Tuesday, September 19, 2017

Scientific American's embarrassing example of the premise story, or "yes, it is just a god damn vending machine."

When criticizing the overhyped for being overhyped, one always runs the risk of becoming part of the problem. That said, the badly conceived vending machine startup Bodega is just too good an example of what's wrong with the way we discuss technology and business. It is one of those cases where multiple lines of bad journalism and bad investor logic converge.

For instance, one of the most popular and damaging tech narratives is the premise story. Some company or entrepreneur makes some ridiculous claim about how his or her (the former most of the time) new product or service will revolutionize everything. This claim will almost certainly never come to pass but it gets a tremendous amount of buzz and provides a great excuse to discuss something a writer would like to talk about. The result is a premise story.

The premise story provides a highly useful loophole for tech and business writers. An article about a proposal that is obviously fatally flawed needs to address the lack of viability, but when the proposal is framed as a premise ("if we have [Mars one/Hyperloops/Bodega boxes], what will life be like."), The author can slide by with the mildest of conditional statements.

The trouble is that the implications of an event that almost certainly won't happen might occasionally be interesting in a thought experiment sense but are, by definition, trivial. Since the authors of these pieces want to be taken seriously, they inevitably if implicitly lend an unwanted degree of credibility to these impractical schemes.

Case in point.  Krystal D'Costa, blogging for Scientific American, clearly wanted to write something about real bodegas, their history, their cultural significance, the roles they play in their communities, this is an interesting and admirable topic, but in order to make it marketable and bring in the traffic, she needed to tie it to one of the silliest startup ideas to come out of Silicon Valley in recent memory. (At this point, I am supposed to say "since Juicero," but, at the risk of praising with faint damns, Bodega's business plan really is worse.)

Thus we get these opening paragraphs:
Need deodorant? Or craving potato chips? Don’t want to run to the “corner store” to get it? Well, soon enough you may only need to go as far as a deluxe vending machine. A start-up is looking to put "pantry boxes" stocked with the non-perishable items you might find at a corner store in key locations, like your gym, or a dorm common area or apartment lobbies. The name for this venture? Bodega.

As the name implies, the creators want to replace brick-and-mortar bodegas—the name given to those ubiquitous corner stores in New York City and Los Angeles—with these pantry boxes. They believe that with time, and data, they can tailor these boxes to meet the specific needs of an immediate area by tracking purchases and restocking those items that are most commonly purchased there, which is what seems to distinguish these boxes from normal vending machines. (The logistics of this restocking has not been disclosed as it is presumably a large part of the business plan.) But the idea has drawn a public outcry—and for good reason: It presents a very real threat to smaller mom-and-pop type stores, but it also appropriates the name “bodega," without acknowledging the cultural and social capital that these spaces have.

That “very real threat” is what makes this a topical, click-worthy story; it is also complete bullshit. In terms of functionality, these pantry boxes are nothing more than fairly crappy vending machines. There is nothing you can get out of them that isn't available from existing machines and a great deal that you can get from other machines but you can't get from them. No cold sodas. No hot beverages. No frozen treats. (We won't even get into many of the more elaborate, robotic offerings in Japan and Europe.)

The pantry boxes fair even worse compared to bodegas and other convenience stores in terms of selection. In addition to all the things listed above, they offer a small fraction of the nonperishable goods available in even a tiny store, not to mention no produce, alcohol, or tobacco products.

In lieu of any real improvement in functionality, the founders offer magical heuristics. Ex-Google employees from Silicon Valley using artificial intelligence to disrupt an industry. Even "data" is used here in the magical rather than statistical sense. Large, well-established vending machine companies have tons of quality data going back forever and constantly being updated. Bodega is actually working at a huge disadvantage here, one that would take years of widescale operation to erase.

One of the most telling lines in the excerpt above is the speculation about the logistics of restocking. This is, of course, an enormously challenging business problem. It is one of the primary reasons why vending machines have not already carved out a much larger segment of the retail market.  D'Costa assumes that the company has a workable solution to the problem and is just keeping it to itself. There is absolutely no reason to believe this – – Silicon Valley startups routinely jump into press releases and funding pushes with inchoate and/or unworkable business plans – – but believing that the details have been carefully thought out is essential for preserving the premise.

The sad part is that the rest of  D'Costa's article is interesting and well worth reading. It's a shame she felt the need to prop it up.


  1. Mark:

    You write, "Large, well-established vending machine companies have tons of quality data going back forever and constantly being updated. Bodega is actually working at a huge disadvantage here, one that would take years of widescale operation to erase."

    But is it possible you're missing the point here? Maybe the business plan is for these people to get tons of funding from their friends, and then use that funding to buy some vending machine companies and get that expertise?

    1. I suspect the focus is less on who they buy and more on who buys them.