Welcome to The 9 To 5, our ongoing series where we highlight and interview creative professionals doing great things. Today’s profile is on Dmytro Marushkevych, Director of Analytics & Optimization at Rosetta.
1. How did you end up focusing specifically on the data aspect of marketing as a career path?
I have graduated college with an MSc in Theoretical Physics. Have not worked a single day as a physicist as I have quickly realized that there is really no middle ground in Physics: you either are a brilliant scientist (think Steven Hawking) or teach kids. I did not have the brilliance of Stephen Hawking, or even Sheldon Cooper from “The Big Bang Theory”, and dreaded teaching. I have ended up going into product marketing in Telecom, where knowledge of Math coupled with the way Physics teaches one to think proved to be very useful.
Data does not lie. Interpretations do. The most common mistake is looking at the data without understanding the context, and the business need.
Working in the product marketing, I have always wanted to find out how successful are my campaigns in driving conversions, how I can make my communications more relevant and how I can improve the campaign performance. Well, there was no one who could help me answer these questions, so I got access to the data warehouse, remembered statistics and learned SQL. This is how my career in analytics begun, and I never looked back.
2. How in-depth can your data and analytics at Rosetta go in helping you understand who customers are and what customers do?
This is actually the core of Rosetta’s offering. We start with collecting all the data to understand the customers, their needs, attitudes, and behaviors. This understanding is then translated into insights, communication plans, data-driven targeting and messaging. Our Personalization Engine allows us to customize all aspects of communication in order to create long term relationship with customers. This engine had been originally developed for Activision, the company behind the popular Call of Duty video game. At the time, it let us take a very deep look to identify the different types of gamers (from ‘noobs’ to ‘vets’) and behavioral patterns, which helped increase in-game purchases or prevent churn. This deep understanding, used in targeted marketing campaigns, has quickly paid off the time and money invested.
3. Does someone who aspires to be in your position need to have a good knowledge of math, or more of an understanding of how to interpret data sets?
I have heard a story about Sir Isaac Newton and his time in the English Parliament. There, he was known for keeping quiet and only speaking out twice: once to ask to close the window, and second time during the discussion whether extra funds should be used to develop more language or math education. He raised up and said: ‘Math IS the language”. Not sure if this is a true story, but fully agree with the point: Math is the language. You cannot interpret data sets if you do not speak the language. The most important thing for aspiring data-driven marketer or data scientist is the ability to take a business problem, translate it into a set of data questions (including what data I need, what I do with it, etc.) and feed the answers and insights into the marketing campaigns. Actually, a lot of a marketing analyst’s work is actually a story-telling, with numbers, charts and infographics.
4. What are the best ways that the everyday marketer—the marketing coordinator or manager at a startup—can use data (through sites like Google Analytics) to their advantage?
Analytical data (whether web analytics, reports from your Email Service Provider or insights from your customer data warehouse) builds the intimate understanding of your customers and helps to identify good or bad ones (yes, there are bad customers who cost you more than they bring). Such understanding drives better and more relevant marketing campaigns, optimized over time. There are zillions examples, so, let us use already mentioned Google Analytics. There, one would start with looking where customers are coming from (search engines, your marketing campaigns or referrals) and what are the difference in their conversions – so you can try to increase the better performing traffic (say, investing more in Google AdWords or building the email database – whichever brings the best return). Looking at how visitors engage with your site, how they go through the pages and where they drop off will help you to improve your site. Lack of engagement, e.g. high bounce rate is a sign of trouble. The way how Avinash Kaushik, well known Google evangelist, puts it: “I came, I puked, I left”.
5. Are there any common mistakes you see marketers make when it comes to either leveraging data in marketing, or even interpreting data?
Data does not lie. Interpretations do. The most common mistake is looking at the data without understanding the context, and the business need.
To illustrate (not invented, just stolen by me), imagine you are a turkey. Very geeky turkey, so you keep a journal of what is happening to you every day (how well you are fed, etc.). If you look at all the past data and run a predictive model in June, everything looks very nice and promising. And this is exactly the outcome in July, August, and September. Until comes Thanksgiving….
6. What do you know now that you wish someone might have explained to you before entering your position or industry?
Do not blindly trust the data. Always take the time to understand how it is collected, all the limitations, how it is manipulated and what does it mean. This will save you from wasting a lot of time down the road and, more importantly, developing wrong insights and making bad business decisions.
7. What is one piece of advice you have for any young data-driven marketer about to graduate from university?
Good technical basis is important. You should not be a system architect, but good understanding of relational database concepts (SQL), best practices of data visualization (whether it is Excel or Tableau), statistical concepts and various marketing platforms out there (from Email Service Providers to Ad Servers to Data Management Platforms) will go a long way in ensuring successful employability.
Still, the most important skill is translating business problems into data and data insights into marketing campaigns and business solutions.