As data capture and analysis continued to boom worldwide, the beauty industry was now looking to pivot around data further upstream, to help ingredient cataloguing, trend tracking and new product development. But what sort of hurdles faced industry when innovating around data? And was data really the future for beauty supply chains globally?

Speaking at last month’s in-cosmetics 2025 tradeshow in Amsterdam, a panel of industry experts weighed in.

Big data and blank spaces

“One of the things we need to be very mindful of is that there is so much data available now,” said Lorna Radford, founder and Managing Director of UK-based R&D centre Enkos Developments.

And at times, Radford said this could be counter-productive to beauty innovation. Big databases that gathered huge amounts of information on beauty and personal care products, for example, made it very easy for formulators, brands and companies to create dupes and copycats, she said. In addition to this, rising use of data-backed Artificial Intelligence (AI) tools tended to “average everything out”, which was the opposite of what was needed for true innovation, she said.

Industry, therefore, had to approach data “objectively” and seek out the “blank spaces”, she explained. “That’s where the true innovation is going to come from, rather than trying to copy what everybody else is doing. And that’s harder because it’s almost trying to look at the data that isn’t there versus the data that is there.”

Winnie Awa, founder of textured hair platform Carra Labs, agreed: “You don’t know what you’re looking for until you see an anomaly.”

But, Awa added that as soon as those gaps or that anomaly had been identified, it was crucial to add in a “layer of expertise” to cut through any data bias and find relevant insights.

“If you have a lot of unstructured data, sometimes you have to go through a lot of data cleansing before you reach anything meaningful,” she explained. “There’s an excitement around data but we need to remember to draw out the actionable insights of that data, especially when you have terabytes of it (…) It’s very, very important to actually be able to interpret the data and insights and use it to create something exciting, magical or powerful that can change the game.”

AI future - ChatGPT and large-language models

On the topic of AI and rising-star tools like ChatGPT and large-language models, the experts agreed progress still had to be made before these tools were reliant enough for beauty supply chains.

Timo Von Bargen, co-founder and co-CEO of B2B ingredients data platform Covalo, said AI was great for working through “unstructured data”, particularly documentation data in the likes of clinical trial papers, tech sheets, product brochures, ingredient lists and sustainability credentials. Covalo, for example, used AI to help users locate information within these documents as a sort of “pointing” tool, which helped cut down initial search times.

But beyond this, Von Bargen said use of AI remained limited. “It’s still a very nascent discipline now and we are learning a lot.” ChatGPT, for example, interbred “thousands of sources” which resulted in 40-50% accuracy levels of final information at times and required “a lot of data cleansing”.

Radford said Enkos Developments wasn’t able to rely on ChatGPT at the R&D stage yet because when experimenting to input in vivo data into the company’s system, the tool was only 85% accurate, “which isn’t good enough, yet”.

“I’m really keen to see how we can use these models, but at the moment, it’s just not quite cutting it as opposed to using a qualified chemist who knows what they are doing,” she said. “When models are specific enough and developed for a specific purpose, it can be helpful, but trying to use some of the generic AI tools available, if we then have to read through it anyway, it doesn’t save us time.”

However, based on the level of innovation happening in AI, she said the beauty industry would likely be having “a really different discussion” within the next few years.

Customer feedback and consumer needs

On the brand side, Radford said beauty companies had successfully starting using AI tools to aggregate customer product review data and identify common issues or feedback, which helped at product iteration stage. “With product development, it’s not static,” she explained, so live data was “really powerful” when working out how to improve a product.

Awa agreed data was powerful when trying to better understand and respond to consumer needs. Smart data platforms, for example, could pool relevant data and provide diagnostics on consumer concerns, helping to identify relevant products or influence innovation for individuals or groups, she said. But, she added that within this personalisation had to remain a priority. “Yes, data is great, aggregation is great, but we really need to make sure that personalisation exists all the way through.”

“...What we advocate for is understanding the consumer, within their needs and routines, so we can better formulate and create brands that speak directly to them,” Awa said.