Sun scream
Welcome to Best of the Week, written on a bright morning in Evandale, Tasmania.
Today: How Ultra Violette’s co-founder destroyed her brand; some dubious AI claims; and the Unmade Index stumbles into the weekend
Happy Cuban Sandwich Day. I’ve no idea what one is, but it was a minor plot point in yesterday’s new episode of Dexter: Resurrection, so there you go.
It’s your last chance to sign up to a paid membership of Unmade and lock in all of the current benefits. In a few weeks’ time, we’re going to stop accepting new paying members of Unmade. Instead we’ll be offering membership of an expanded Mumbrella Pro as we bring the two brands closer together.
All Unmade membership perks will be carried across, including complimentary tickets to REmade, Unlock and Compass for our annual paying members. These won’t be available to anyone else as part of the new Mumbrella Pro membership.
Your paid membership also includes exclusive analysis and access to our content archive which goes behind the paywall six weeks after publication.
Upgrade now or miss out.
How to kill a brand
It’s the journo in me, but I must confess to a certain bias when I see brands go to war with the watchdogs. I want the watchdogs to win.
It took two months, but this week, Ultra Violette waved the white flag. I was always on Team Choice.
You may recall that Choice published research earlier this year in which 16 of the 20 sunscreen products it tested did not meet the sun protection factor claims on their labels.
On Reddit, it was the AusSkincare community’s 9/11.
Among the worst performers in the tests was Ultra Violette’s Lean Screen product.
The brand’s response became a case study in how not to deal with a PR crisis.
Co-founder Ava Chandler-Matthews took to social media, disputing Choice’s work, and telling her followers she used the product on her own children.
Instead, she argued there were flaws in the testing process. She came across as defensive and in denial.
As a result, Ultra Violette, rather than any of the other brands that failed the tests, became the one associated with the report.
Many beauty products are built entirely on the brands. Companies like Ultra Violette don’t make the products themselves; they’re white label products with production outsourced to others. The only value the brands have are what they create themselves.
This week, Ultra Violette admitted new tests on Lean Screen showed “a concerning discrepancy”. For something supposed to offer factor 50 protection, five of the eight tests were below that, with one scoring as low as 4.
They’ve withdrawn the product.
As Ultra Violette told customers in its statement: “Trust is built and rebuilt over time, and we know we have work to do.”
With a product like sunscreen, the stakes are high. With commenters on the brand’s social media pointing to the consequences.
When the Choice report first emerged, Ultra Violette could have survived as a brand. There’s an entire product range beyond Lean Screen.
Now though, it’s best known as a product that doesn’t work. I don’t see how they can come back from that.
AI and other cloaks
One of the many curiosities of the state of generative artificial intelligence is that a business announcing a new AI initiative may mean a little or a lot. And it’s often impossible to tell from the outside.
There were a couple of examples this week, involving CommBank and B2B publishing house Aspermont.
A weird development was CommBank’s sudden U-turn on AI. Having announced 45 job cuts last month thanks to the success of its AI chatbot in driving down call waiting times, the bank said it was mistaken and would not be making the redundancies after all.
The CommBank statement was vague, saying it "did not adequately consider all relevant business considerations" and "we should have been more thorough in our assessment of the roles required".
It’s possible, even likely, that the abilities of AI chatbots had little to do with either decision. When the bank first announced the cuts, the Finance Sector Union claimed the real agenda was to send the work offshore. The union had already taken CommBank to the Fair Work Commission claiming it had breached an agreement by outsourcing jobs to India.
It’s hard to believe that in the space of a week, new data on the performance of the chatbots emerged, was analysed and led to a reversal.
What did occur though was the final stage of the union-led court case against Qantas for illegally sacking baggage handlers, which was accompanied by a $90m fine.
I have a hunch the decision was a pragmatic one about avoiding union confrontation rather than based on the abilities of the AI.
And Aspermont also made an ASX announcement cloaked in AI.
The resources publisher - which last month disclosed to the ASX that it would run out of money in little more than a year unless the business changed trajectory - put out an announcement to the market.
The final part of the announcement was arguably the most consequential, revealing a capital raising by issuing new stock in the company, bringing in $1.8m from as yet undisclosed overseas investors.
It was hard to tell whether there was spin or substance to the rest of the statement.
Aspermont announced what it described as a “strategic agreement” with mining giant Rio Tinto. It will put its Mining Journal archives into a large language model. This will (and in the announcement, this bit was in bold) see Aspermont “become the leading provider of data and intelligence solutions for the global mining industry”.
Rio Tinto will give Aspermont $550,000 for the privilege of six months of exclusive access to the platform when it’s ready.
What wasn’t disclosed though was how much Rio Tinto usually spends with Aspermont, and whether this would actually represent $550,000 of new money from its client.
The idea of digitising your archive and putting it into an LLM may make sense, but it’s hardly revolutionary. Remember when we took our archive and put it into an LLM called TimBot more than two years ago? I’m kicking myself that we didn’t announce it to the ASX.
And the final part of the Aspermont announcement is the launch of its “proprietary data and intelligence platform” Mining-IQ, complete with “global datasets” and a “suite of analytics products”.
Strategically, the right move. Just look at how much more profitable News Corp’s Dow Jones data division is than its news division. Data is more valuable than news.
But it needs to be a good product, and subscribers need to be willing to pay for it.
Will that be the case for Aspermont? We’ll see.
Unmade Index dips into the weekend
The Unmade Index finished the week on a downer, with most ASX-listed media and marketing stocks taking a tumble.
Among the larger stocks, Seven West Media had the worst of it, losing 3.3%. ARN Media lost 3.2%. Agency holdco Enero lost 3.1%.
The Unmade Index finished the week on 584.9 points, a loss of 0.7% for the day.
In case you missed it from Unmade…
More from Mumbrella…
Opinion: Note to brands and agencies: AI is both an enabler and a smiling assassin
The case of Margaux Blanchard: Publishers fall for AI-written articles
Nine’s A Current Affair boss steps down after Logies misstep
Time to leave you to your Saturday.
If you’d like to hear a little more from me, last night’s episode of MediaLand is now available. Our guest was ABC boss Hugh Marks. With co-host Vivienne Kelly off on some sort of diving expedition, I was joined by my former Mumbrella colleague (and now at the Australian Financial Review) Zoe Samios. As well as quizzing Marks about his plans for the ABC, Zoe and I tackled the latest sports rights deals, the Rupert Murdoch movie, and her unlikely love of wrestling.
We’ll be back with more soon.
Have a great weekend
Toodlepip…
Tim Burrowes
Publisher - Unmade + Mumbrella
tim@unmade.media
