Markus, a data scientist at Smartly.io, develops new product features together with our most advanced customers. He gets kicks out of solving complex problems, building easy-to-use solutions, and giving aha moments to people around him.
Why should a data scientist collaborate with customers?
“The best ideas come when you’re talking with customers,” Markus states without hesitation. “My work is about understanding the challenges our customers have with online advertising, and turning them into simple, automated analytics and optimization solutions.“
“At Smartly.io, engineers work closely with the most advanced and demanding customers. We learn from their pain points, validate new ideas, and develop and test new features with them. If a prototype works nicely, we’ll productize it and scale it to all our 400+ customers. One of most motivating aspects of my work is how quickly I can see my work in production and being used by our customers worldwide. I can prototype something quickly and already have customers test it the next day.”
Any examples of how this works in practice?
“Take our Predictive Budget Allocation (PBA) feature for instance. PBA automatically reallocates ad budgets to the best performing ad sets and campaigns. I noticed that many customers were optimising budgets by hand day after day. I did some research, and realized we could make an automated feature that would cut out the need for manual, repetitive work, and give better results in the end.”
“The first version of Predictive Budget Allocation was made about 18 months ago for a customer trial. It worked, so we developed it into a product feature, and scaled it to all customers. We ended up using Bayesian multi-armed bandit methods. Since then, we have expanded Predictive Budget Allocation with new features like automatic budget scaling and ported the implementation from R to Python. PBA is now one of the most loved Smartly.io features.”
Why do you like being close to customers?
“I come from a very technical background, so of course I enjoy all the brain-twisting technical challenges that come by in product development. That being said, I’m getting more and more passionate about working with people. I’m at my happiest when I can first discuss an idea with customers, get them excited, and then go to my Bayesian modelling with R and Python to make it happen.”
“And it’s not just about solving the problem. It’s about making the solution so simple that is easy to use for everyone. There’s little value in a feature that customers don’t understand or can’t use. I wouldn’t personally enjoy building features if they didn’t truly help everyone in practice.”
What does an engineer–customer relationship require to work smoothly?
“It is easy for an engineer to work close to customers when you have a good product that customers actually like. I like working at Smartly.io because we have a very open and candid way of interacting with the customers—we really, want to help our customers and shoulder their burdens. Also, our customers are eager to help us make our product better. They are happy to run tests with us and validate new ideas in cooperation. In my previous experience from product companies, it’s very rare that customers work so closely with engineers.”
“The company culture makes it also very easy to foster a tight feedback loop between customer teams and engineers. At Smartly.io, being a full stack data scientist means that I can develop a feature from mathematics to back end code and all the way to UI that customers interact with.”
"A full stack team also stands for a culture where everyone speaks the same language: customer teams are tech-savvy and understand the product very deeply, while engineers understand the customers, and their business. This culture makes it easy for us to cooperate and get stuff done quickly and smartly.”