How Data Science Turns Marketing Campaign Analytics Into Strategy
May 21, 2021
4 minute read
Data tells a story. When it comes to digital marketing, brands have the ability to collect an extensive amount of data at various stages of the customer journey. Data science helps us turn this data into actionable insight that results in a greater return on investment (ROI). Campaigns that leverage data-driven personalization report 5-8x ROI for their campaign spend.
How does data science work? Frequence data scientist Megan Duncan stops by the blog to explain how companies can aggregate and measure data to show true campaign performance and turn analytics into strategy into ROI.
Megan, how do you contribute to Frequence?
I lead the data science team, which means I analyze data from marketing campaigns to drive insight and see how we can improve on future campaigns. A data scientist looks to predict the future and makes predictive models using regression (mathematically sorting out which variables have an impact), machine learning, and other advanced statistical methods.
That sounds complicated! Do you have special training?
Yes, many years, actually! I have both a Master’s Degree and a Ph.D. in Statistics. My team has varying degrees, including Ph.D. in Mathematics, Ph.D. in Astrophysics, M.B.A.
So, why do we need data science?
Traditionally, the way to measure the success of a brand’s online marketing has been to measure quantity — clicks, conversions, new visitors, etc. But that doesn’t tell us enough about a customer’s full experience with a brand. We want to understand the quality of the engagement and level of commitment to the brand, too, so we measure video completion rates, view-throughs, click-throughs, viewability, invalid traffic, and site visit lift. By combining quantity and quality, and understanding the relationship between them, we get a deeper understanding of a customers’ entire experience with a brand. We also look at cost and how we can pass savings on to the advertiser – we want to reach a customer by engaging them with an ad at a lower price while getting the same value from their engagement.
What’s different about Frequence in the data science space?
We focus on local advertising. The common challenge with local advertising is you have a smaller sample to learn from. It’s hard to tell based on one small business in one location how another small business in another location is going to act. Other agencies may look at data on a national level and apply it to local campaigns. We stand apart in that we analyze tens of thousands of campaigns, including current and historical, and apply algorithms on top of the national data to see how campaigns perform in aggregate, then use that data to drive insight towards the local business.
How do you know it’s working?
The results speak for themselves, honestly. We’ve developed an algorithm that’s constantly adapting to provide the best quality ads, and we are proud to say that over 99% of our campaigns deliver in full.
How does your team support Frequence’s all-in-one platform?
We contribute to three main areas: proposals, campaign management and operations, and reporting. For proposals, we contribute algorithm development to help predict or recommend budgets. Most of our work falls into campaign management and operations, which is what I described earlier as using data to drive insight for future campaigns. With reporting, we help customize dashboards with specific data or perform ad-hoc reports based on performance insight that’s important to a particular business.
What do you like best about the work you do?
From a data science perspective, I like having access to data that is granular enough that we can pick out very small, very specific things to help campaigns succeed. For example, we can see people were more interested in an ad for a campaign on Website A and less interested in Website B. We can then adjust the campaign to advertise less on Website B and more on Website A. Data tells a story, and it’s fun to read it and put it to work.
The biggest thing I enjoy though is truly my team. Every data scientist on the team has a different background, and we bring all of that knowledge together to not only solve a problem but improve upon it. Leadership works hard to help us figure out ways we can contribute the most, and I value the ownership over my work that gives me.
What do you do when you aren’t data-sciencing?
My husband and I have a new baby at home, so we enjoy having new experiences with him. Right now, solid foods are a huge hit in our house (she laughs). Before the baby, we spent a lot of time exploring on our road bikes. We hope to continue that as our son gets older. But for now, the statistician in me can’t help but make sure he’s learning his numbers.