I am one of the incredibly fortunate people to call Hugh Dubberly my mentor — he “discovered” me, believed in me, and teaches me more every time I encounter him, even having known him for 20 years.

When I was working at Dubberly Design Office, he introduced me to Beatrice Warde’s essay, The Crystal Goblet. It’s a lovely essay that is only reproduced on the internet in the most ironically horrifying formats (in fact, I once saw it painstakingly set in the form of a goblet, poor Beatrice). It’s a perfect analogy for good research moderation. It’s in my head every time I sit down with a research participant. The reasons won’t make sense unless you read it, and it’s short, and it should be somewhere on the internet that isn’t unbearable to look at, so here you go, and I’ll pick up after Ms. Warde:


by Beatrice Warde
Excerpt from a Lecture to the British Typographers’ Guild

Imagine that you have before you a flagon of wine. You may choose your own favorite vintage for this imaginary demonstration, so that it be a deep shimmering crimson in color. You have two goblets before you. One is of solid gold, wrought in the most exquisite patterns. The other is of crystal-clear glass, thin as a bubble, and as transparent. Pour and drink; and according to your choice of goblet, I shall know whether or not you are a connoisseur of wine. For if you have no feelings about wine one way or the other, you will want the sensation of drinking the stuff out of a vessel that may have cost thousands of pounds; but if you are a member of that vanishing tribe, the amateurs of fine vintages, you will choose the crystal, because everything about it is calculated to reveal rather than to hide the beautiful thing which it was meant to contain.

Bear with me in this long-winded and fragrant metaphor; for you will find that almost all the virtues of the perfect wine-glass have a parallel in typography. There is the long, thin stem that obviates fingerprints on the bowl. Why? Because no cloud must come between your eyes and the fiery hearth of the liquid. Are not the margins on book pages similarly meant to obviate the necessity of fingering the type-pages? Again: The glass is colorless or at the most only faintly tinged in the bowl, because the connoisseur judges wine partly by its color and is impatient of anything that alters it. There are a thousand mannerisms in typography that are as impudent and arbitrary as putting port in tumblers of red or green glass! When a goblet has a base that looks too small for security, it does not matter how cleverly it is weighted; you feel nervous lest it should tip over. There are ways of setting lines of type which may work well enough, and yet keep the reader subconsciously worried by the fear of “doubling” lines, reading three words as one, and so forth.

Printing demands a humility of mind, for the lack of which many of the fine arts are even now floundering in self-conscious and maudlin experiments. There is nothing simple or dull in achieving the transparent page. Vulgar ostentation is twice as easy as discipline. When you realise that ugly typography never effaces itself, you will be able to capture beauty as the wise men capture happiness by aiming at something else. The “stunt typographer” learns the fickleness of rich men who hate to read. Not for them are long breaths held over serif and kern, they will not appreciate your splitting of hair-spaces. Nobody (save the other craftsmen) will appreciate half your skill. But you may spend endless years of happy experiment in devising that crystalline goblet which is worthy to hold the vintage of the human mind.



As a researcher, I believe my job is to be a crystal goblet: to be there as little as possible. To make space for the research participant to fill. To, “reveal rather than to hide the beautiful thing which it was meant to contain…”, indeed to be, “worthy to hold the vintage of the human mind.””

All of our most fundamental rules about research moderation hold this as our north star. To wit:


Seriously. Less talking. Listen. Help the participant feel that what they’re saying is of interest. Your job is to be there as little as possible. You can tell if you’re doing this right by getting a transcript done of your research session. (We use Rev to do this quickly and reliably, not most cheaply.) Look at your transcript. You’ll see good researchers have a sentence, half a sentence that trails off… and then paragraphs of response. This is a sign that you’ve created space for that person to fill. You are almost transparent. All these other rules are really techniques for shutting the f#¢& up…


People are uncomfortable with silence. Breathe, count to 5 in your head, wait. They will be moved to fill the silence with that thing they weren’t sure they would share, the thing beyond their pat answer. See what they really have to say.


This creates an immediate uncomfortable silence, and also keeps you from putting words in their mouths. So instead of, “Would you say this experience was good?” (this is a terrible technique) or even, “Would you say this experience was good or bad?” (also not good) instead, “Would you say this experience was…?””


See Steve Portigal’s iconic TiVo story. If they say “TIE-vo”, you say “TIE-vo”. If they say “wirey thing-y”, don’t say “dongle?”. Say “wirey thing-y”. Remember, you’re trying to be there as little as possible. By using their words, you are inserting yourself less as a presence.


Remember: attitude is a terrible predictor of behavior. Whether or not they tell you they’ll buy or use a product in the future has almost nothing to do with whether or not they actually do. (If you must, better to ask them to predict what other people like them will do.)


The Crystal Goblet analogy shows up after research too: when you’re considering what you’ve learned, be careful that you don’t insert your own ideas into what you heard. This is called confirmation bias, and it’s hard to avoid. Well, it’s really impossible to avoid. Just being aware will help keep you from turning into a flagon. Including other people who don’t care as much about the outcome can help too.


The folks I share this idea most often with are Product Managers. More and more they’re getting out there and talking to users. This is GREAT. A+++. Five gold stars. But I worry about the garbage-in, garbage-out phenomenon. Great Product Managers are usually successful specifically because they’re charismatic, they believe in what they’re doing, they get other people excited about their ideas too! So they have to flip a switch, when talking to users and trying to learn about them, to become a crystal goblet. Otherwise Product Managers just learning about how persuasive they can be in the moment.

My favorite trick for addressing all these problems (at least to some degree) for Product Managers: Don’t test your own product. Just as we don’t have Developers test their own code, because they have a fixed idea of how it works, Product Managers can swap and test one another’s projects.

Would love to hear other researcher’s thoughts on this analogy, more rules this does (or doesn’t apply to), and from Product Managers about whether this works or resonates with them!