-
Will formulators become obsolete? What do you think?
Someone asked me at a recent conference whether formulation jobs would become obsolete in a few years. It turns out that P&G is already using computers to create new iterations of their household cleaning products.
I explore the idea of whether cosmetic chemists will become obsolete.What do you think? Are we formulators on our way out? -
9 Replies
-
depends on what’s required from the job; if it’s very simple, low-risk work like changing a fragrance or surfactant levels, I can see how that might be automated
if the job involves more advanced tasks, e.g. matching an existing product with a spurious/incomplete INCI list, then the sheer amount of available materials, the assorted aspects of process engineering associated with them, and the occasional thinking outside the box that’s required means that a human mind would be needed to create something in a truly elegant fashion
plus, if you want new products which meet the customer’s requirements, are stable, are reasonably within budget, and can be made correctly and consistently on your plant, that’s apparently beyond the capabilities of some humans - so there’s no hope of teaching a computer to do it any time soon!
-
Unless the computers learn how to feel, show emotions and understand the complex relationship between looks, texture, smell, etc of a product I don’t see how formulators will be come obsolete.
But I have to say I’m looking forward to a day when I can plug a formula in a computer and it can at least predict stability and compatibility. -
I think you nailed it on everything. There are a couple of things I’d like to add:
1) Objective vs. subjective evaluation - “Cleaning” lends itself well to objective evaluation. Add a specified amount of a standard soil, clean for a fixed amount of time, determine how much soil and/or detergent residue is left. Easy to put numbers to this. “Elegant skin feel” - not so much. Who decides? How do you assign numerical values to this? Without values, computerization/automation is useless. For much the same reason, DOE/Six Sigma techniques don’t work well for cosmetics.2) Market size/type. I started off my formulation career working on fabric softeners for Unilever. Their pilot area alone was larger than any cosmetic companies main liquid manufacturing area I’ve ever seen, and their pilot tanks were 2 or 3 times larger. Cleaners, detergents, etc. are an enormous market. Cosmetics, in contrast, is a small-volume, high-profit industry. When I started, I was told that $0.10 cent savings per gallon in the cost of a fabric softener formula would save the company a million dollars/year - and the fabric softener volume was less than a quarter of the laundry soap volume. With those kind of savings possible, it makes sense to invest in an automated lab to engage in the formulating equivalent of “high-throughput chemistry” to sort through all possible formula variations in order to optimize their formulas. In contrast, the amount of money saved from optimizing, or making a small variation on a cosmetic formula, frequently pales before the amount of revenue you’d get from introducing a new/unique formula - which computers currently cannot make, or identify. -
“Elegant skin feel” - not so much. Who decides? How do you assign numerical values to this? Without values, computerization/automation is useless. For much the same reason, DOE/Six Sigma techniques don’t work well for cosmetics.”
Maybe this part wouldn’t the hardest one. Shiseido for instance is trying to put this in numbers:“Research regarding cosmetic affective evaluation using electroencephalogram analysisFor affective impression related evaluation such as texture of cosmetics, highly-trained specialists are in charge. However, in order to further comprehend how customers actually feel in an objective manner, we have been conducting the joint research with Professor Masahiro Nakagawa, Nagaoka University of Technology. It does not utilize the electroencephalogram in the low frequency region (α wave band, θ wave band, etc.), which is utilized in the conventional electroencephalogram analysis, but it is the technique to detect the customers’ predominant impression when using cosmetics based on fractal dimension analysis of electroencephalogram in the middle frequency region before the customers become aware of the sensitivity.To be specific, we ask the customer to picture sensitivity words such as “refreshed”, “sticky” and “moist” and record the electroencephalogram at that moment. After this, when the same customer actually uses the cosmetic product, we take her electroencephalogram and analyze it by sensitivity word pattern, enabling us to comprehend how much the customer actually experience those impressions.”
Source: http://www.shiseidogroup.com/rd/development/brainresearch.html
Maybe in the future a system could help to formulate a non-stick, smooth formula (elegant…) without to have to try it on many subjects.
-
Simple analogy:
We now have the interwebs … is bulk rate junk mail still showing up in you s-mail box? Do people still drive to work?If the computers are programmed to believe manufacturers’ claims on the capabilities of their ingredients … I rest my case. -
I think a lack of quantification of many characteristics of cosmetic products is the main hurdle that computer formulators will have to overcome.
-
I’ve got to add, though - having a lab automated enough to replace a lab technician or two would be truly awesome. The idea of being able to just load chemicals into a device that would weigh, heat, mix, cool, and then clean automatically gives me chills.
-
Only if customers become obsolete.
-
Exactly DavidW - The day our customers realize they don’t need another shampoo formulation - we are obsolete.
Log in to reply.