Latte Art Calibrations

So (over a span of several weeks) during the training for latte art, I have asked the folks whom I report to if we can have specific targets of art styles to emulate. The answer has been no, and that instead we should deliver what is considered "pretty to the customer," which eventually led me to think of how we should design a survey with minimum bias to determine what latte art customers would find "pretty."

Should it be like Facemash, where we structure a primarily binary choice between two milk drawings selected to isolate a particular parameter? Or should we simply throw random sets of images at customers and as for a "prettiness rating," from 1-5. Perhaps in the latter case, customers should be first asked to identify their "prettiest," and "least pretty," image from the given set of images, so that we can normalise the distribution of their ratings.

Just a few thoughts which I really shouldn't be bothering with on my day off from work. I'll try to avoid thinking about this, and to come back to it when I am on shift.
Conversation continue, none the less, and since it is a on a narrowband medium (Whatsapp), I offer that "pretty latte art," on Google is probably a safe mean, less then etching and hard foam. Also, a couple of popular tutorials published by other cafes was brought up, so I will be emulating those directly.

Statigram searches by hashtag, sorted by likes, also works. This was insight from a colleague.

Day 20 on the job purely as a trainee barista. Things look swell, albeit a little mundane. But that is where the rest of my hobbies come in, I suppose. Here's to a simple life. For now.

No comments :

Post a Comment