Posted by: nickgerlich | March 12, 2008

Fries With That

It was about 2am. South suburban Chicago. I was giving Brian a lift home following another late night jamming with our bandmates. We were yet another garage band (long before it became a software title), not so cocky as to give up on our day jobs, but belting out chords and choruses as if our lives depended on it.

White Castle was open, as it always was. Every drunk in Chicago knew they could count on the Slyder House for a late-night drying out session. The Porcelain Palace served nasty little grease burgers that somehow took the edge off the evening’s liquid diet.

And so it was another typical Friday night. The crew was motley, on both sides of the counter. If you decided that high school wasn’t your schtick, you might get lucky and become a burger flipper or order-taker there.

Brian and I took our place in line, not because we were hammered, but because we had worked up an appetite playing tunes. Being young, bulletproof, and ten-feet-tall, we really took no notice of the garbage we were putting in our bellies. At 23, everything has food value, even if it is dubious.

I placed my order and moved aside, jingling the spare coins into my pocket. Brian stepped up to the plate and rattled off his stomach’s desire.

Fries“I’ll have six burgers and a large Coke.”

Would you like fries with that?

“No, thanks.”

Large or small?


My ears perked up, but I bit my tongue. Brian was clearly tired, and White Castle’s Recommendation Engine had just scored a direct hit. I wanted to see how long it would take before he noticed.

We walked to an empty booth and placed our trays down. I hesitated to start eating, wanting to see Brian recognize his faux pas. It didn’t take long.

“Hey, I didn’t order fries!”

Um, yes you did. I heard you say “large.”.

“Did I pay for them?”

Yep, like a good little robot. They could have sold you the store and you wouldn’t have noticed.

Ah yes, suggestive selling. It’s part of every fast-food employee’s training. Always try to upsell the customer. Do a quick scan of what they ordered, find something they omitted, and then suggest it.

In an e-commerce world, suggestive selling immediately presented itself as a problem. How in the world could a computer recommend something to a customer it cannot see? How could a machine make a sensible upsell recommendation without it being laughable or offensive?

Mathematicians and programmers have labored long into the night to develop recommendation engines that truly do make sense, and also increase the odds of add-on purchases. Much of Amazon’s success is based on the algorithms they have employed to quickly tell you “People who bought this book also bought…” And to take it a step farther, it explains those emails Amazonians receive with out-of-the-blue recommendations. The value of their relational database is beyond measure.

Netflix also employs a recommendation engine to help people make their next selection. So critical is this to their success that they even launched a competition in late 2006 to award $1 million to anyone who could improve the existing engine’s efficiency by 10%. Netflix wants its customers to be happy with their movies, so it is critically important to make good recommendations.

But how good can a machine be? And could it ever stand up to my wife with her feminine wiles, insights, and hunches?

Fortunately for me, I have proven to be pretty elusive and unpredictable for my wife, who often has no clue what I would want to read next. But Amazon has me pegged pretty closely. While it may not be able to detect eyerolls, smirks, and vocal tone, it does know nuance. And it knows books.

Which may explain why, although I don’t have any fries laying around here, I do have a lot of books that just magically show up.

And my wife just rolls her eyes. “I knew you would do it again.”

Dr “Gotta Find A Better Hiding Place Next Time” Gerlich


Leave a Reply

Please log in using one of these methods to post your comment: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s


%d bloggers like this: