Content marketing

Brand safety or ad performance? Here’s why it’s a false choice

The latest research into brand safety highlights an apparent dilemma for marketers. It’s clear that they are hugely concerned about brand safety for advertising. It’s clear that their concerns are justified. However, it’s equally clear that a lot of marketers believe their hands are tied: they can only go so far to ensure that their brands only appear in an appropriate context.

We’ve been sold the assumption that brand safety inevitably involves a trade-off; that you can’t improve it without undermining the performance of your campaigns. When the digital advertising platform Sizmek surveyed 522 advertising decision-makers across EMEA and the US in August, 64% said that achieving brand safety for their campaigns negatively impacts on performance and 71% said that it was difficult to achieve reach while delivering ads to the right audience in the right context. This suggests that at least two-thirds of marketers have been convinced that putting their brand in risky environments is the only way to drive the results that they need to report to the business.

If you buy into this assumption as a marketer, then you are left wrestling with all kinds of unattractive questions: how safe do you need your brands to be, and how much performance are you willing to sacrifice in order to secure that safety? How much can you risk reporting higher costs for clicks, leads and impressions back to your business, in order to prevent the less visible damage that your brand might suffer from appearing alongside inappropriate content? How much of a drag on ROI can you risk in order to insure against the PR firestorm that might occur when a brand safety issue erupts?

These feel like impossible choices – but I would strongly argue that they are false choices. They’re based on a misguided perception of the value of the performance that brand safety supposedly undermines.

We’ve been too quick to accept that brand safety comes at the expense of meaningful results. Based on my experience as both a Campaign Manager and a Marketer, I’d argue that looking past initial vanity metrics quickly changes this view.

Before I explain how, let’s take a quick moment to clarify exactly what we mean by brand safety in marketing – and what the latest data shows about the scale of the brand safety issue.

Most marketers would define brand safety as avoiding their brand appearing in inappropriate environments or next to inappropriate content. The question of course is, what constitutes ‘inappropriate’? In January this year, the AI-powered image-analysis platform GumGum and Digiday Media asked marketers to rank the brand safety issues that they were most concerned about. In descending order of concern, these were: hate speech, pornography, violence, fake news, divisive politics, disasters and tragedies, and competing brands. Vulgar language, drugs and other issues scored much lower in the survey – but will be considered more important by some brands and marketers. Other concerns are more specific to particular brands. A sports nutrition brand might consider a news story about performance enhancing drugs in sport to be an inappropriate environment – but plenty of other businesses would not. News coverage of a celebrity scandal might be inappropriate for a brand with that celebrity as a spokesperson – but less so for others. Context matters in defining inappropriate.

Another question for marketers to consider is how close the association to these different forms of content needs to be before activity becomes ‘unsafe’. Does a problem only occur when a brand appears next to hate speech or fake news – or is it unsafe simply to appear in an environment which is associated with these things? After all, doing so involves supporting a publisher that is providing a platform for those views.

Because brand damage can take so many different forms, brand safety is an issue that concerns almost every marketer. In an eye-catching topline statistic from an Oath survey of advertising decision-makers this year, 99% said they were concerned about their brands appearing in unsafe environments – and 58% said their concern had increased compared to a year ago. Over a third (38%) of the advertising decision-makers that Sizmek spoke to said their brand had definitely run ads on harmful web pages, and a further 5% weren’t sure whether it had or not.

Reducing brand safety risks involves a range of different actions to address the different ways that a business might be buying digital ads. When Sizmek asked advertisers to define what a brand safety strategy involved, 60% identified blocking the most harmful pages, 50% named managing a blacklist of sites, and 48% described delivering only on pre-defined site categories. However, the broad nature of brand safety risks, and the unpredictability of some social and user-generated content platforms mean that these are partial solutions at best. They’re not going to catch every issue. Things aren’t helped by the complexities of the digital supply chain for marketers buying ads programmatically, on wide-ranging open auctions. That’s why 76% of advertising decision-makers say they’re prioritising getting more transparency from their digital inventory, and 64% are prioritising reducing the number of vendors that they work with.

Of course, the most effective way to get more transparency, clarity and simplicity into the digital advertising supply chain is to be specific about which publishers and platforms you partner with. When you’re making active decisions about which environments to include rather than which to exclude, your default setting is far safer from a brand point of view. And the value of choosing specific partners increases hugely when those partners deliver a consistent context for your ads. In GumGum’s ranking of platforms considered brand safe by advertisers, LinkedIn scored +45 while all other major social platforms received double-digit negative scores. This isn’t a surprise when you consider the consistent professional and appropriate context that the LinkedIn feed provides.

The Sizmek study identified a big swing towards marketers adopting a more prescriptive approach to targeting that involves identifying the right contexts first, and then exploring how to reach relevant audiences at scale within those contexts. In fact, 45% define brand safety as delivering ads only on premium publisher sites, and 85% say they are planning to scale this form of contextual targeting.

This though, is where that assumption about safety compromising effectiveness kicks in – and where marketers are told they have to worry about costs going up and performance going down as a result of their more selective approach.

Here’s why that concern is overblown, and why compromising brand safety for the sake of it is almost always a misguided strategy. Before you do so, ask yourself these three key questions about what your campaign is trying to achieve – and how it’s trying to achieve it:

Is the performance that you are taking brand safety risks for really worth it? As B2B marketers and their businesses place greater emphasis on revenue-based metrics, it’s becoming clear that optimising around click-through rate (CTR) and cost per lead (CPL) can become dangerously misleading if you’re not also measuring the value of those leads. Most B2B businesses are better off generating fewer leads at a higher CPL but converting those leads to customers at a higher rate. This delivers a lower cost per sale, a higher real Return on Investment (ROI), and keeps sales teams a lot happier as well.

If you’re generating more clicks at a lower cost but with little transparency on the context the clicks come from, how relevant are those clicks likely to be for your business? Are they delivering the lead quality that makes the sales pipeline add up? And are they therefore worth the brand risk you’re taking?

On the face of it, targeting the exact individuals you want, wherever you can find them on the web, seems a lot more efficient and scalable than trying to find enough relevant people across a few, more specific sites. However, the theory of how you find and buy an audience is one thing. As a marketer, it’s important to dig a little deeper into how much you really know about who you want to target.

Any targeting approach involves assumptions about who is really relevant to your campaign, and who isn’t. Behavioural targeting picks up on potentially valuable signals of intent but excludes audiences who may go about their buying research differently – and will inevitably include plenty of people who sent behavioural signals but aren’t actually in the market for your products. Demographic targeting makes assumptions about what your typical buyer looks like. However, any business has plenty of buyers who don’t fit neatly into this model. Effective targeting recognises these different assumptions for what they are.

When you target individuals with no regard to context, you’re betting the campaign on the completeness and accuracy of your targeting assumptions. When you target people within specific environments that you know are relevant, you’re increasing your chances of finding the right prospects. And with this increased confidence, you can loosen your definition of your target audience slightly, buy at a lower CPM or CPC as a result, and find more reach and scale cost-efficiently. Experimenting with different ways of targeting your audience, and then tracking the effectiveness of these different approaches, can help you expand your understanding of your audience and deliver high-quality leads you’d have ignored otherwise. Before you decide whether it’s really more efficient to target an audience at scale regardless of context, ask how confident you are in your starting definition of your audience. Exploring different targeting approaches within a brand safe environment can enable you to find reach and scale while discovering new, valuable segments. The Sizmek study found that a majority of marketers are now favouring this approach. When asked to choose between “right person” and “right place” for their campaigns, context takes priority.

Taking a riskier approach to brand safety doesn’t just cause problems when your ads appear next to something hateful, disturbing or tragic. Playing in the environments where this is more likely to occur is doing damage to your brand even before something goes badly wrong.

A growing body of research shows that the context in which an ad appears has as much, if not more, influence on perceptions of the brand than the content of the ad itself. As marketing blogger The Ad Contrarian puts it, the message of what your ad says is often seen as a less reliable indicator of your credibility, trustworthiness and value than the signal sent by where your ad appears. When an ad appears in unsafe or potentially inappropriate environments, any clicks that it generates come at a cost that’s not included in the CPC. What perceptions of your brand have been formed by the far greater number of potentially relevant people who saw your ad but chose not to click on it?

Compromising on brand safety in order to safeguard campaign performance is a trade-off that makes less and less sense the closer you look at what you’re getting in return. If you’re a B2B marketer aiming to deliver on metrics that mean more to your business’s bottom line, then there’s really very little to lose by advertising only in relevant environments where you can be genuinely confident of safety. To do otherwise involves taking more risks than is often realised for less reward than is typically claimed.

I predict that we’ll see brand safety becoming less of a dilemma for B2B marketers the more we succeed in aligning our businesses around the need for revenue-based metrics. If sales and the C-suite are on-board about the value of prioritising quality leads over quantity (and in many cases, this involves pushing at an open door), then the maths only points in one direction from there.

Before you feel pressured into a false choice on brand safety, try running the numbers on the real value of clicks from proven safe environments compared to unsafe ones. The dilemmas over performance will be a lot easier to resolve once you do.