When I started working with marketing and market research technologies almost a decade ago, few people were talking about “Big Data” and the buzz at Nielsen was about “Measurement Science,” not “Data Science.” 

Regardless of which buzzwords you prefer, some elements of the data architecture are true today as they were in the very beginning. 

Don't believe me? I'll let these quotes from some notable historic figures frame my argument. 

Tie Data to Important Business Outcomes

“If one does not know to which port one is sailing, no wind is favorable” — Seneca

I start all consultative projects by asking the client how they would define success and establish quantitative metrics that would demonstrate that success. 

Even more so in machine learning, you need a target variable to model your dataset against. If you don’t have a clear understanding of what outcome you want to achieve, then all outcomes are equally valid. 

Anything Can Be Quantified

“Measure what is measurable, and make measurable what is not so”unknown

Sentiment, engagement, happiness and loyalty all have metrics that can be examined. Even qualities with a more subjective definition can be quantified if they can be clearly defined.  

In business practices, metrics for employee engagement and customer loyalty and enthusiasm can be expressed in numbers. This is done by understanding the factors that contribute to these metrics and studying them, or studying the factors that detract from those outcomes. 

Putting the Answer Before the Data

“He uses statistics as a drunken man uses lamp posts — for support rather than illumination” — Andrew Lang

More commonly seen in politics, this also presents itself in businesses when a leader comes to a conclusion and seeks data to support that idea. You'll often see it in marketing teams that are tasked with developing their own attribution scores. 

Market research often calls it an “Is my baby pretty?” study. If you have an idea about what is going on in your business you may have a hypothesis, but not an answer. You should search for data that suggests that your theory is wrong as often as you search for data that supports you. 

Machine Learning Doesn't Replace Human Effort

“On two occasions I have been asked, 'Pray, Mr. Babbage, if you put into the machine wrong figures, will the right answers come out?' I am not able rightly to apprehend the kind of confusion of ideas that could provoke such a question” — Charles Babbage

Data cleaning, validation and transformation makes up 80 percent of my work. 

And I don’t expect that this will change. 

Do not let timeline-sensitive project managers downplay this step. This is really important especially as data science efforts are trying to build models that will automate major business processes. 

New machine learning platforms boast being plug-and-play, requiring little to no support and the ability to create actionable insights all by themselves. Usually the technical documentation tells a different story.  

If you plan on building intelligent machines, plan on employing talented humans to build and train them.  

Having Humans Do Tasks That Should Be Automated

“If one wants to make a machine mimic the behavior of the human computer in some complex operation one has to ask him how it is done, and then translate the answer into the form of an instruction table. Constructing instruction tables is usually described as ‘programming.'" — Alan Turing

Many business analyst waste their talents in processes that should be automated. I’ve seen some processes as simple as, “I take the values from a spreadsheet that someone else gives me and I enter them in another spreadsheet and send it on.” 

When I hear of this kind of process I picture water flowing down a river until it reaches a dam, where a bunch of people wait with buckets to carry the water 10 feet to dump it into another river. Now substitute data for water. 

Business processes like these waste human capitol on tasks that computers were designed for.

Lay a Foundation

“The scientific man does not aim at an immediate result. He does not expect that his advanced ideas will be readily taken up. His work is like that of the planter — for the future. His duty is to lay the foundation for those who are to come, and point the way" — Nikola Tesla

Digital transformation is a vantage point. None of the companies I’ve ever worked with are 100 percent ready to implement fully dynamic personalized content. Many are years away and some don’t even plan to get there. 

Regardless of where you stand on the digital maturity curve, you and your department should strive to set your company up for more robust personalization in the future. When your company is ready to move along that pathway of digital experience, your team will be ready.

Serving your customers well requires knowing them well. Giant, complex systems require a well-developed infrastructure that is best placed at the outset. 

I’m hoping to see a lot of fascinating changes in the digital experience marketplace, but I already know that some things will never change.