The Truth About What Works in Digital Marketing
Does digital media deliver a better return on investment than TV?
Patti Wakeling, global director-media insights at Unilever, neatly sidestepped that query put to her by a Coca-Cola exec during a presentation before the Advertising Research Foundation in March.
It is a good question, but here's an even better one: Does it sell soap at all? That head-scratcher is cropping up more often from major and minor players in packaged goods, a category accustomed to measuring marketing success not by conversation, community-building, virtual transactions or sales leads but by precisely how many packages of Dove went from shelves to shopping carts at Piggly Wiggly.
The answer is yes, but it's taken a long time to divine it.
In 1994, Ed Artzt, who was then CEO of the world's-biggest ad spender, Procter & Gamble Co., called for ad agencies to board the interactive train or be left in the dust. It sounded an alarm. Eighteen years later, it seems P&G and its competitors have hit the snooze button a million times (roughly all the nine-minute intervals from then to now). By the most generous estimates, the all-in costs of digital make up around 10% of P&G's U.S. marketing outlay and a similar share for P&G's big global rivals, Unilever and L'Oréal -- Nos. 2 and 3 in global ad spending, respectively. Digital has its place for them, behind TV and magazines.
It would be easy to look at this and conclude that digital will never play a central role for consumer packaged goods, and that the industry's data-driven executives must have caught on to the fact that digital just doesn't work that well for them, despite much talk of late about how efficient it is .
But the reality is far more complex. Several studies suggest digital sells as well or better than TV for CPG brands. Despite the slow uptake through nearly two decades of hype , packaged goods may now be on the upward slope of a hyperbolic adoption curve, since digital as a percentage of marketing spend has doubled or tripled at P&G, Unilever and L'Oréal over the past few years.
Moreover, digital media appears on the brink of some measurement breakthroughs that could have huge ramifications for data-driven CPG marketers. That data revolution, however, is proceeding on separate tracks with vastly different implications for how digital will be used.
The second track is use of Big Data/advanced analytics to find targets based on purchase or other behavior that suggests purchase potential and then measures the impact of the digital advertising on offline sales (more on that later). This uses digital much like the interactive direct-response medium envisioned by Mr. Artzt and as it's applied by other industries that sell directly to consumers.
A third way of using digital -- to build advocacy or reach consumers at the moment of consideration through search rather than as an awareness builder -- is also taking hold among CPG marketers. But large-scale research into the ROI of these often-social-media approaches isn't widely available.
Detailed multibrand studies of ROI for social media and social-media advertising are still scarce, in part because spending is low and effects occur over a more extended period than with traditional media advertising, so it's even harder to measure a sales response.
Digital "allows us to bring additional reach with a nice combination between traditional TV and digital," said Marc Menesguen, managing director-strategic marketing for L'Oréal. "For the same budget, we can increase the reach of our campaign."
Unilever's Ms. Wakeling, in her presentation at an ARF conference in March, outlined various ways that the company had used Nielsen's Online Campaign Ratings to extend the reach of brand advertising across its target demographics. That said, she didn't do the digital-media industry any favors when she avoided the question from a Coca-Cola executive of whether digital overall delivered better ROI than TV.
Late last year, Unilever Chief Marketing and Communications Officer Keith Weed outlined numerous cases where digital displays and social-media programs had delivered strong ROI of $3 or more per dollar spent.
Bridging the two, Unilever Senior VP-Global Media Luis Di Como frames the issue somewhat differently: Digital isn't really a substitute or extender for TV, but something entirely different."We don't think that TV and digital are comparable," Mr. Di Como said in a statement. "They are two very different media that deliver different benefits for our brands. TV offers a unique scale of reach: The impact that a creative [TV commercial] will have in driving brand awareness and recognition is unsurpassed. What is interesting and valuable about digital -- and social media in particular -- is its unique quality of providing 'social context,' which essentially is an endorsement of our brands from our fans to their friends. Therefore, we see TV and digital as complementary -- not necessarily in terms of reaching different audiences, but rather to engage with our consumers in different ways."
P&G, for its part, is more willing to give credit to digital for delivering better ROI -- though it's not giving up TV anytime soon in the name of reach. Digital may be efficient, but that 's still in large measure because it doesn't have so much scale yet in P&G campaigns.
"The ROI work that we've done so far has demonstrated that digital is better than TV," said P&G Global Brand-Building Officer Marc Pritchard in an interview. "While you can still get very broad reach from TV, that to an extent dilutes some of the ROI." Digital still has "a smaller range of reach, but it's more targeted, and that 's what gives it a higher ROI."
He said P&G will continue to try to drive greater ROI by shifting funding to digital as it becomes more ubiquitous and generates greater reach, but he predicts TV will also get more digital. Reckitt Benckiser a few years ago shifted around 20% of its U.S. TV budget to online video, and this year around 40% of the company's media impressions will be in digital media (though given lower CPMs, that doesn't translate into 40% of spending), said Laurent Faracci, general manager of Reckitt Benckiser USA.
But RB has stopped looking at digital so much as a TV replacement.
"Now we are embracing digital media a different way, which is a space made of conversation," Mr. Faracci said. "We need to understand what are these conversations and how can we be part of them."
Mr. Faracci sees the need for a "common currency" that allows measurement of digital similar to other media, but he said, "You need to be careful in the media space today about seeing everything through the eye of TV. I think looking at digital just as a reach complement … fails to take full advantage of the digital assets that are there. A GRP doesn't give you the measure of engagement, and engagement in digital tends to be higher."
Mr. Briggs' analysis of purchase-intent surveys and spending data suggest that social media provides the best bang for the buck of any medium for CPG and non-CPG advertisers alike. But that bang fizzles pretty quickly -- at levels of around $1 million each for an automotive and food brand he analyzed.
The other big value for social lies outside media: market research. "Insights now more often than not come from digital, from data mining, from social monitoring, from social listening and understanding the conversations that are happening out there," said RB's Mr. Faracci. L'Oréal, which hiked digital spending 45% to around 8% of its global marketing budget last year, is also using digital for more than a TV-reach extender. Digital lets the marketer make a stronger emotional connection, Mr. Menesguen said, and the company has invested heavily in developing microsites to help capture the growing amount of consumer search activity on beauty topics as it uses digital to shift funds from awareness further down the purchase cycle to consideration.
Whatever the role, considerable research does point to digital advertising selling soap, cornflakes, shampoo , etc. A study released last year by ComScore and Dunnhumby USA, based on pooling data from ComScore's U.S. digital-audience panel and loyalty-card purchase data from Dunnhumby across hundreds of brands, showed a 21% average lift in offline sales from digital ad campaigns.
Two years earlier, ComScore and Dunnhumby found a 9% average sales lift for 600-plus CPG campaigns, compared with 8% for TV campaigns across 130 measured over several decades by SymphonyIRI. The digital lift came in three months, vs. 12 months for the TV campaigns.
Digital may work faster in part because it's more targeted, said ComScore Chairman Gian Fulgoni. But the other reason digital worked faster in the study points to another key difference: In CPG, digital is a conduit for price promotion.
"When we look at what's in these digital ads and compare them to television, half the time digital ads mention price or price discounts or coupons or special packs or some other kind of promotion incentive," Mr. Fulgoni said. "And you only get it about 8% of the time for television. If you give people a price incentive, you're going to get a much faster response."
That also helps point to another bucket where CPG media dollars will ultimately flow into digital media -- trade and consumer promotion, Mr. Fulgoni said.
"The average CPG company today is spending somewhere around two-thirds to 70% of its marketing budget on trade promotion," he said. That would mainly cover money deducted from net sales rather than listed as advertising spending for public companies. "The retailer spends that money somehow. Print traditionally has been the medium of choice for communicating price information. And what I think you're seeing is the beginning of the movement of some of that spending to digital."
It's only the beginning, he stressed, given that newspaper free-standing inserts continue to be the primary medium for both coupons and retailer circulars, despite the long decline in newspaper readership.
Microsoft used a predictive-targeting algorithm based on ComScore and Dunnhumby data to target people based on purchase data, something the three companies said doubled the sales lift from digital ads to 42% among people who saw the ads -- even if it lowered the number who saw the ads to a more targeted group.
Big Data's growing ability to track the offline purchase impact on people exposed to online ads is helping to forge the return data loop that seemed so powerful when Mr. Artzt started talking about interactive advertising -- really interactive TV, as he envisioned it in the pre-browser days -- so many years ago.
Catalina Marketing recently joined the growing number of automated trading systems that have sprung up at agency holding companies. These desks buy digital media in real time to target individuals using purchase or other behavior-based criteria, then measure the offline purchases that result.
The idea is to get beyond demographics, contextual targeting and all the other proxies that have long stood in for the real objective for most marketers -- CPG or otherwise -- of selling more stuff.
These systems employ anonymized analytics mashing up a variety of online and offline opt-in databases, including loyalty data from Dunnhumby and Catalina, among others, adding in detailed data from such companies as Acxiom and Experian along with opt-in databases that link them all to individual computers.
As Brian Lesser, CEO of WPP's trading desk, Xaxis, said at a recent Association of National Advertisers Financial Management Conference, such systems can tell when someone who sees a toothpaste ad online buys it offline, and adjust ad buys in real time accordingly.
Of course, for all the analytical progress, it hasn't helped digital that some of its most bold support in packaged goods has been associated with failures. Pepsi Refresh, rolled out with much fanfare two years ago, saw a diversion of TV funds to a program where social media was the centerpiece. In the ensuing period, Pepsi lost its No. 2 soft-drink rank to Diet Coke and most of the executives involved have shifted attention elsewhere. Pepsi has refocused its ad spending on more traditional media.
Marketing Evolution's Mr. Briggs was involved in work on Pepsi Refresh and is limited by nondisclosure agreements on how much he can say about it. But he said the experiment was worth trying.
"Yes, Pepsi Refresh was probably looked at from other marketers with a groan, and fear of making the same mistake in a new medium," Mr. Briggs said. "But I think the better response is to ask: What are we doing to push the envelope and learn from what works and what doesn't? I can guarantee you, Pepsi is better off for having made mistakes and learned from them than if they had not tried the Pepsi Refresh campaign."
While Mr. Briggs didn't comment directly on Pepsi Refresh, his analyses would suggest that the issue wasn't spending on social media -- it was spending too much on social media at the expense of other media more likely to produce results.
P&G Chairman-CEO Bob McDonald stepped up his talk about the efficiency of digital earlier this year just as the company's results were slowing. While the post-hoc argument that shifting money to digital may be to blame, it's worth noting that Unilever and L'Oréal executives have been just as vocal about the power of digital in recent months, while posting accelerating top-line results that beat P&G's.
There are some exceptions to that last statement: Recent Forester research showed beauty to be a rising category for online search by consumers. And bargain-minded consumers have always searched for online coupons and offers for a variety of CPG brands.
"For CPG, digital does still have good effectiveness and value in the decision process, but it's not the most-effective driver," Mr. Briggs said. "You still look at things like television and magazines as pretty effective, and digital plays a role, but it's not the prime driver."
In his spending analysis for a mature CPG food brand compared with a mature automotive brand, Mr. Briggs, as previously stated, found social media delivers the swiftest returns for both even though effectiveness tops out at around $1 million. Cable TV provided the next fastest lift, but effectiveness topped out at around $10 million for CPG and $12 million for auto.
Digital provided a slower lift, along with network TV for both, but they were the only media that showed continuing effectiveness even beyond the $20 million spending mark -- suggesting that big spenders have considerable money to plow into both. Indeed, digital provided the overall highest improvement in purchase-intent scores for the food or auto brands at high spending levels -- albeit stronger results for automotive than food.
While that 's only one among hundreds of brands and categories in CPG, that suggests spending -- and spending big on digital -- can indeed sell soap.