Reflect.com

» see reflect.com as it lived, on the wayback machine

~ the last words ~

reflect was a us-only personalized-cosmetics web store run out of san francisco by procter & gamble. i never visited it. nobody i knew did. it was the kind of dot-com that existed inside an audience i was never part of, sold a category i would not buy, and shipped only to us addresses. its cultural footprint is mostly in business-school case studies about mass-customization that did not pay off.

procter & gamble launched reflect in late 1999 as a separate subsidiary, with co-investment from institutional venture partners and redpoint ventures. the model was unusually ambitious for a beauty site. customers filled out a web form about their skin and hair and stated preferences, then picked colors and fragrances, with control over ingredients and even the product name. roughly fifty thousand permutations of cosmetic and personal-care items were possible. each order was assembled to spec and shipped from a dedicated p&g facility. price points were set near department-store brands rather than mass-market drugstore tiers, which made the product a premium experiment rather than a volume play.

by 2003 the model had clearly not broken even. p&g had invested around sixty million dollars across three funding rounds. the company tried a pivot into brick-and-mortar partnership distribution with department stores, but the custom-fulfillment economics did not scale into retail any better than they had scaled online. p&g closed the site on june 17 2005 and folded the team back into the main beauty division. some of the personalization research migrated into a covergirl product line called custom compacts, which was itself short-lived. the reflect.com domain went dark.

~ the record ~

~ the pitch ~

reflect was pitched internally at p&g as the company's chance to learn what one-to-one customization looked like at industrial scale. the bet was that a brand that could profile every customer and ship a unique product would have the kind of margin protection that mass-market beauty no longer had. for a few years inside p&g it was treated as a research lab as much as a retail operation. the people working on it were genuinely solving hard logistics problems around small-batch manufacturing. the only thing they could not figure out was how to get the per-unit cost down to where a mass audience would buy in.

~ what got buried with it ~

most of what reflect actually built lived only inside p&g's internal knowledge base after the shutdown. the customization infrastructure went unused. the recommendation engine did not get carried over to any of the surviving p&g brands at any visible scale. the covergirl custom compacts product borrowed the idea but ran on a much simpler personalization model and was discontinued within a few years. by 2010 reflect was a footnote, cited in business-school books as the cautionary example of a corporate-backed e-commerce experiment that proved its hypothesis was wrong.

~ the funny part ~

the part that aged best is that p&g actually shut reflect down at the right time. most dot-com experiments of the 1999 cohort died in 2000 or 2001 along with the bubble. reflect lasted into 2005, more than four years past the crash, because p&g could afford to keep funding the lab. when the company finally closed it, the writeup was not a panic exit. it was an internal review concluding that the long-term return on a custom-cosmetics web channel was not going to justify the operating overhead. for a category whose obituary section is mostly about reckless burn, reflect's quiet corporate shutdown was the unusual one.

@}-->--

              )
             (_)
            .-'-.
            |   |
            |   |
            |   |
            |   |
          __|   |__   .-.
       .-'  |   |  `-:   :
      :     `---'     :-'
       `-._       _.-'
           '""""""

« previous  ·  random  ·  next »