“Data-Driven Thinking” is written by members of the media community and contains fresh ideas on the digital revolution in media.
Today’s column is written by Adam Heimlich, senior vice president of media at GALE Partners.
There’s a crisis of trust in the market for digital media services. Advertisers are finding ways to pay their agencies less, often by reducing their scope. But it’s worse than that.
The digital marketers who work in-house at big brands are also seeing their budgets reduced, redirected or eliminated. There’s no question that the prospects advertisers want to reach are online, addicted to mobile devices and streaming tons of video. Yet at the C-level, the burning question is, often, whether dollars aren’t better spent in broadcast TV and CRM.
Management consultants, software companies and financial analysts are intensely curious about this crisis, as it suggests tremendous opportunity. Media-buying services are a multibillion-dollar industry. Where will this work go?
It will go to whomever can earn trust as they buy media. But trust isn’t a simple matter of proving, over time, to be an honest, ethical individual within a reliable organization. Not in this case. The same crisis affects all players – from the triopoly, to the holding companies, brands’ digital teams, the tech platforms and every breed of strategist for hire. Our advice is assumed to be self-serving. It’s not like this in other professional services, and it didn’t used to be like this in media.
What have we done to deserve such distrust? I think it’s a growing pain. Digital used to be hot, but now we have to confront the fact that we invest without a model. We can’t keep operating without a verified claim to predictability.
Here’s an analogy: In the early days of catalogs, you could convince a retailer to print and send some because hey, folks order from catalogs nowadays. But eventually the direct mail model made cataloging much less risky. TV, too, was no doubt a special case for years before branding theory made it the second form of investment with predictable results.
Digital investments are absolutely not predictable according to the DR or the branding model, though plans are presented as if they were. Everybody whose money is being spent is aware, at some level, that something is very wrong. It’s like the first reel of a horror movie: They can’t say why, but they know they must get out.
Digital marketers, mostly young and poorly trained, are not in a position to say The Emperor Has No Clothes, because they don’t even know – to extend the metaphor – what clothes are. They’ve never been close to proper DR or branding. You can’t prove the sort of fiduciary responsibility digital marketers need to demonstrate without some grounding in how media buyers earned trust in the first place.
Here’s why we need to abandon the DR and branding models like a sinking ship:
DR means the target responded directly to the ad. Before digital, DR buys were always based on either a list (direct mail or telemarketing) or a strict geographic segmentation (in the case of call-driving radio and TV ads). So, the response could always be measured by comparing actions among people who were and weren’t exposed to the ad. Reducing spend to the maximum level driving profitable incremental response is the DR model.
Search marketing corrupted the DR model by taking the list of users who searched relevant keywords out of buyers’ hands. It’s hard as hell to isolate the effect of search ads on site or store visits. This is by design. Most display and social sellers copied it. Because when advertisers can’t ascertain who is and isn’t responding to their ads, they’ll play along as if everyone is responding, until they wise up.
Today, uncertainty prevails. A consultant can walk into any top consumer brand and see reports showing ROI of $5, $10 or $15 per dollar spent on digital ads. Ask why more isn’t invested in such a reliable, money-making machine, and you’ll hear that no one actually believes those numbers, and though we actually have no idea what returns we realize on our “DR” investment, we’re confident it’s positive because the possibility it’s negative is too disturbing to contemplate. (I humbly suggest it’s not as disturbing to contemplate as epidemic mistrust in the industry to which one owes one’s livelihood.)
In many cases the practice of digital DR media-buying is more like a video game than responsible investing. Any advertiser can investigate this by asking their hands-on-keyboard team what they did today, and why. You may find buyers spend as much as they can on tactics they know lack incremental value (brand search, remarketing) to justify a relatively small and scattered investment in what they believe to be effective (non-brand, prospecting) to end up with a blended “ROI” near their goal. It means hand-raisers get pummeled with ads while everybody else – as many as possible – get only enough exposure (ideally one ad) to secure credit if they happen to convert. Traditional DR buyers have found that convincing people typically takes five to 10 exposures to an offer. There’s no incentive to show people five to 10 ads in digital DR.
If this sounds corrupt to you, I would bet you’ve never investigated a large advertiser’s digital “awareness” buy, which dwarfs digital DR in both waste and sloppiness.
Branding investment strategies hinge on aligning with famous content. The ideal is to sponsor yet-to-be-famous content at upfront rates, reaping a cost-per-reach benefit if the show is a hit.
But an upfront strategy makes close to zero sense in digital, where you can measure cost per reach in real time every day. The bulk of digital branding dollars – the deep end of the enterprise digital budget pool – is committed without regard for efficiency. People pretend as if the importance of the branding discipline precludes looking at the meter as your dollars flow. It’s a bug, not a feature of a blithely negligent culture that agencies should have rebooted at the first sign of client distrust.
When they bother to look, advertisers see that the assumptions behind their digital branding tactics don’t hold. The typical practice of spending tens of millions on custom ad units on relevant, premium sites, with demo targeting, is not optimal. The effectiveness of “high-impact” ads is rarely worth the extra cost over standard, contextual relevance is poorly verified and also overpriced and reaching high concentrations of a target audience (e.g. men watching sports) isn’t more effective than reaching them where they are less concentrated (e.g. men watching cooking shows).
Our failure to apply consistent measurement to digital branding destroys the opportunity for digital tools to earn back their cost via efficient reach.
Targeting is where many digital branding planners verge on malpractice. Accuracy against a Nielsen audience has never and will never be proven to be worth more than accuracy against a behavioral segment. No one wants people older than 54 or younger than 25 to never see their ads – in TV, this range is used because targeting is imprecise. User-level precision changes everything! The proposition that, say, a chocolate company cares more about targets’ sex or age than whether or not they ever purchase chocolate is ludicrous. Chronic passivity about the opportunity for a chocolate company to focus its fame-making investment on likely chocolate buyers is a great example of why digital buyers are not trusted.
Though there’s no turnkey solution to the problem of investing without a model, there is an easy first step for any service provider: Stop hiding how ad tech works. It’s extremely relevant that ad tech enables user-level targeting across audience lists that you, enterprise advertisers, can house or even own. Ad tech makes all audiences addressable, which makes the job of finding the most efficient way to influence any audience doable.
Once you have lists of who you’re addressing (identified anonymously or as permissioned, it doesn’t matter), you can control for, isolate and measure response. That’s true with regard to conversions, and no less true for brand lift. Striving for predictability across a set of addressable audiences is work budget owners will get behind.
This approach is better than phony DR and branding because it takes audience and reach as givens and makes influence the variable to chase. You can go as big as the biggest branding campaigns and have exactly as much control of exposure as a direct-mail advertiser working off a name and address list. I’ve seen this knowledge lead marketers toward new capabilities: managing one predictive audience strategy, consistently, across planning, creative, CRM and media.
Not that it’s easy to innovate within a marketplace fouled by distrust. It beats the alternative, though.