Kevin Hartman: “We are at a greater risk of data being manipulated.”
Kevin Hartman considers himself both a scientist and an artist. He believes that data must be able to create something impactful and unlock experiences that uplift today’s consumers. That sentiment drives his work as the Director of Analytics for Google’s Consumer, Government & Entertainment sector and also his practice as a seasoned teacher of digital analytics.
Kevin’s most recent book, Digital Marketing Analytics: In Theory And In Practice, stands out for detailing three functional skills of an analyst that combine the art and science of data.
Through a remote interview, Kevin spoke to us about his day job, all things data as well as his role as an educator, which gave us some solid takeaways:
- Every business executive, besides digital analysts, must be able to detect manipulated data.
- The current measurement infrastructure will soon become extinct due to increasing privacy concerns.
- Virtual reality can be tapped to collect data and create more personalized experiences for consumers.
- Companies must first invest in a good analyst before investing in expensive software.
- Nobody is a born analyst. It is a gift that can be learned.
On investing in good analysts
What’s your advice to companies who are sitting on a trove of data?
I’ve got three pieces of advice for those companies. The first one is: hire a data analyst. The second one is: hire a second data analyst. And the third is: hire more data analysts.
In my experience, the companies that have struggled with turning data that they've collected into insights are simply those that don't have the human resource they need to do that work. Companies will fall into the trap of buying technology and feeling like now they're done.
But it just doesn't work. It's about having people who can bring functional skills to bear against the data that's being collected and tell stories that fit the business context. That’s what has separated a successful company from one that struggles in this area.
Would you say analysts are more important than softwares?
Absolutely. My good friend, Avinash Kaushik, who is a brilliant data strategist and evangelist around all things digital has this concept of the 10/90 rule. The idea is that if you have a budget for data analytics, you should put 10 percent of that budget against tools and 90 percent of that budget against people to operate the tools.
Investing six and seven figures into technology implementations with no one to run that, is akin to buying a sports car and not putting wheels on it. It just doesn't make any sense.
Today, particularly with the amount of data that is out there and being generated, it's irresponsible to put someone at the head of that data who doesn't have the skill needed to really turn that into an insight.
Are there any offline industries that should smartly invest in data analysis?
Any industry where the conversion, the purchase, is still done offline. So you can think of consumer packaged goods. You can think of auto. You could even, although this has changed with COVID, think of theatrical, wherein you’re trying to drive people into a movie theater. Those industries have a greater challenge in acquiring data.
There are companies inside these industries who I would say are way ahead of the curve and those that are way behind the curve. Those that are behind the curve, look at data challenges and think I'm not going to be able to collect data and make sense out of it. And they don't invest in the resourcing that is needed to execute the work.
Or there are companies that face that challenge and don't become trusted partners with people who could help them, like Google, like Facebook. When data is difficult to acquire, you need to become a great partner. Because you're not going to be able to do it all your own. So these partnerships should be parallel to you.
On the moral responsibility of respecting data
Would you say we are in danger because data can be manipulated?
We are at a greater risk of data being manipulated. That is the downside to greater data accessibility and availability. Now everybody has the ability to be an analyst and you can use data for whatever means you really want.
I had a mentor, Michael Foz (sic) who's now the CMO of the city of Chicago. He’s a longtime president of FCB ad agency in Chicago. And Michael used to call it data interrogation, which is to torture the data until it tells us a story we want it to tell us. That was not what we were there to do, but people are able to do that.
So there is a deep responsibility that analysts need to carry. To tell stories that organically come from data that are truthful, not the stories that they want to tell or the stories that their clients want to hear.
There's also a great responsibility for those marketers who are receiving the stories to know enough about data. To be able to identify when something is off and to challenge and ask those questions.
Shouldn’t everybody learn to detect manipulated data?
Absolutely. In my book, Digital Marketing Analytics, I talk about three functional roles of an analyst, and I don't expect a C level executive to play those roles, but they should have an appreciation for the story that data can tell and have an understanding of where the data came from, and how the data was collected.
Also, having an appreciation for bias in data, understanding how bias gets into data and how you account for that. Because there has to be trust. I do believe that it is the responsibility of any executive, anyone who is a customer of data, to understand when something sounds too good to be true, because it almost always is. They should be able to identify those red flags.
It doesn't matter if you have an “analyst” in your title or not, it's really important for everyone to have an appreciation for analytics, because it does impact every aspect of business, every aspect of our lives outside of business.
On the future of Cookies and Privacy
What would you ask a magic crystal ball about the future of data in business?
The measurement infrastructure that we've used literally for decades is under threat, and frankly, very close to being wiped out entirely. And that will change dramatically how we operate. I would ask the crystal ball where we're going, where we'll be in five years.
What exactly is threatening our measurement infrastructure today?
What I'm talking about is the Digital Cookie infrastructure that we've been using since it was invented in 1995 to track consumers to get a sense for how they're interacting with ads. The threat has been the shift to privacy: Apple's introduction of ITP, Mozilla's introduction to ATP. These are things that are wiping up Cookies.
Now Google's commitment in the next two years is phasing out all cookie usage as well in Chrome. That will hamper advertisers’ ability to measure to target and to frankly create personalized web experiences.
Digital advertising today is what keeps Google search-free. The ability to watch YouTube videos for free, Tik Tok, Snap, Facebook, it goes on and on. Those platforms exist today because they are funded by the advertising that marketers use on their platforms.
Are you suggesting that consumers should not shy away from Cookies?
The whole purpose of Cookies is to create better experiences for you as a consumer. The reason we track you is so that we can create ads that we hope are more relevant for you and that you will like.
If you are concerned about advertisers using cookies to identify you, then you can easily block those things, but you'll just wind up seeing more generic ads.
But my hope is that we can reach a solution that gives consumers privacy concerns that they deserve and yet allows advertisers to create personalized experiences for them that they should find more enjoyable.
Will this push towards privacy take us back to a time when ads were irrelevant?
Yeah and the reason that YouTube introduces skippable ads, the reason that we measure and allow consumers to either engage in ads or not, is to put the onus on the advertiser to create a good ad.
At the end of the day, that's what we want the advertiser to create: a good personal experience for the consumer. And without this measurement piece, without the data, we're not able to do that.
On the future of collecting data
Is there a type of data that is currently unmeasurable?
I think that almost everything is measurable.
The one thing that I would like to get more understanding of is the way consumers weigh options, and what makes them choose the things that they choose. I am fascinated by the human condition of why we do the things that we do. And I don't know if we actually ever or always know why we do the things we do.
Many times the reason we did what we did comes from a gut instinct, this part of our brain that has no capacity for language. So it's really difficult for us to measure even though we can get to proxies.
I would love to get deeper into that. I think it has to do with neuro study and understanding more of brain function. I know that our tools are evolving quickly, but, today, that kind of measurement is not available at the scaled level that we would like it to be.
What about access to people’s brain waves and studying their emotions?
I think it would be so much better than where we are today if we had that level of insight. But having people plug themselves in to their computer is probably not the experience they want either.
Something that provides both the physical experience that a consumer wants and yet allows us to collect information that can create these rich, personalized experiences for them, would be really fantastic.
What are the most cutting-edge things today you guys at Google are working on?
We are doing quite a bit around automation. Quite a bit around creating more personalized experiences for consumers by using machine learning and other large scale analytics tools to get a good sense of what distinguishes one person's decisions from another person's.
The thing that I think is really fascinating and is somewhat futuristic is the use of virtual reality and how you're creating experiences around consumers. Then tracking how they interact in those experiences. That's something that gets used sparingly today. I think we'll see much more of that in the future and I hope that Google continues to be a leader in that space.
How will such advanced data collection shape the digital experience?
The idea is that it would just make for a more seamless experience for you as a consumer. The actions that you are taking will be better understood by the platforms and by the marketers and others that can provide you those things.
Think about how Google Search has evolved. There was a time where I would need to conduct several searches before I really got to that thing that I was looking for. Now with some of the Autofill technology, I get a few letters in, and suddenly it knows exactly what I was looking for. That’s that idea of that trajectory that we're on.
On teaching and mentoring analysts
What drives you to teach? Why do you teach?
Having these teaching experiences is a fantastic forcing mechanism for me to stay present and to stay on top of the things that are happening. The idea of staying current enough to stand up in front of classes and appear to know what you're talking about is a tremendous challenge in this industry which is moving so quickly and changing.
Also, you forget sometimes that there's an enormous world out there with people who see things very differently or haven't had the kind of experiences that we might take for granted. In my role in Google you get very much locked into the Google way of thinking.
I get so much value from the interactions with students and hearing them talk about analytics from their perspective. It just makes me so much better in my day job at work. The ability that I have to connect to people that I'm working with, outside of Google, has improved that much more, because I do have that empathy.
Have you seen your students benefit from your teaching?
I literally work with thousands of students at this point. I've seen students go from jobs they were terribly unhappy in, to jobs that they love.
I've seen students come into programs without a job, and leave to go on to places like Google, like Facebook, to other important roles in organizations because of the things that they learned and the programs that they were involved in.
The biggest thing that I take pride in is seeing someone who comes in with this faint idea of how analytics should be used and then seeing them progress into someone who truly understands what it means to be an analyst. That for me is the greatest transformation that I get to witness.
Do you think education is rapidly changing these days?
I feel that the quality of education has improved dramatically. And that corresponds to the amount of attention and participation that we're seeing in classes which are much bigger than they were when we started.
I was lucky to be involved with Coursera when it was just getting off the ground. So in those early days of online education, the programs looked very different from where they are now. And participation is very different from where they are now.
This whole experience with COVID has students, at least in the United States and in most places in the world, away from university and back home into an e-learning environment. I'm sure it’s going to have traditional schools rethinking what it means to be a student.
So, it's changing both in terms of the expectations that students have for the programs as well as the quality of the programs. It's a natural evolution, but with this COVID experience, I don't think it'll ever go back to what it was before.
Some people can naturally “see” and visualize data patterns. Can that gift be acquired?
Data visualization is, without a doubt, a learned and acquired skill. There is a right way to do it and there are so many wrong ways to do it. I've seen people come in with no experience and they've ended programs that I've worked with them on and courses that I've taught. It's because they learn them.
It's really where three functional roles come to play. First, you need to have the technology skills to acquire and manage data. Second, you need to put those numbers into a business context. It’s a learned skill of understanding business strategy. Then finally, you need to be able to communicate those things to others.
Give me someone who has a less natural ability to communicate data but is willing to invest and learn and do things the right way. They will always be more successful than that person who is a natural storyteller yet doesn't approach data visualization the right way.
On the upcoming ELVTR course
What do you want to tell your future students who’ll join your course soon?
There’s never been a time that digital analytics has been more important. And there's also never been a time that digital analytics has been in more of a crossroads.
In this course, you'll expect to get a great understanding of all things digital, all ways that analytics can be used to promote business and marketing and help you make better decisions. I'm just very excited to be able to work with students who are curious and want to learn more about the field.
What is the favorite toolbox that you usually use and that you usually teach?
There are a number of tools that we wind up using through the courses that I teach. Students continue to use them in their professional careers outside of the courses.
It's things as simple as SQL, it’s a great place to start. R is a fantastic program that I advocate for, as well as Python as another option to do deep, rigorous analytics. Then, a tool like Tableau allows you as a data analyst to design images and communicate visually.
But truthfully, I provide more of a framework for you to think about which tools fit which use case and provide you with a set of tools that you could use to solve problems. Then it's up to the analyst to just experience them, get used to them, cast off the tools that they don't wind up liking. And then really lean into the tools that they like very much.
It’s being able to master a tool and see the fruits of your labor. You can take data and convert it into something that is really impactful.