Tim Hoolihan
Name: Tim Hoolihan
Current Job: Centric Consulting: Cleveland Practice Lead, Data & Analytics
Favorite restaurant in town? Cleveland: Mabel’s BBQ, Akron: Ken Stewart’s Lodge
Favorite thing about Cleveland? The park systems we have are amazing, I’m a fan of the towpath in particular
Q: Tim, you just joined Centric Consulting 7 months ago as their practice lead for data and analytics. What does that role entail? Centric is a national consulting firm that focuses on local presence. I work for the Cleveland Business Unit, which provides services for a variety of public and private organizations in Northeast Ohio. However, we have access to consultants on the national teams that may travel in or work virtually. This helps Centric provide a variety of expertise with flexibility and scale. My job is to build out a practice locally that specializes in data analysis, visualization, and data science for organizations in the area. Our work ranges from data definition documents to data warehouses, to machine learning and more.
That role starts with meeting with clients to understand their opportunities and define what kind of project we can collaborate on. The next responsibility is to build out the right team to complete the project. Finally, I work with the team and project manager to ensure each project stays on track.
I enjoy the variety of clients and projects we work with, and the chance to collaborate with so many great Cleveland organizations. Data Science is an evolving field, and there are always new things to learn.
Q: Before Centric, you were at a product company called DialogTech, what was your role there? For background, DialogTech helps marketers track conversations across channels. For considered purchases (like auto sales, real estate, or legal services) it is common for consumers to research online but then move to telephone research before completing the sale. DialogTech’s product suite helped marketers tie phone calls to online marketing efforts.
My role was to build a data science department. Up until that point, the product suite was known for tracking users across channels. But the journey stopped there. It would be like Amazon tracking it’s marketing from search engine to an amazon page, but then ignoring whether a customer actually purchased or not. The data science team built models that helped close the last bit of the loop. These were deep learning models, using tensorflow, keras, pytorch, and other related technologies.
I assisted the team in a variety of ways, including some of the technical work and prototyping before we had fully hired the team. However, I focused on strategic things that helped leverage the smart women and men of the team. For example, since most of our Machine Learning models were supervised learning, I spent a lot of time and effort building out a team and process for part-time workers to listen to and label calls. This allowed our modelers to focus on the problem, rather than collecting data. I also worked closely with our product department to make sure we were aligning our models with customer’s needs.
Q: What are some of the trends you are seeing in the big data space? Cloud migration is a trend across IT disciplines, but I think it is emphasized even more in analytics. Building an on-premises big data solution these days is fairly difficult to justify.
In addition, I would say as many organizations are stepping back from Hadoop, we are seeing new approaches to big data. In other words, hybrid approaches using the right tool for the right job. Removing structure from data that is already relational is probably not a great idea. But neither is forcing raw data into a relational structure just to be stored. For example, rarely should you store images in a database.
Finally, I would point out solutions like Snowflake are gaining popularity. Tools like that use some of the clustering techniques of Hadoop under the hood, but provide an interface familiar to those who come from a relational world.
Q: Many companies are constantly collecting data, but not very good at utilizing it properly. What are the most common mistakes you see from companies when it pertains to utilizing their data sets to the best of their ability? The first problem I see is that data solutions are built starting with engineering. A developer surveys what data is available and starts creating solutions or reports. As any good UX person would tell you, ask instead what does the end-user need? How often is this insight useful? Then work backward from there to build an appropriate solution that will create value.
The next problem I see is incorrectly capturing data. This problem can manifest itself in many ways, but I’ll highlight a few examples. We find clients that track a health score for their accounts. But they don’t capture those over time. If you don’t retain historical data, you can’t analyze trends. Another example is some consistent definition of subjective measurements. Back to our client health score, what does that mean? And does each account manager calculate it the same? Better yet, if it’s calculated based on a few qualifying factors (number of employees, size of the account, recency of contact), why not capture those factors directly and automate the calculation? Your information and insights are limited by the quality of your data.
Finally, I’d like to comment on the term “data-driven”. It has become a bit cliche. The intent is good, but I think a mature organization is “data-informed”. Models aren’t perfect, and they may be missing context. I like to illustrate this with a narrative of two companies that have SaaS products that are tracking page usage statistics to decide which features to focus on and keep. The “data-driven” organization cuts the dashboard that users only use the first week of onboarding and fall off ever using. The “data-informed” organization talked to their Sales department and found out that a dashboard is key to 95% of sales demos and made an exception to keep that page, despite the low metrics. Human insight and context still have value in making decisions.
Q: What is it like working for a virtual company? How does that impact your time at home? While I’ve had the ability to work from home before, this is my first time remote full-time, as we don’t have offices. There are some great perks that come with that. I started the week my kids went back to school, and it was easy to step away for 10 minutes to see them off to the bus on their first day. My dog, a Boston Terrier named Bogey, can come visit my office anytime he wants. Many days, I’ll go for a walk or run over lunch. That said it comes with challenges. It can be hard to “turn off” work at the end of the day. And our company has to be intentional about getting time face to face, etc. However, as I do this interview, America is dealing with the COVID-19 pandemic. I consider myself lucky that I was already adjusted to work from home and do not have some of the child-care challenges others are facing with schools closed, etc. Centric has been very supportive of employees through this, and supporting each other is really part of our company DNA.
In better times, I like that occasional working from coffee shops and finding other ways to stay in touch with the Cleveland tech community. It’s a great community. I appreciate the chance to be a part of this interview as one of the many ways Cleveland works together. Thanks for keeping the conversation going in Northeast Ohio, Ari!