Editor’s note: This article is part one of a multi-part series co-written by DEG Creative Special Projects Manager Maril Hazlett.

Right brain versus left brain. Facts versus emotion, analysis versus creativity, and ugh. Haven’t we all had enough of this old-school, polarized, either/or thinking?

3 Ways AI Impacts Marketers’ Ability to Create Personalized Experiences

Well, we certainly had. “We” consists of Shrey Bhatnagar, DEG, Linked by Isobar insight analyst, and Maril Hazlett, DEG creative special projects manager.

Which is why we created this blog series. Our purpose is to have conversations that explore the common ground between our two teams: analytics and insight and creative.

TL;DR: Cool big ideas, awesome projects, and lots of potential.

Maril Hazlett (MH): Hi, Shrey!

Shrey Bhatnagar (SB): Hey, Maril. Should we start by talking a bit about our backgrounds and experiences?

MH: Perfect. You first.

SB: Ha, thanks. Well, I’m an electrical engineer by background but fell in love with data analytics when I took an advanced marketing analytics course as a part of my MBA. As the course progressed, I started pulling in publicly available datasets for cricket players and ran analyses to perfect my fantasy team. I was having so much fun doing this that when the opportunity for a full-time job at DEG popped up, I grabbed it with both hands and made a run for it.

I still remember the first time we got to work together on a project that primarily involved clients who operated in the highly regulated industry of finance and insurance. We talked at length about the challenges you faced, and I couldn’t wait to help understand the audience better.

MH: And I’m currently on the creative team, after a long, complicated journey through the digital landscape. Also, from a past life, I have a Ph.D. in history, and qualitative analysis is totally my jam. Which is part of why I started constantly bugging you for more, and more, and MORE information on my clients and their target audiences. And you were so awesome and patient about it.

SB: Well, it helped me to know what you needed. The fact that you even brought it up was amazing, because I don’t know that a lot of analysts entirely get what creatives do.

MH: Trust me, that goes both ways. So let me break it down. The creative process is huge—we imagine and develop journeys, stories, branding, content, you name it. My area is technically copywriting, but I end up doing a lot of creative strategy; I think our entire team does. I might start with concepting, get pulled into an ongoing implementation or work on an optimization. Basically, we take business goals and audience analysis from strategists, brand guides from clients, let it all bake in our brains, then boom, we go crazy with awesome ideas.

When I decide what to write, or how to write it, or what kind of creative direction to offer designers, I’m balancing a ton of qualitative factors that require a slightly different take on the numbers. – Maril Hazlett

After the crazy, we reel it back in, focus, and get real. Afterward, in my case, we also read the analytics reports and/or track down the analyst and ask lots of extra questions, to find out more about how our amazing creative performed, for whom, in what channels and contexts, etc. Then I know what to do better the next time.

Now, can you say something about what analysts do?

SB: With how different our outputs are and based on the official nerd status conferred upon analysts, one would imagine that we are very process-driven, and our job involves drawing sharp lines that split things clearly into black and white. This couldn’t be farther from the truth. In reality, that sharp line is very blurry, and we work with our own shades of grey. Strategists from different channels and divisions help us with bringing that line into focus and further defining those gradients.

The beginning of our process, very much like yours, can be best described as organized chaos. We are provided with large dumps of data consisting of multiple datasets from disparate sources. We start off by consulting with the data provider to understand the structure, granularity, and data collection methodologies that were employed to curate the diverse datasets.

This step is followed by a data cleanse to account for anything missing, outliers, and anything funky that might bias our analysis in any particular direction. Beyond this stage the path forward splits into innumerable directions based on the type of data provided and the questions we want answered. Then, we summarize our findings into a report and mark it complete, or so you would think. Actually, this is where the real fun starts.

Learnings from the data analysis only reveals part of the story. Think of it like the plot summary from a movie script. You get an overview of the different roles played by cast members, the sequence of events, who did what, and a gist of their motive. What’s missing is the in-depth reasoning behind the characters’ actions. And that’s precisely why we need creative and strategy. Collaboration helps us with filling in the blanks with invaluable context and answering the most important question of “why?”

MH: That. Is. So. Cool.


MH: It gave me all sorts of ideas and now I want to go off in a really crazy direction, but I think it’s probably best if we give a concrete example of a project we worked on together.

SB: Indeed. Wow. Where do we start?

MH: We could start slow and pick a single channel example, like a mini-case study? And build up to multi-channel in other blogs?

SB: Yes, that.

2019 Super Bowl Spots: Studs, Duds, and Unpopular Opinions

MH: So once upon a time, there was this copywriter, working on paid and organic social for a multinational brand in a very traditional industry. This company was fairly new to translating its brand creative presence for social media, yet it wanted to hit some pretty aggressive KPIs. The company also had identified at least 10 target audiences in varied industries, most of them pretty technical. Even given guidance from strategists, this copywriter—obviously, me—needed to drill way deeper into post categories in order to understand which versions of creative performed the best with each audience. Otherwise, I was just firing darts at the wall, which is inefficient and costly.

However, the kind of questions a creative asks—the clients and strategists don’t necessarily ask them, and creatives don’t always get the results we’re looking for from traditional analytics reports. When I decide what to write, or how to write it, or what kind of creative direction to offer designers, I’m balancing a ton of qualitative factors that require a slightly different take on the numbers. I need to know not just what performed, but why, and I need to look at a larger sample than a few top-performing outliers. I also need to see what did not perform, too.

So I had a problem—and you solved it. Explain how.

SB: Thinking about that project reminds me of the wise words from my marketing analytics professor: “There are lots of complex analyses you can perform, generate predictive models, use machine learning to train AI, but in marketing one of the simplest and most impactful solutions is segmentation and targeting.”

The problem at hand (ignoring all the analytical mumbo jumbo) can be boiled down to increasing user engagement on the social posts we developed for the client.

Like you said earlier, reviewing overall account engagement or even the top performing posts helps little in understanding opportunities for improvement. By no means am I saying that those metrics or showcasing the top post have no value, because they most definitely do. The account metrics aggregated to a monthly or quarterly granularity provide invaluable insights into account health and the top performing post is indicative of current trends in the client’s market.

But, let’s be real, not all social posts are created equal. What may be valuable to me may not be as valuable to you. In today’s digital age the end user’s time is at a premium, where you only have a moment to make your impression and capture their attention. Therefore, optimizing the content and marketing to the moment is key for driving the desired conversion (in our case, engagement).

The beginning of our analytics process, very much like creative, can be best described as organized chaos. We are provided with large dumps of data consisting of multiple datasets from disparate sources. – Shrey Bhatnagar

Following the wise advice from my professor, the first step to this analysis involved breaking down not just our contactable population into distinct segments, but also the content we were putting out for consumption. As we hadn’t employed a targeting strategy with our historical posts, they served as the perfect dataset to set a baseline for average engagement in each post category.

MH: Therefore, you set up this awesome monthly report for me! I am in that data literally every day, searching and analyzing and considering it from every angle. I don’t just limit myself to post categories, either. I look at cross-industry trends and overall storylines. I can zoom in and out on different angles of how we achieve voice and tone, and I can tell where we might need to push the brand a bit, too.

I gotta tell you, my next goal is to find room in the budget to analyze more of the design. For example, lifestyle images versus infographic treatments, how different audiences react, to what extent the reaction may be channel-specific, etc. As long as I have that engagement number, I can consider a number of creative variables that contribute to content performance. Or lack thereof. When I aggregate and examine low performing posts, I can definitely see patterns there, too. Patterns that pain me, yes, but it’s also helpful to know what NOT to do.

Oh…just thought of something. Is there an argument here that since I consult the data to create MORE data, I am actually corrupting the data?

SB: Don’t go there. But I do want to say—back on targeting. These engagement benchmarks we generate do not stay static. With every revision we learn something new and keep building on it, serving as fodder for the next test. The predetermined categories do not stay the same either. For example, if we hit a ceiling with improvements, we’ll go through the process of re-categorization of posts again and the process repeats itself. New trends begin to emerge in terms of cohesion and separation of interactive engagement from different target audiences.

Something to always remember is to keep track of a variable we have little control over, such as seasonality. In certain cases, a winning strategy may seem to falter despite substantial proof of its success in the past and this is where reviewing those high-level account trends is pivotal to checking yourself … before you wreck yourself.

Being Everything to Everyone Is Killing Your Creativity

MH: Ha, ok. And I have a caveat, too. Creatively, you obviously can’t be bound by past performance alone. It’s only a guide. You always need to be ready to take a chance and head out in a totally different direction. The analytics are there to help keep you in touch with your audience, not to constrain you. Constraint equals boring, and no one engages with that.

SB: Actually, let’s make that our conversation for next time. What NOT to do with creative analytics.

MH: Sounds good.

Keep in touch.

Stay up-to-date on the latest digital trends, DEG news, and upcoming events by subscribing to DEG's newsletter.


Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>