Concept Mining™ & Productive Analytics™

Coal Miner with Tagline

Hi, my name is Les Guessing and my mission is to help evolve the relationship between Data and Creative(s).

Wait, what? Data… in the Creative Department? WTF?

I know at first that might sound like a strange mashup. But take a minute to think about what Copywriters, Art Directors and Planners do every day. We’re essentially trying to find a way to interest people in some piece of relevant information. We are a $500 billion dollar global industry that’s about the communication of information.

Data can help you say better things. And say things better.

Well, information is just refined data. So while $Billions are being spent on ad tech that uses data to programmatically buy media, the goal of advertising isn’t to buy media. It’s to produce great, influential Creative work. I believe data’s natural home is in the Creative and Planning Departments.

Creative use of data can give your agency or company a measurable advantage.

Also, if you’re a Copywriter, Art Director or Planner, your job is basically to sit around and think, right? Doesn’t it make sense that a bunch of super-smart Machine Learning algorithms that can kinda-sorta help you do some of that thinking, might be a good thing? Learning what algorithms might be helpful for Creative people (out of the thousands of algos available) is something I spend a lot of time thinking about.

meth lab

What if we could take raw data and use science (Data Science) to cook it down into something powerful?

In fact, I believe there are some algorithms that are basically ways of reducing large amounts of unwieldy data into a more crystallized form – one that can be used for Creative purposes. (Imagine telling your client that you read thousands of relevant documents in order to arrive at their Creative strategy and the Creative work you’re about to show them.)

I may be the only Emmy-winning Copywriter / Creative Director in the world using subsets of Artificial Intelligence (Natural Language Processing and Machine Learning) along with programming tools and techniques from Data-driven Journalism, to turn raw Data into advertising content.

emmy statue


Predictive Analytics has been a part of the ad biz for awhile now. I do some of that. However, it’s probably best to think of my main pursuit asProductive Analytics”.

I like to dig into data, especially text data, with an eye towards producing strategies, concepts and content.

I love finding “Creative data opportunities” – often they’re sitting right under your nose. I want to help make Creative work better, Creative people’s lives easier and more fruitful, and help you attract and keep happy clients.

How does that work exactly?

First you have to either have data or get data. If you have data, I can mine it for concepts. If you don’t have data, I can write a web scraper in Python to get data from the web. I’ve also created an audio/video transcription web app using Node.js (JavaScript) and IBM Watson that’s about 90-95% accurate. So we could, for instance, mine the text of a large number of relevant Ted Talks, YouTube videos, etc.

Learn More

An important phase of doing Data Science is called EDA (Exploratory Data Analysis). During that phase, a Data Scientist is looking for signals in the noise – possible relationships in the data.

I’ve worked as a Copywriter/Creative Director in Los Angeles for over 25 years. So my mind is naturally oriented towards finding ads, jokes and stories in the data. I try to “solve for interesting”. And when I find it (or it finds me) I write a quick ad and/or strategy from it. I’m not looking to do “great work” at this point in the process, we all know that can take a lot of time. I’m looking to show “proof of concept.” This gives Creative Directors and Planners more options – options based in data, to choose from. By turning it into an ad immediately, you can get a better sense whether a possible strategy has real potential to turn into a great ad. (Who among us hasn’t worked with beautifully written creative strategies that, in the end, turn out to be un-executable?)

My relationship to data during that process is not unlike the way an improv performer asks the audience for, “A noun”, “A verb”, or “A relationship” and creates material from it. (In fact, Parts-of-Speech Tagging is just one of the many tools that Natural Language Processing can give you.)

Improvisation request

Why finding ideas in data is such a good idea.

Machine Learning is about extracting “intelligence”. Data can ground Creative in something tangible, something you can point to with clients and colleagues and say, “This has at least some objective truth to it.” That can result in more agreement and less middleware inside of an ad agency. And can make it harder for clients to throw good data-driven ideas into the trash. Meaning fewer soul-crushing rounds of creative to get to “yes”. Cover of book "Moneyball"(Remember, “Moneyball” was about using data to exploit the inefficiencies in the system.)

Finally, good writing is specific and good data is specific. “Three people walk into some place” is never going to be as funny as… “A Redneck, the Dalai Lama, and Sasquatch walk into a Lesbian bar…”. Good data can make for better punch lines, stories and ad content.

By mining data, Creative ideas can truly come from almost anywhere, from anyone, in the world. Especially through mining text data.

Remember, the world’s largest database, the Internet, is 80% unstructured text.

This can allow you to think outside your insights. I’m often surprised by how data can suggest ideas I would never have considered.

We often talk about “getting to know and understand a client’s business”. For me, there’s no better way to do that than brainstorming with relevant client data.

I love to learn. Which is why I love this area. It’s vast. And I have so much to learn. One of the things I’d love to learn is how my passion could help you and your clients profit.

I’d appreciate a brief opportunity to talk with you by phone. Or even make a short presentation in person (especially to your Creative and Planner folks) and answer any questions you might have.

Thank you,

Les Guessing
Creative Algorithm
Twitter: @LesGuessing