AI for Marketing: Regaining Business Traction During a Pandemic

By Kelly McKeown, VP of Reve nue Marketing, conDati. Lincoln Merrihew, Independent Consultant

Businesses are often caught off-guard by major disruptions to their markets, or to the market overall.  Some things can be planned for, some cannot. In both cases the winners are those that can size up–and adapt to–those disruptions the quickest. That’s where Artificial Intelligence (AI) came in for a maker of non-slip shoes that needed to navigate through the fall-out from the COVID-19 pandemic. While this may seem like an uninspiring product category, the trends were dynamic: demand fluctuated wildly as hospitality and restaurant service industries shut down and then re-opened then shut down again in some cases. What was inspiring was how quickly they were able to rebound and the competitive advantage they gained thanks to AI agility. As a result, they quickly pivoted their marketing strategy, leading to much greater ad spend efficiency while simultaneously setting new records for revenue.

The Shift to Digital Marketing

COVID-19 has disrupted the world like few natural disasters have both in breadth and duration. One of the major fallouts was a huge change in consumer spending patterns^[1], which led to a global recession. And because the pandemic lacks any recent baseline it’s been challenging for businesses to predict how, when and if their business will recover. Impacts in the USA vary by State^[2], further disrupting business tactics and requiring regional precision. And the latest spikes of new cases suggest that there could be several waves in the recovery curve.

Independent of starts and stops associated with the recovery is the extent to which consumers return to pre-recession products, categories and brands (also known as “consumer elasticity.”^[3]) That means that even as quick service restaurants and other businesses that need non-slip shoes re-open, there is no guarantee on the volume of business they’ll get, making demand planning even more difficult.

All in all, COVID-19 one of the most unpredictable and diverse disruptions in business ever. The only confirmed variable is that–with stores closed and consumers quarantining–companies turned more than ever to digital to drive business.^[4] Therein lies a tremendous and unique opportunity to gain market share with tools that provide agility and the right insights.

Analytics-Forward Companies are Best Positioned to Thrive

Companies that will gain share will do so by leveraging science and analytics. But not all companies will have that foresight. We see businesses falling into one of three groups:

1. Hope and Pray: These are companies that are generally risk averse. They’ll wait on the sidelines or perhaps play follow the leader. Or get left by the wayside.
2. Risky Business: Here we have businesses that take chances but ill-informed ones. That may be the result of no data, stale data, incomplete data, or mistaken cause-and-effect connections.
3. Strategic: These are the fast movers that will chart fluid and dynamic responses by leveraging intelligence layered over rich datasets, moving from data to insights to action.

Leveraging Digital Data to Drive Strategy

Kelly McKeown

The strategic companies, like our non-slip shoe manufacturer, are positioning themselves to not only survive but thrive in these times. A significant portion of their marketing budget goes toward their digital presence – paid media and providing a great web experience. These companies unify their web analytics, ad spend, and revenue data into a single data asset that then has advanced marketing analytics applied to it. In this case, they used conDati’s RevenueLift™ platform – providing data unification, visualization and application of AI and machine learning in one solution. With accurate, complete, granular, well-organized, real-time data at their fingertips they have the confidence to make smart decisions at high (e.g. channel) and low levels (e.g. micro-segments such as geography and income within ad groups) and speed up decision cycles.

Getting off on the Right Foot

Even before the pandemic, the non-slip shoe company had already successfully used the conDati RevenueLift AI engine to optimize it’s spend on search. Like many companies, as soon as the pandemic really hit they cut back sharply on spending.  Search volume immediately collapsed as did revenue. They knew they had to act so initiated a two-phased recovery strategy with the help of conDati.

Phase 1 began about a month in, and entailed spending much less than normal but focusing heavily on campaigns with the greatest ROI. That also included learning more about how, when, and where the market reacted to their efforts.

Phase 2 was where it got interesting: As the economy began to show signs of recovery, they applied the learnings of Phase 1: Rather than just restore search spend to prior levels they focused on the paid marketing campaigns that generated a 4x or more return on ad spend (ROAS), and doubled down on campaigns with ROAS above 10x. They dropped the underperformers. They gained these insights using conDati’s RevenueLift Optimizer.

RevenueLift Optimizer’s multi-step algorithm approach focuses on diminishing marginal returns and the statistical significance of conversion rates. It does this by analyzing targeting variables such as demographics, geographics, timing, creative, and device across campaigns and combines that with saturation levels for every ad group. That many variables means there may be 100s of thousands of combinations—far too many to manage and action without AI like RevenueLift. And this processing happens in hours or days, not weeks or months.

The output is a recommended set of high potential microsegments that will generate the greatest return as well as clear bid adjustments to capitalize on those. The recommended bid adjustments reallocate spend from lowest to highest potential microsegments without the need to increase spend. Clients–like the non-slip shoe company—easily upload the recommended bid adjustments into their ad platforms in minutes. AI also means new optimization recommendations can be activated every few days. This means near-real-time responses to intelligence and the greatest opportunity to dynamically drive revenue lift.

Lincoln Merrihew

So what happened?  Client revenue increased in both phases with accelerated growth from Phase 1 to Phase 2, and rebounded incredibly in less than 60 days. In fact, revenue reached a 2020 peak and was up 20% year-over-year in June. They did that following a basic construct: do as much as possible with AI alone (i.e., optimizing without spending more on ads), saving time and money. You only up marketing spend when the marginal returns on optimization alone start to show.

AI Every Day

AI’s best application is for solving complex problems and when the volume of information available is simply too much for standard tools to process. Unfortunately, COVID-19 is likely to be the AI poster child for months to come. But it’s not just natural disasters; it’s everyday business turmoil: The business environment is constantly being disrupted by new players, new platforms, updated regulations and changing consumer tastes. AI helps you ride out any storm giving you confidence to make decisions and action on them quickly to adapt to whatever comes your way quickly and efficiently.

Interested in hearing the success story directly from the non-slip shoe company? Watch the webcast.

^[1]https://www.whitehouse.gov/articles/depth-look-covid-19s-early-effects-consumer-spending-gdp/

^[2]https://tracktherecovery.org/

^[3]https://www.quirks.com/articles/unpacking-consumer-elasticity-in-a-covid-19-world

^[4]www.nbcmiami.com/news/coronavirus/consumer-spending-is-already-rebounding-in-these-states/2239346/