Foundational changes are underway at PepsiCo as the global food and beverage giant tries to account for a year laden with disruptions to its business. Increasingly central to the marketer’s strategy for thriving during and after the COVID-19 pandemic is an in-house technology called ROI Engine, a machine learning-powered measurement system that PepsiCo says allows it to make smarter decisions in measuring campaign effectiveness and return-on-investment (ROI), hence the name.
The company behind household names like Pepsi, Doritos and Gatorade has in recent months accelerated a shift away from linear TV toward streaming, while expanding rapidly into online sales channels that are becoming enshrined in consumer shopping habits by the pandemic, particularly direct-to-consumer (DTC) offerings. These are areas the marketer was investing in well before COVID-19 decimated most in-person venues, including sports stadiums and restaurants, where PepsiCo sells many of its products. Still, the extent to which the company has reallocated resources to digital since March is noteworthy and stands to transform its business in significant ways even after the health crisis ends.
“This is what enabled us to confidently shift investments away from linear TV into more addressable mediums. We saw the bump in ROI,” said Shyam Venugopal, senior vice president of global media and commercial capabilities at PepsiCo, in reference to ROI Engine. “This has allowed us to take a really different approach to our upfronts. This has significantly scaled up our addressable mix in our portfolio as we go into 2021 across every single one of our brands.”
As packaged goods marketers looming and a proliferating number of retail media networks on which to advertise, PepsiCo is propping up ROI Engine as a tool that will let it nimbly navigate a fast-changing media landscape. While the solution is now largely centered on building a more efficient marketing model, executives said they hope it will eventually guide everything from in-store activations to product formulation in another signal that CPGs are aligning future bets around data-driven capabilities.
“Ultimately, our vision is, how do you leverage this engine to drive all of our commercial decisions, whether that is marketing pricing, in-store activation or go-to-market strategy,” Venugopal said. “We don’t see this being just a marketing tool; we see this being a complete enterprise tool guiding all of our commercial investment decisions.”
Keeping pace with COVID-19
ROI Engine’s genesis predates the pandemic, with the idea around the technology first emerging a year and a half ago amid a broader reassessment of PepsiCo’s marketing measurement, per Venugopal. The move to rework measurement and place a bigger spotlight on in-house technology followed years of growing internal investments at PepsiCo, as well as complementary hiring initiatives in specialties like data science and software development.
“We had measurement systems that were primarily outdated. Things ran maybe at a national level once a year [and were] primarily outsourced and even inconsistent across PepsiCo,” Venugopal said. “The first thing we needed to start figuring out is how do we build a consistent measurement approach globally so we can leverage the same tool, the same capability to make consistent investment decisions.”
The pandemic has reinforced the utility a holistic internal data engine can provide, executives said, as PepsiCo — like others in the CPG business — had to move much of its operations online on a truncated timetable due to COVID-19. In May, the marketer to meet consumer demand for snacks and shelf-stable goods that can be purchased from the safety of home. The websites — PantryShop.com, which offers top-selling pantry staples like Quaker oats, and Snacks.com, which is largely for Frito-Lay products — represent the types of experiments ROI Engine has enabled PepsiCo to put together more quickly.
“The shifts to e-comm broadly have been accelerated by our ROI Engine findings,” Venugopal said. “We were already shifting and this has significantly accelerated those [bets].”
In terms of bolstering marketing efficiency, ROI Engine has already saved the organization millions of dollars, executives said, without breaking out specific figures. A key factor helping to cut back on costs is the speed at which the solution operates, allowing PepsiCo to read the results of a campaign that’s still in-market and make adjustments to drive more tangible impact on sales rather than having to wait months to conduct a post mortem.
“A few weeks in, you’re actually starting to see outcome signals that our marketers can then leverage to say, am I seeing the real commercial outcomes out of this particular campaign?” Venugopal said.
“It also sounds pretty simple, because in a DTC world, this is exactly what happens on a regular basis. But in our world where most of our sales still happen through traditional retail terms, this is fairly revolutionary,” he added. “When you make even simple optimization decisions that make things 5-15% better, that translates to multi hundred-million dollar outcomes for the organization.”
Making the leap
ROI Engine is now up and running across PepsiCo’s North American division, with pilots in Europe. The marketer aims to have the solution fully operational across another top 18 markets in the U.S., Europe, Latin America and Asia by the end of 2021, according to Venugopal. Given that PepsiCo spends so much on advertising and promotions — , according to Statista findings — transformations spurred by the ROI Engine carry considerable implications for players outside company walls.
The marketer this year has recalibrated its TV advertising and prioritized streaming ahead of the holidays, conducting pilot programs with platforms like Roku. PepsiCo marketers for a long time held an “informed gut feeling” that TV wasn’t providing the same returns as digital, according to Venugopal, and ROI Engine added new credibility to those claims and the confidence to actually make the shift away from linear media, which has historically been a bedrock for many of its brands.
“If you looked at our learnings let’s say back in late 2018, someone might have said … I think digital should be working but digital ROIs are way poorer than TV ROIs. That was sort of the truth that we were living in,” Venugopal said. “Actually, outside of a few specific campaigns here or there, our digital ROIs are way better than some of our linear TV ROIs.”
Other high spenders in the CPG category, like Procter & Gamble, have in a bid for more leverage in a TV market that might not recover from COVID-19. However, in putting some of its traditional media buys in the rearview, PepsiCo sets its sights on a digital ecosystem that’s also rife with challenges.
CPGs are contending with new media networks from retailers like Walmart, Target and Kroger. These fledgling offerings provide an alternative to the established duopoly of Google and Facebook, while carrying over that earned those platforms the “walled garden” moniker. ROI Engine helps to level the playing field in negotiating with the media networks, PepsiCo executives said, because the company has a much clearer sense of how campaigns should ideally be performing.
“By having an engine like this that’s trying to get apples-to-apples comparisons of different media channels, it gives us the ability to talk to a retailer and say, if you want us to invest in your media platform, this is the type of return we expect, this is the kind of performance we expect,” said Michal Geller, senior vice president of global e-commerce marketing and product at PepsiCo. “[We] can have that kind of conversation in a really clear way — not opinion-driven, but very much fact-based.”
Even as it represents an ambitious wager on in-house technology, ROI Engine hasn’t deprecated PepsiCo’s work with external agencies, per Venugopal. At least on the media side of the business, the executive said the relationships have become stronger due to the more transparent data tradeoff and the ability to keep a closer ear to the ground on campaigns.
“Back in the day, both us and the partners were dealing with a measurement system that was running once a year, maybe once every two years,” Venugopal said. “What we have [now] is more of a granular and always-on measurement ecosystem that’s enabling both us and the partners to make better choices and a more adaptive, agile set of choices from an investment perspective.
“No longer do we have to wait an entire year to figure out what happened,” he added.
Leave a Reply