Digital Marketing Analytics: Understanding marketing-specific data
If you have a background in data science, but not marketing, the best place to start is by familiarizing yourself with common digital marketing data sources, types, and terminology. The most common category, known as “clickstream data,” is aggregate data generated from the ways users arrive at and interact with a website, mobile app, or digital marketing campaign.
Clickstream data forms the foundation of digital marketing
analytics. And with the prevalence of integration capabilities between popular
tools, clickstream data is often collected from numerous sources, including
paid digital advertising, organic search campaigns (SEO), social media
platforms, email campaigns, and user activity on websites, with all of these
sources consolidated into a single analytics tool. Simple analysis of this type
of data is called “descriptive analysis,” because it is simply describing what
has happened.
The good news is that a lot of these terms, such as clicks,
click-through rates (CTR), impressions, bounces, and many others, are the same
across channels. A great glossary of these terms can be found here.
Machine Learning:
Move from looking back to planning ahead
Clickstream data is a great introduction to marketing analytics, but machine learning is the future. Machine learning is a type of artificial intelligence that replicates manual data mining, so that a computer program can automatically identify trends and changes, and can even adjust programmed actions accordingly. Having an advanced understanding of machine learning will propel your resume to the top of every employers stack.
A common example is your Facebook feed. When you “like”
posts from certain people more than others, Facebook automatically starts
showing you more posts from that person higher on your feed. This is how
machine learning has moved digital marketing data analysis towards “predictive
analysis,” i.e., predicting what may
happen, as opposed to descriptive analysis summarizing what has happened.
Then there is “prescriptive analysis,” which takes machine
learning to the next level by suggesting potential changes and actionable steps
to produce more favorable results. This is one of the most advanced skills in
marketing analytics. But it requires more than a strong background in data
science – it necessitates knowledge and skills that fall outside of the
umbrella of data science and analytics.
Business Acumen: Knowing
your company’s niche
Mike Dickenson is the CEO of Ironbridge Software, a company that specializes in analytics-based business solutions for consumer packaged goods (CPG) and retail companies. This specificity is important to note: they don’t tout themselves as a business intelligence solution for any-and-all clients. Since their founding in 1989, they’ve kept their focus to what they know. In Dickenson’s own words from an American Marketing Association interview:
"As a data scientist, you have to know your industry. I know statistics, and I know how to produce nice charts. I could move over and work on Wall Street, but I wouldn't know the real tricks in the industry."
Having a deep understanding of the business you’re working
in seems like a no brainer, but the still-nascent understanding that businesses
have of data scientists when hiring them can easily lead to a mismatch of
talent that doesn’t benefit either party. This leads into the next skill,
communication.
Using Effective Visuals: Data analysis is complex. Your job is to make it simple.
As a marketing data scientist, having a solid understanding of the business you’re working in is useless if you aren’t able to communicate your findings effectively. And this doesn’t mean simply having basic good manners and a cordial nature.
One of the most crucial ways data scientists communicate with
the rest of their company is through visualization. The ability to present a
succinct, impactful, and most of all actionable analysis of data, hinges on how
the data is visualized. Can the late-middle-aged CEO with a limited
understanding of marketing or data analysis look at your presentation and
understand which variables are being used, why, what they mean, and how they
translate into an actionable business strategy? Even if the answer is “Maybe,
but I’ll be there presenting to explain it to them in greater detail,” you will
want to go back to the drawing board.
Effective communication of your reports begins with an
advanced understanding of data visualization tools, such as SAS VisualAnalytics, Google Data Studio, Tableau, and many others.
Think Outside Your Toolbox:
Familiarity with other teams’ tools helps guide your reporting
Aside from an expert understanding of the tools you will need as a data scientist, you will also need a basic familiarity with the tools used by your marketing and development teams. This will give you a deeper understanding of past strategies and the data being aggregated on the marketing side, and will enable you to understand the technical limitations on the development side.
For example, a brilliant, data-driven marketing strategy focusing
on mobile video is useless if the development team is currently struggling to
make the website fully responsive, and you were unaware of the fact that emails
are unable to support video. This could mean that your plan is simply not
possible, or perhaps worse, that it requires intensive additional work from the
development team. Then the next time you step up to present, the developers
will be ready to push back at the slightest hint of an incompatible strategy.
Even if your plan is technologically feasible, having an
understanding of basic coding languages like HTML/CSS and JavaScript, as well
as the limitations of your specific web platform and digital marketing tools,
will streamline the process for everyone involved.
No comments:
Post a Comment