Understanding the limitations of ad targeting

Understanding the limitations of ad targeting

Ad exchanges and programmatic buying can improve ad targeting, but there's danger, too, writes Jane Bainbridge.

It’s the type of crisis that wakes marketers in the middle of the night in a cold sweat… your online brand advertising pops up next to a hate speech or a piece extolling violence toward women, and the issue goes viral.

However, this was no mere nightmare, but the reality facing several brands when the #FBrape campaign started in late May, orchestrated by a coalition of groups including Women, Action and The Media and The Everyday Sexism Project. They expertly exploited the speed of propagation possible in the social sphere and spurred Twitter users to directly challenge the affected advertisers, urging them to pull their Facebook ads and thus pressuring the social network to take action against gender-based hate speech posted on it.

More than 60,000 tweets later, advertisers from Nationwide to Nissan pulled their ads, although others, including Dove – particularly pilloried because it was so at odds with its "Real beauty" positioning – toughed it out. Facebook eventually said it would review its policies. At the end of June it announced it was changing its systems to restrict where ads appear on the site. No ads will run on pages containing violent or explicit sexual content.

Nonetheless, what this campaign brought to light was a much broader issue about advertiser targeting. In the midst of this crisis it became clear that most consumers did not care why an ad appeared where it did or about the mechanisms behind ad serving, and so held the advertising brands directly culpable. It also raised questions about how expert many marketers are on such issues, and the quality of the advice their agencies provide (see Phil Georgiadis' advice, below).

Is targeted online advertising a myth?

Digital marketing promises advertisers so much in terms of targeting, but the shift from a one-to-many to a one-to-one model can be problematic. So, given current automation, is targeted online advertising really a myth?

In the main, no – as Richard Sharp, man­aging director of audience-targeting platform ValueClick Media says: "Anything is possible, but at a price." Even so, he believes targeting has been oversold and overpromised. "The technologists have got ahead of the marketers; it is all possible, but in the real world it’s not practical."

He argues that advances such as programmatic buying and exchanges have to go hand-in-hand with experienced and knowledgeable teams on both the client and agency side, and calls on advertisers to take responsibility. "Some understand exactly what’s going on while others go to the occasional conference, read the odd article, pick up a few buzzwords and think they know what they’re talking about."

It's really hard for a marketer to understand what is affecting their placements. You can't rely on a black box to make it work.

Magnus Fitchett, marketing strategy director at interactive agency SapientNitro, similarly urges the human touch to add a layer of vetting and control. "There is so much data you can overlay across your impressions to be more personalised, it’s really hard for a marketer to understand what is affecting their placements. You can’t rely on a black box to make it work."

The industry is moving away from blind networks to more transparency, especially transparent DSPs (demand-side platforms). Ben Cockrell, data strategist at media agency MEC Global Solutions, says targeting techniques have improved and, because of the huge numbers of media-owners, there is no alternative but to be reliant on automated networks – although there are inevitable consequences.

"The non-hierarchical and dynamic nature of the internet makes it impossible to control with certainty where your advertising will appear," he adds. "It’s important that the industry continues to take steps to mitigate against the risks – ad-verification technologies are a step in the right direction."

Preparing for the worst

Balancing risk is an important aspect of online advertising, especially on social networks. Marketers want the reach and scale they proffer but must be prepared in case it goes wrong.

"Facebook is a great solution; it’s very cheap and easy. But the content is not policed – it’s not written by journalists and doesn’t go through an editing process. You have to make sure with your intermediary you set and define your boundaries properly," says Sharpe.

Brands need contingency plans ready for the worst and, just as the internet means mistakes are disseminated fast, it also means ads can be pulled quickly, and appropriate responses issued by agile brands. Speed of response is vital, but requires detailed contingency planning and escalation processes with clear lines of responsibility.

Brands have to work out scenarios: what if my brand is associated with porn [due to ad placement]? It's not about withdrawing advertising; no one person can change anything, but, as a profession, we have to move forward.

Anne McCreary, digital strategy director at media agency Carat, says: "Brands can’t afford to not be in those spaces so they have to work out scenarios: what if my brand is associated with porn? It’s not about withdrawing advertising; no one person can change anything, but, as a profession, we have to move forward."

In the case of the #FBrape campaign, not all brands withdrew their ads. None, however, is keen to talk about the strategy behind the decisions. So while Nissan pulled its ads promptly, it would not offer any comment for this article as it was "keen to draw a line under it". Similarly, Nationwide would not expand on its original statement welcoming Facebook’s commitment to improve its response to violations of standards.

"Brands are like cats, they have nine lives. So even for Dove, which has done so much right… it’s seen as no better than all the rest. A few moments of poor positioning really is detrimental; Dove has lost one life," says McCreary.

A representative of the skincare and beauty brand says in future it will extend its Facebook targeting. "We were shocked to see our ad on an inappropriate page and spoke to Facebook immediately, who removed the page. We will continue to carefully review and revisit our advertising and marketing activities to reduce the chance of this happening again."

However, this level of refinement in targeting is not currently possible – particularly when it comes to Facebook targeting, which focuses on people, not content.

Nick Baughan, managing director of media agency Maxus UK, agrees that brands have to move fast when ads are served up against inappropriate content. "It’s relatively easy to blacklist an entire inappropriate website, but harder to proactively blacklist individual URLs that are part of a reputable site," he warns.

Online protection

If at one end of the spectrum you have networks with very little moderation, and at the other, sites with full editorial control, there is a wide choice of more moderated and protected online communities offering brands a greater degree of risk management than complete UGC sites. Frequently, the digitised platforms of traditional media offer this. Sam Finlay, head of digital advertising at IPC Media, is, not surprisingly, a proponent of the protection that traditional media online offers. He also argues for more nuance by advertisers in their targeting.

"From a direct-response point of view, data-targeting is hugely beneficial for re­targeting or upselling. But for branding, there is a debate about what’s more influential – dem­ographics and data on users or environment and mindset when seeing campaigns. Data-targeting is missing the environment to some point," he says.

The #FBrape campaign has been a successful exercise in consumer activism: it highlighted the level of (sometimes violent) anti-women content online; forced brands to publicly address the perceived endorsement resulting from ad placement and made Facebook review its evaluation of, and guidelines regarding, hate speech on its network.

Facebook will remain strong, however; with more than 6bn active monthly users, few brands would ignore the platform, regardless of the risks.

Rob Jewell, chief executive of social marketing software company Spruce Media, says: "Facebook is the first channel where true identity-buying is possible, since users are inputting actual data about themselves, and the ads are served to logged-in users. It offers reach and scale plus identity-buying, which is really the perfect competition to TV advertising."

What is required is better understanding and communication between marketers and media agencies on the limitations of targeting, and robust response strategies for when another social-media ad crisis occurs – as it inevitably will, when ad sites are at the mercy of what the user posts on the page.

Inside the trading desk

Phil Georgiadis, chairman, Walker Media

Trading desks level the playing field for media agencies because agreeing a rate for each impression negates the advantage of buying in bulk. Publishers get to sell off their excess impressions for the best possible rate, and clients see reductions in their CPMs and CPAs… so what’s not to love? That’s fine, as long as you’re happy that you know the answers to these five questions:

1. How is the trading desk funded?

You are paying your media agency a clear fee or commission. You may even have a clause in your contract guaranteeing you the negotiation of best rates possible. So do you know exactly how much your agency’s trading desk is costing you? You are almost certainly being charged a mark-up on the price paid for impressions by the trading desk, usually justified by the effort of enriching the data into effective inventory for the client. Do you know how much this is? Is it broken out in invoices by the DSP?

2. Are you happy allowing your trading desk to "arbitrage"?

In other words, to pre-buy inventory in bulk and then sell it on to you? Traditionally, media agencies have acted as an agent, buying the right media for the specific objectives of their clients. In March, GroupM’s Rob Norman accepted that [GroupM’s audience-buying platform] Xaxis trades through arbitrage, meaning that media is not bought to order but according to the trading desk’s deals. Does this give you sufficient buying neutrality?

3. How is your first-party data handled?

At least one trading desk allegedly requires its clients to sign a contract that allows the learnings from client data to be used to inform strategies for other clients, competitive or not, buying through the same trading desk. Are you comfortable with your customer data being used like this?

4. How is the trading desk safeguarding your brand?

What safeguards does it have in place to ensure that your brand does not appear near inappropriate content? And how will you be compensated if these guarantees are broken? What rules have you agreed in terms of placement, and which verification tools are being used to ensure that your campaign is delivered as agreed?

5. What technology is used, how often is it reviewed and are the tech-provider contracts exclusive – and if so, why?

Trading desks use a wide range of third-party technology and these will shape and influence the buy. For example, some trading desks have had a heavy reliance on Google technology. It is important that clients understand these relationships, and how they may affect delivery.

One last point: in a world of reducing CPMs and better CPAs, you might be tempted to say "let sleeping dogs lie". People tell me those chuggers working for charities "pay back"… just as your CPA, through a trading desk, might appear to. But what about the people you irritate along the way if your approach to selling loses you some potential high-value prospects?

Start Your Free 30-Day Free Trial

Get the very latest news and insight from Campaign with unrestricted access to campaignlive.com, plus get exclusive discounts to Campaign events.

Become a subscriber


Don’t miss your daily fix of breaking news, latest work, advice and commentary.

register free