What 2020 Analytic Trends Will Remain Trendy in 2021?
Not all predictions come true. So it goes with articles predicting trends for the new year.
I wrote about 2020 analytic trends in December 2019. Of course the COVID-19 pandemic upended my predictions — as it did those from analysts in every industry you can imagine. My prediction of the opportunity to extend analytics through DOOH ads was certainly a bust when public confidence on the nation's pandemic response diminished consumer interest in outdoor spaces.
For analytics practitioners the disruption from the COVID-19 pandemic means two things: stop following the measurement playbook, and mark up the playbook text with adjustments accounting for the recovery. If you are adjusting your playbook, consider these trends that are still viable for strengthening your plans.
Opportunities Still Exist to Get Customer Experience Optimization Right
Managing customer experience is still a proven strategic advantage for connecting repeat customers and leads to sales. But Gartner has predicted a coalescence of tech and human behaviors that will require deeper scrutiny to identify the opportunities. Gartner noted the emergence of an "internet of behaviors," a combination of people-centric tech such as facial recognition and data to map where beneficial customer behavior occurs in a sales process. The whitespace for marketers is applying experimental campaigns to their brand's current channels to see how customers will appreciate the benefits. Scott Clark's post on how brands can best reevaluate opportunities to strengthen customer experience offers some suggestions on where to start.
Related Article: Optimize DX – and Get Real – With Customer Experience Management
Know What Ads Drive Customer Spending
Just about every media platform — from podcast to streaming — offers an ad service. During 2020, Hulu introduced an affordable ad service, Spotify bolstered its services to appeal to podcasters, and the leaders in digital ads (Google, Facebook, Amazon) all strengthened their market share. All of this means an overwhelming plethora of choices to engage customers. Marketers will invest in analytics usage to refine ROI from channels and manage an advertising budget being more heavily scrutinized by their executive management.
Better Application of Predictive Analytics to Ecommerce Sales Analysis
Digital ad spending temporarily declined in 2020 but is expected to regain ground in the year ahead. Meanwhile, marketers will encounter new advances for customers to conveniently act on an ad message. Shopping extensions that display offers within a stream, like Instagram Shopping and Stories — a carousel of images or videos on social media — are examples. This means attribution of channel activity to sales will be scrutinized more closely than ever to understand customer activity. This also means predictive analytics is needed to link channel activity to complex behaviors such as customer churn. Google's updated Google Analytics 4, for example, is revised to better emphasize measurement based on customer activity rather than website/app functionality that has to be interpreted.
Privacy Concerns Will Refocus Analytics on Forensic Tasks
I once mentioned that analytics will head back to its origins as a monitor for diagnostics issues. While their original purpose was for website functionality, today's analytics solutions must satisfy privacy governance concerns, such as the recently passed CPRA guidelines. That means you will turn to analytics that complete diagnostic tasks related to data collection, processing, and elimination. Organizing these tasks ensure processes align with data collector responsibilities first introduced when GDPR was enacted.
Related Article: California's CPRA: It's Time to Cut Ties With Old Data
Using DataOps for Data Quality Management
The diagnostic element of privacy will shine a light on data quality management. Programmatic ads will make that spotlight larger. A wider scope of programmatic advertising will require dataop methodologies to ensure that placement quality errors are not systematic and compliance mishaps due to overzealous customer personalization are few.
Data observability, for example, can highlight where monitoring and alerts can best take place. Testing will introduce some developer workflows but overall marketers should experience the blend of dataops and marketing knowledge to best solve data quality challenges.
Paying Down Technical Debt
Martech stack investment will evolve to address privacy protocols and other technological trends such as the discontinuation of third-party cookies. Because of this, discussions on technical debt will certainly rise. Questions like "What is the most cost efficient way to remove outdated solutions?" will push marketers to be better informed. Technical debt left unchecked creates an untenable drag on productivity. Yet the cost for replacing some tech and training analysts can also be prohibitive. Look to work alongside IT teams to identify inefficiencies and highlight recommendations.
Improve Your Storytelling to Improve Your Data Visualizations
Data visualization choices for marketers are broader than ever because techniques have gained a wide scale application. The need for better storytelling will bring more focus into data visualization. This means you must learn how to leverage open-sourced languages, as well as platforms with user interface that differ from traditional analytics. You can better craft stories around your data, letting users avoid into ancillary analysis details.
Simplifying Dashboards Will Simplify Your Workspace
Analysts usually explain metrics relative to a dashboard application - discussing conversion reports in Google Analytics, for example, mean understanding how the solution measures conversions. The pandemic has introduced some complexity in how to communicate dashboard details effectively. Thus analysts will seek solutions that offer simple UI and reporting features that minimize distraction in explaining UI-related jargon.
Sharing Is the Right Remote Workflow
Remote working has altered how analytics teams collaborate on data. To keep your team moving forward, value team dialogue regarding the collaboration on data so that more workflow discoveries and analysis techniques can be shared. For example, if the team creates an programmatic data model that overlooks an important objectives but can provide a cogent argument for why the steps were taken, consider listening to the explanation well before making a decision to reject it or explore how to best use the model.
Related Article: How to Better Communicate Analytics Reports While Working From Home
The Opportunity to Pursue Diversity and Inclusion Through Data
2020 saw an enormous spotlight placed on how diversity and inclusion are managed among corporate ranks. Addressing diversity through data is possible —Microsoft CEO Satya Nadella explained why addressing diversity is vital to machine learning strategy. Yet it can be a struggle to get the right message. Google's high-profile firing of Timnit Gebru, a Black artificial intelligence researcher, for her refusal to retract a research paper that stated AI discriminates against darker-skinned people, demonstrates how initiatives can be adversely impacted. HBCU 20x20, a STEM mentorship network for minority students who attended historically black colleges and universities, discontinued its partnership with Google as a sign of solidarity after a former Google diversity manager added her experiences to Gebru's on Twitter.
Look at opportunities in analytics and machine learning to build meaningfully diverse teams that blend inclusion and soft skills programs. Doing so provides better accuracy and productivity benefits as Harvard Business Review pointed out back in 2016. Moreover, you should expect competitors to strengthen diversity and inclusion as a competitive advantage.
The strategic opportunity for developing machine learning from collected data still exists. But business shutdowns during the pandemic activity paused operations. Operational data is not representative of normal operations — and the nature of "normal" will change as a restart for many machine learning initiatives become based on real-world activity. Yet the reset gives those who are still new to tech subjects time to focus on learning techniques and implement learned applications (David Blankley offers a few great tips on how to get started in this earlier post). Education and data go hand in hand, so expect professional development to include data education every year.