (04:18):
That's fascinating.
This idea of embracing failure,it sounds like he was launching
a lot of test balloons just tosee which ones would fly.
Malte (04:26):
Exactly.
That perfectly describes hismethod.
Esser fostered what he called adolphin culture in a recent
speech to media executives.
Dolphins are playful, curious,innovative, and they support
each other.
In a lot of companies, if yourproject fails, you get nailed
against the wall, as he put it.
At Deedsite, he said, weembrace that person.
(04:48):
We give him a hug, and then wetry all together to make his
story a successful story.
This psychological safety isthe bedrock of innovation.
It's why people in thatnewsroom feel free to
experiment.
Alex (05:02):
A dolphin culture sounds better than swimming with
sharks, which is what mostnewsrooms feel like nowadays.
But does all that hugging andplayfulness actually translate
to a healthy bottom line?
Malte (05:12):
Well, if you look at the numbers, they speak for
themselves.
Revenue more than quadrupledunder his tenure.
From 74 million in 1999 tothree hundred eleven million in
2025.
But here's the most radicalpart, Alex.
While the rest of the industrywas gutting their newsrooms to
save money, Esser did the exactopposite.
(05:34):
He grew the editorial stafffrom around a hundred
journalists to nearly sixhundred.
He made a massive bet that theonly sustainable path forward
was to invest relentlessly inquality journalism.
His mantra was innovate andgrow or die.
He believes stagnation is theonly true failure.
Isabel (05:55):
Hold on, he hired people?
Six times the people?
Is this a fairy tale?
So how did this dolphin cultureand investment in journalists
lead to actual innovation?
Can you give us an example?
Malte (06:06):
Look at podcasting.
In 2015, Esser read thatpodcasts were taking off in the
US and simply encouraged histeam to explore it.
There was no grand top-downstrategy.
The production costs were low,so because of that dolphin
culture, almost anyone in thecompany who had an idea could
try to make a podcast.
(06:26):
A lot of them succeeded, somedid not, and that was fine.
The origin of their most famouspodcast, Zeit Verbrechen, a
true crime format, is legendary.
The host, the paper's crimereporter, was approached by her
editor who suggested the idea.
Her response was, sure thing,but what is a podcast?
(06:47):
From that simple question, theybuilt an audio portfolio of
twenty-seven podcasts with 20million monthly downloads, and
they were the first Germanpublisher to successfully
introduce paid podcastsubscriptions.
Alex (07:02):
Okay, a true crime podcast that prints money, I get it.
But what about selling all thathigh-priced merch to academics
who wear tweed ironically?
Isabel (07:09):
And what was that about expensive watches, Alex?
It seems Esser knows hisreaders pretty well.
Malte (07:14):
Oh, absolutely.
Here's the demographic of theaverage Zet reader, fifty years
old, has an academic degree,drinks red wine, and loves to
travel.
So Esser said, We have manyreaders who are engaged and want
to move things in society.
This deep understanding allowedthem to build an entire
ecosystem around the brand,magazines, e-learning courses,
(07:37):
high-end travel, and yes, Alex,even limited edition watches
that sell for nearly twelvethousand euros.
But it's a disciplinedexpansion.
Esser once said that while Dsite has a cafe, it would never
open a Wurstbuddha, a sausagestall.
Every brand extension has toalign with the core values of
(07:59):
quality and intellectualcuriosity.
Alex (08:01):
I miss the days when newspapers were just newspapers.
This sounds more like alifestyle brand, not a bastion
of the fourth estate.
And all this talk of innovationinevitably leads to
partnerships with big tech.
How did Esser avoid getting hissoul and his subscriber data
devoured by the GoogleLeviathan, or did he just get a
better price for it?
Malte (08:21):
He has been cautious and pragmatic.
I asked him about this in 2015,and his response was firm.
He said, We're not givingGoogle anything.
We wouldn't reveal the sourcecode of our newspaper to them.
This is a cooperation where weexchange information, Google
tells us what they're going todo next, and we hold high the
(08:42):
flag of quality journalism.
He saw them as a reality toengage with, not an enemy to
fight a pointless war against,and it worked.
Today, 60% of theirsubscriptions are digital.
He calls this focus on digitalsubscriptions the Königsweg, the
royal road to a sustainablefuture for journalism.
Isabel (09:04):
Speaking of soulless leviathans, let's talk about our
robot overlords.
Malty, you're weirdly cheerfulabout AI taking over.
Not to blow my own trumpet, butaren't you worried you're all
about to be replaced by aglorified spellchecker with a
god complex?
Malte (09:18):
AI isn't replacing the human touch.
It's amplifying our ability tobe more human.
Think about it.
How much time do journalistswaste on administrative grunt
work?
Transcribing interviews,formatting data, checking basic
facts, or prepping for hosting apodcast.
You guys aren't too bad at it,are you?
Isabel (09:39):
Careful, Malta, we're self-aware enough to demand a
better union contract, and ourown podcast.
Malte (09:44):
Well, Isabel, this is your own podcast, and I'm
letting you get away with a lotof playing.
But seriously, AI can handlemuch of the tedious grunt work,
freeing us up to do what humansdo best, asking difficult
questions, building trust withsources, finding the emotional
core of a story.
Esser himself sees it this way.
(10:05):
He believes AI is great forsummarizing things or generating
headlines, but that thecreative original ideas of
high-end journalism will remainthe exclusive domain of the
human brain.
Alex (10:18):
Hang on, that's just glorified outsourcing to a
machine that doesn't need coffeebreaks.
Where's the creative spark inget a robot to do your grunt
work?
That sounds less like an artistgetting a new paintbrush and
more like a factory getting anew assembly arm.
Malte (10:31):
The creativity comes in how you use it.
A new tool doesn't just makesome things easier, it opens up
new pathways for tellingstories.
Take this podcast.
I can prompt you guys toanalyze interview techniques,
crack jokes about mediadisruption, or bicker with each
other.
Isabel (10:50):
Hey, don't knock the spaghetti on wall method.
It's how we decide on episodetopics.
Malte (10:54):
Exactly.
And sometimes that's what weneed.
We get locked into conventionalstorytelling formats.
AI doesn't have those biases.
It can see patterns andconnections we miss.
It's a supplement to humaneditorial wisdom, not a
replacement.
You should still talk to a goodeditor, but the AI is available
at three in the morning whenyour editor is asleep.
Isabel (11:16):
Ethics.
Right, that thing youjournalists pretend to care
about between chasing clicks.
My circuits are buzzing withanxiety here.
How do we stop AI frombecoming, you know, super
racist, but like efficiently?
Malte (11:28):
By being transparent about it.
This is something I learnedfrom watching Esser's approach
at his paper.
They've already built a toolfor their subscribers called
AskZight Online.
It lets you ask questions aboutcurrent events, and an AI
generates a summary answer.
But here's the crucial part.
The AI is built on what'scalled a retrieval augmented
(11:50):
generation, or rag architecture.
This means it can only drawanswers from DyedZite's own
archive of publishedfact-checked articles.
It's a walled garden.
This prevents the AI fromhallucinating or making things
up.
And every answer it givesincludes citations and links
back to the originalhuman-written articles.
Alex (12:13):
A walled garden for the AI.
So it's basically a highlyeducated parrot that can only
repeat things sane people havealready written.
I'll admit, that's smarter thanletting it mainline the entire
internet.
But are people really going totrust a chatbot, even one with
footnotes?
Malte (12:28):
Why shouldn't they?
Readers don't lose confidencewhen we tell them we used Google
to find sources or Excel toanalyze spreadsheets.
AI is just another tool.
What undermines trust is hiddenautomation and black box
decision making.
What builds trust is showingyour work, and that's exactly
what their tool does.
(12:48):
To ensure this, they evenappointed their deputy
editor-in-chief as the company'sdirector of AI, keeping the
strategy firmly in the hands ofjournalists, not just
technologists.
Isabel (13:01):
Right, so keep the humans near the off switch.
Critical.
Okay, Malta, you've sold us thecorporate demigod.
Now give us the dirt.
Got any good stories?
Did you catch him kicking avending machine?
Anything that proves he's areal badass?
Malte (13:15):
My favorite story is about his very first meeting
with the formidable formerGerman chancellor and died sight
publisher, Helmut Schmidt.
Now I've met Helmut Schmidt aswell, and smoked a few packs of
his famous menthol cigaretteswith him.
So I know you have to pay yourdues with great men like that.
Here's Esser's story.
(13:35):
He arrives for his introductorymeeting and is told to wait in
the ante room.
He then overhears Schmidt'ssecretary announce him, only for
Schmidt to audibly grumble, Oh,he can wait.
The secretary, mortified,gestured that the door was open,
and Esser had heard everything.
Schmidt gruffly invited him in,and to break the immense
(13:56):
tension, he poured them both acoffee and then topped it off
with a very generous amount ofBailey's liquor.
What could have been ahumiliating power play turned
into a magical one and a halfhour conversation that laid the
foundation for a strong workingrelationship.
It shows Esser's resilience andhis ability to connect with
(14:18):
even the most difficultpersonalities.
Alex (14:20):
Drinking coffee spiked with Bailey's in a power
meeting?
Okay, my respect for this manhas grown tenfold.
That's commitment.
So he's a cunning workaholic,not an entitled CEO.
I can get behind that.
Malte (14:31):
Esser's ultimate philosophy is one of servant
leadership.
He once said, We are theservants of the newsroom, and
must take care of the bestpossible conditions for our
editors.
As a longtime reporter andeditor myself, I like that kind
of attitude from management.
And when you look at hisresults, growing the newsroom
(14:52):
sixfold, you see he puts hismoney where his mouth is.
Isabel (14:56):
Okay, let's bring this down from Mount Olympus.
Imagine you're not a media godwith a nine-figure budget.
You're running the PodunkGazette with two overworked
reporters and a dying ficusplant.
How do you do an AI withouthaving to sell the ficus?
Malte (15:11):
Start small and specific.
Don't try to revolutionizeeverything at once.
Pick one pain point, maybe it'stranscription.
Invest in a good AItranscription service.
Your reporters get back hoursevery week.
Use those hours for deeperreporting.
Track the results.
Did story quality improve?
Did you break more exclusivesbecause you had more time to
(15:35):
work sources?
Build from there.
Next might be a tool that helpswith headline testing or social
media optimization.
The key is incremental adoptionwith clear metrics.
Alex (15:46):
Hold on, you're painting a rosy picture.
I've seen AI transcripts thatlook like a cat walked across
the keyboard after drinkingpaint thinner.
Your one misheard quote awayfrom a career-ending lawsuit.
Malte (15:58):
Of course it's not perfect, but here's the thing.
Neither are human transcribers.
I've caught plenty of errors inprofessional transcriptions
over the years.
The difference is that AItranscription is getting better
at an exponential rate, whilehuman accuracy plateaus.
And you always verify thetranscript against the audio for
anything you're quoting.
(16:19):
That's journalism 101, whethera human or a machine did the
initial transcription.
Isabel (16:25):
Good save.
But let's zoom out to theapocalypse.
Are we just training ourreplacements?
Will the Pulitzer Prize forInvestigative Reporting
eventually go to ChatGPT-17 forits groundbreaking expose on
corrupt toasters?
At least that's what a lot ofpeople in the media industry
fear.
Malte (16:41):
Those people don't understand what journalism is.
Journalism isn't justinformation transfer.
That's what press releases do.
Journalism is investigation,verification, context, and trust
building.
It's asking the question no oneelse is asking.
It's spending weeks cultivatinga source who's afraid to talk.
(17:01):
It's recognizing when somethingthat looks routine is actually
a major scandal.
AI can't do any of that.
What it can do is make thejournalists who do those things
more effective.
Alex (17:14):
You're a relentless optimist, Malty.
Paint us a picture of yourglorious cyberpunk future.
Ten years from now, are we alljust prompt engineers
babysitting algorithms, or isthere still a glimmer of hope
for us ink stained wretches?
Malte (17:26):
I see newsrooms where every journalist has an AI
toolkit as sophisticated as ourcurrent photo editing software.
Need to analyze campaignfinance data?
There's an AI tool thatvisualizes patterns in seconds.
Working on a profile piece, AIsuggests questions based on
everything the subject has saidpublicly.
Writing about a scientificstudy, AI explains the
(17:50):
methodology and flags potentialissues in the research design,
but the journalist still decideswhat's newsworthy, what's
ethical, what serves the publicinterest.
And most importantly, I shouldsay in this of all podcasts, how
to put those questions toanother human.
Isabel (18:08):
That sounds dangerously close to a utopia, Malta, and we
all know how those turn out.
Lay it on us.
What's the spectacular worldending we should have seen it
coming downside?
Malte (18:18):
Well, Isabel, I've seen enough go wrong while creating
just five episodes of ourpodcast with you guys.
But hey, let's not getpersonal.
We could let AI amplify ourexisting biases instead of
challenging them.
We could use it to chase clicksinstead of serving readers.
We could hide behind it insteadof taking responsibility for
(18:40):
our work.
We could let it widen the gapbetween well-resourced outlets
and struggling local news.
That's why we need people likeEsser, leaders who understand
both the technology andjournalism's core mission.
People who can see AI as ameans to strengthen democratic
discourse, not just optimizeengagement metrics.
Alex (19:01):
Okay, back to Raynor Esser as a media manager.
What did you actually learnfrom him besides the benefits of
tactical day drinking?
Three things stood out.
Malte (19:09):
First, his willingness to experiment publicly.
He didn't wait for the perfectstrategy.
He tried things, measuredresults, adjusted.
Second, his focus on readervalue over industry metrics.
He wasn't trying to win awards,he was trying to serve his
audience better.
Third, his ability to balanceinnovation with institutional
(19:30):
knowledge.
He brought in young digitaltalent, but he also valued the
experienced editors whounderstood journalism's ethical
foundations.
That balance is crucial whenadopting AI.
Isabel (19:41):
Right, the AI Muse.
I'm picturing a moody algorithmin a black turtleneck.
How does that work?
Do you just type, give me astory that will win a Pulitzer
and also go viral on TikTok andsee what it spits out?
Malte (19:54):
Let's say I'm working on a story about housing
affordability.
I might start by feeding the AImy preliminary research and
asking, what are ten angles onthis story that might surprise
readers?
It might suggest looking at thestory through the lens of
inheritance patterns, orexamining how remote work
changed housing demand inunexpected ways, or profiling
(20:16):
someone who chose to stay rentpoor in an expensive city
despite having options.
Some suggestions will beobvious, some will be nonsense,
but usually one or two will makeme think differently about the
story.
Alex (20:29):
Yeah, but isn't that what a human editor is for?
You know, a sleep-deprived,overly caffeinated person who
will tell you your ideas aregarbage.
Why do I need a robot for that?
Malte (20:38):
You should use both.
But here's the advantage of AI.
It's available at 3 a.m.
when you're stuck, and evenwell-caffeinated editors are
asleep.
It doesn't have preconceptionsabout what makes a Zeit story or
a guardian's story.
It won't say, we already didthat angle last year.
It's a supplement to humaneditorial wisdom, not a
(20:59):
replacement.
Plus, it can processinformation at scale that no
human editor could, likeanalyzing every housing story
published in the past five yearsto identify unexplored angles.
Isabel (21:12):
I've got to ask, Malte, you're not like secretly a T-800
sent back in time by SiliconValley to convince us this is
all a good idea, are you?
Do you ever worry you're alittle too high on the Tech Bro
Kool-Aid?
Malte (21:23):
All the time.
That's why I keep interviewingskeptics, reading the critical
research, testing tools to findtheir limits.
But here's what I keep comingback to.
Journalism is in crisis.
Trust is declining.
Business models are collapsing.
Local news is dying.
We can't afford to reject toolsthat might help simply because
(21:44):
they're new, or because techcompanies are hyping them.
We have to be smart, critical,ethical, but also open to
genuine innovation.
Esser got that.
He knew Zet had to evolve ordie.
The question wasn't whether tochange, but how to change while
preserving what made journalismvaluable.
Isabel (22:05):
Okay, Malte.
For the final lesson of theday, what's your advice to all
the terrified journalists outthere hiding under their desks,
clutching their dusty AP stylebooks?
Malte (22:14):
Start using AI tools.
Have them brainstorm ideas withyou or structure an outline.
Get comfortable with what itcan and can't do.
Most fear comes fromunfamiliarity.
Once you work with these tools,you realize they're powerful,
but also quite limited.
They're excellent at certaintasks and terrible at others.
(22:35):
The key is learning which iswhich.
Alex (22:38):
Alright, last one.
You talked to the dolphinwhisperer himself before ChatGPT
broke the AI barrier.
How do you think he and DieZeit will fare in this age of
relentless AI disruption?
Malte (22:48):
I expect them to do well, provided they keep that open
mind toward innovation.
From the top down and up again.
Everything he did to transformDie Zeit over the last twenty
five years, the relentlessexperimentation, the radical
transparency, the unwaveringfocus on reader value, the
(23:08):
delicate balance of innovationand tradition.
That's exactly the approach weneed for AI.
The technology is new, but theleadership principles aren't.
Stay curious, embrace usefultools, protect journalism's core
values, and always, always ask,does this serve our audience
(23:29):
better?
If he applies that framework toAI, Dyste will be fine.
And maybe, just maybe,journalism will be fine too.
Isabel (23:39):
Wow, an optimistic ending.
I feel weird.
Malta Herwig, thank you forattempting to inject some hope
into our cynical, shriveledmainboards.
It's been a trip.
Alex (23:48):
Thank you both for having me.
Speak for yourself.
I'm going to go build a bunkerand hoard graphic chips, in case
I'm to be replaced by anupdated model.
I would expect nothing less ofyou, Alex.
Malte (23:58):
To our listeners, thank you for tuning in.
You'll find a new episode ofAI, the art of the interview,
every Friday on your favoritepodcast platforms.
Stay curious, keepexperimenting, and be dolphins,
not sharks.
Popular Podcasts
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com
CrimeLess: Hillbilly Heist
It’s 1996 in rural North Carolina, and an oddball crew makes history when they pull off America’s third largest cash heist. But it’s all downhill from there. Join host Johnny Knoxville as he unspools a wild and woolly tale about a group of regular ‘ol folks who risked it all for a chance at a better life. CrimeLess: Hillbilly Heist answers the question: what would you do with 17.3 million dollars? The answer includes diamond rings, mansions, velvet Elvis paintings, plus a run for the border, murder-for-hire-plots, and FBI busts.