Navigating a Pandemic with Predictive Analytics
Can predictive analytics help navigate the media landscape throughout this pandemic? Challenging. Unprecedented. Abnormal. There have been dozens of words used to describe what we have all experienced this year. Being flexible and adapting to the minute-by-minute changes this new way of life has thrown at us has extended far beyond managing Zoom school and working from home: How does an industry that relies on mostly brick-and-mortar sales distribution completely flip their model?
While learning and adapting should be a constant in what we do, it has been even more critical in approaching marketing and advertising in 2020, and that isn’t changing anytime soon. This is especially true for clients whose business relies on in-store shopping. Being nearly nine months into this “new normal,” we’ve been looking back at what we’ve had to do and where we’ve landed and how we’ve walked along this path with our clients.
From the end of Q1 2020 through the start of Q3 2020, our approach and execution had to be agile, adaptable, and flexible as we guided our direct-to-consumer clients who rely heavily on brick-and-mortar sales to a pure ecommerce brand. This shift became the foundation that would inform future strategies and project business outcomes for the remainder of 2020, through the holiday season, and into 2021.
Because of our consultative relationships with our clients and our solutions-focused strategic approach, we’ve seen some clients reporting double- and triple-digit business growth, even during peak lockdown times early on in the pandemic. As retail locations began to reopen in parts of the country, business as usual would have been quite unusual, and we were prepared to address those challenges head-on.
How did we do it?
As most businesses were shifting to a 100% remote working environment, it was critical for us to stay connected with our clients seamlessly, as we couldn’t gather in a room to collaborate, we couldn’t engage in face-to-face meetings, and the daily connections in making business decisions and discussing current trends looked vastly different. And none of this shift in communication could come at the cost of delivering superior outcomes for our clients. As our clients’ business shifted from physical store sales to ecommerce-only, our teams had to also evolve to operate in a digital-only world.
Risk Assessment Re-Establishes Baseline
Leveraging historical performance data from our MarketsmithIQ Trade Area Model, we determined the impact of physical brick-and-mortar locations closing at the individual store level, and how business expectations would shift from retail to web. Each media tactic was evaluated based on ROI impacts, and how each sales channel impacted those ROIs to re-establish performance baselines as we shifted to a web-focused strategy.
Defining Key Business Performance Levers
The first several weeks of the pandemic were most critical in achieving future success. We worked closely with client teams to lay out the potential business levers, from media spend by tactic to impact of COVID-19 by state, all the way to promotional penetration at a localized level. These business levers were then evaluated in near real time over the following four to eight weeks, identifying the core business drivers during the initial shutdown, as well as defining new, improved variables feeding our predictive analytics models.
Evolving Predictive Analytics
Our team of analysts had to reshape the structure by which we were measuring media results, based on the measurement and evaluation of the new performance levers identified through the strategic pivot of our clients’ business models. This extended beyond ecommerce to encompass how performance looked after physical stores began to reopen. Our media reporting evolved, going beyond just media attribution to account for performance trends as markets began to step into a new phase of business as usual. This new foundation of measurement ultimately fed into our data modeling solutions for deeper insights and more accurate predictions.
Incorporating this new structure of media measurement into our media mix modeling tool was critical in accurate modeling for future business success. Continuous evaluation into media budget risk and reward against different business models – ecommerce-only vs. a combined brick-and-mortar/ecommerce model – and the revision of those modeling inputs allow us to view predictive analytics models against different future scenarios as the ever-changing day-to-day life during this pandemic evolves.