Demand for data scientists continues to outstrip supply, with the field topping LinkedIn’s list of “Most Promising Jobs of 2019” (up from ninth last year), thanks to 4,000+ openings (+56% YoY) and a median base salary of $130,000. Likewise, Gartner research (available to clients) states marketing analytics now accounts for the largest share of the marketing expense budget at 9.2%, contributing to intensifying demand for high-quality analytics talent.
Gartner’s 2018 Hype Cycle for Data Science and Machine Learning (available to clients) states 77% of senior managers view data science as delivering significant value or as being essential to the success of their organization. However, much of the current and future workforce lacks the skillset employers need to fill data scientist positions. Additionally, research suggests the effect of marketing analytics on company-wide performance over the past five years has failed to live up to expectations.
Is it a fool’s errand to try to hire data scientists who don’t exist and invest in marketing analytics despite stagnating performance?
To overcome the issues facing data science and analytics in marketing we must clarify answers to three questions.
Gartner research (available to clients) provides a useful framework for understanding the core responsibilities of a marketing data scientist:
Measurement: Determining the impact of marketing efforts and ad campaigns.
Optimization: Recommending changes in tactics or spending to improve results.
Experiments: Designing and executing tests to isolate causes.
Segmentation: Identifying groups and subgroups of customers and prospects.
Predictive modeling: Building computer models to improve response rates.
Storytelling: Communicating messages derived from data to inspire better decisions.
If your business is looking to hire a data scientist, use this framework to scope out the job position. As The Next Web points out, business success or failure can depend on how well a company interprets and acts on its data. However, a shallow understanding of what exactly data scientists do, coupled with analytics becoming mandatory for a variety of business functions (not just marketing), has led to companies filling seats with people lacking education or experience in the field.
In the long term, this quick fix might actually work: people with questionable qualifications will fill data scientist positions, and some of them will become experts over time as they learn on the job. Forbes estimates that by 2029 data scientists jobs will no longer exist. Rather than the current hiring frenzy being a flash in the pan, data science expertise will increasingly complement communication, domain knowledge, and business-strategy skills—becoming an indispensable requirement for productive workers across departments.
In the short term, this may cause headaches as people without data science and analytics expertise fill positions and then inevitably make mistakes, fail to meet employer expectations, or both. Understanding what the job really demands and your own company’s ability to nurture someone new to the field is essential to successfully hiring data scientists and analysts.
Marketing analytics success begins with clean and reliable data. According to the Harvard Business Review, a shortage of experienced data scientists is compounded by vast amounts of messy, inscrutable data. Too often, marketing analytics efforts are hampered by large, poorly organized data sets that are difficult for analysts to extract insights from.
Most companies collect data across departments using different systems and by defining different variables. This makes it difficult to easily combine and analyze company data holistically. Businesses must have a plan for integrating data prior to collecting it, or grapple with the costly and time-consuming processing of retroactively standardizing data for analysis.
Accepting that more isn’t always better from the beginning can help mitigate these issues. Rather than capture everything, determine your business strategy first then decide what data is required to monitor its success.
Additionally, create a detailed customer journey map that outlines every potential engagement point between your business and its customers. Tying data to customer touch points will give marketing analytics the context it requires to inform business strategy.
Abundant marketing analytics software choices can easily become overwhelming. Rather than get caught up in an unending platform comparison, aim to select a single tool that will enable the core responsibilities of a marketing data scientist (even if you haven’t hired one yet).
As you evaluate products, the following questions can help determine which marketing analytics software has the right features for your business:
Measurement: What metrics does my company use to measure marketing campaign success (e.g. click-through rate, bounce rate and page views)? Does this marketing analytics tool allow me to track our success metrics?
Optimization: A data scientist or analyst will be responsible for recommending changes to marketing strategy however, software should provide tools for measurement, experiments and predictive modeling to help facilitate this process. If you anticipate finding a full-time data scientist will be difficult for your business, ask vendors: does your software offer? This emerging technology attempts to automate insight discovery.
Experiments: Does this marketing platform offer A/B testing to ensure the best version of a campaign receives investment? Will this solution allow me to test all the campaign types my business will be running (e.g. email, landing pages, CTAs)?
Segmentation: Will this marketing analytics tool help me define and target audience segments to improve the ROI of my campaigns? How are audience segments created and maintained? Does this platform offer dynamic lists that are updated depending on predefined rules as prospect data changes?
Predictive modeling: Does this platform offer predictive modeling to help my business prioritize contacts based on their likelihood of turning from prospect to customer? Can this tool use predictive marketing to provide users with personalized experiences (e.g. content, offers, pricing) that increase the likelihood of conversion?
Storytelling: Does this marketing analytics tool offer data visualization and reporting features to help communicate messages buried in data to less technically minded stakeholders? Are reports and dashboards generated automatically or do they require manual configuration?
As is the case with data, more software isn’t necessarily better. The questions above can help you determine which marketing analytics software features are the most useful for your specific business needs. Additionally, if you’re able to successfully onboard a qualified (or not so qualified) data scientist, leverage their expertise and preferences to help inform your software purchasing decision.
Ultimately, you should strive to run a single unified marketing analytics platform and maintain a clean database that has been optimized for analysis. If you’re ready to enable marketing analytics using software, try GetApp’s marketing analytics software comparison tool for help finding a platform that suits your business’s data science needs.