The Drum Awards for Marketing - Extended Deadline

-d -h -min -sec

Ad Retargeting Prospecting

The prospecting model that offers a fresh take on the ‘Open Vs. Closed’ debate

Author

By Ronan Shields, Digital Editor

September 10, 2015 | 8 min read

Over 300 UK advertisers are participating in a scheme where they will pool their first-party data in order to collectively improve their online ad targeting, a model that poses as a new perspective to the ongoing ‘open vs. closed’ debate in ad tech.

Programmatic is now the mainstream with over half of all display ads bought using such technologies in the UK (see chart), which means that marketers have to get more imaginative in how they use it. But this poses marketers with many dilemmas in how they use data (and indeed use it in conjunction with third parties), spurring one of the key debates in the ad tech sector today.

How best to use programmatic?

The use of programmatic media buying technologies – commonly referred to as online behavioural targeting - has been synonymous with retargeting web users that have visited a website, but failed to convert. In the early days of programmatic this tactic was widely deployed often using crude targeting methodologies, and sometimes with negative effects.

A 2014 study from InSkin Media and Rapp Media showed that 55 per cent of consumers are put off buying an item they have previously expressed an interest in online, if they are retargeted with ads multiple times after initially researching it.

Furthermore, only 10 per cent of 1,600 20-60 year-olds surveyed in the study said they were more likely to buy an item after being served multiple times with ads for products based on their prior surfing behaviour.

Compare this with a more recent Forrester study asserting that 67 percent of marketers believe that attracting new customers, or “prospects”is one of their biggest challenges, with an additional 64 per cent claiming they need better data for the purposes attracting such prospects.

Prospecting for new customers

Step forward AdRoll, a company historically known as a retargeting outfit, which is now proposing to help advertisers target new audiences online with a ‘first party data pool’ that poses as an alternative to the ongoing ‘open vs. closed’ debate in ad tech.

The company has today (9 September) announced that it has convinced over 1,800 advertisers, 300 of which are UK-based, to participate in its method of prospecting which uses the above methodology to build an “IntentMap” using its AdRoll Prospecting product.

Under this prospecting model, AdRoll asks advertisers to opt-in to the data pool, which then predicts the audiences’ intentions using algorithms, and the advertiser is then matched with the most suitable audience(s) to serve with their ad.

Michael Bertaut, AdRoll's managing director for EMEA (pictured above), explains further: “Many of the advertisers we’ve worked with in the retargeting space have said they want to be able to bring new advertisers to their website.

“The way the prospecting works is that each advertiser will opt-in their anonymised cookie set, and then we’ll be able to build a profile based on the highest performing [audience] segments of that advertiser.”

“We’ll then identify someone [from a relevant audience segment] from another opted-in advertiser with a very high probability of being interest in that product [or ad],” he adds.

A billion-strong first-party data pool

Forrester’s study, mentioned earlier, also found that although marketers are aware of the importance of intent data for their campaigns, 54 per cent surveyed were unable to integrate this intent data into their current targeting technology. This figure comes despite the fact that 67 per cent of participants believed that integrating intent data for prospecting and retention would give them an advantage over competitors.

Bertaut further explains that with over 20,000 advertisers participating in the scheme, AdRoll’s “cookie pool” was over 1 billion-strong, and that the combination of high-intent, first-party data and predictive algorithms has shown strong performance results.

Those participating in the initiative’s beta-phase, with advertisers reporting an average 30 per cent increase in engagement, equating to 11 per cent lift in conversions, compared to basic retargeting, the company claims.

But can a ‘cookie pool’ target mobile users effectively?

However, with the advertising industry constantly seeking ways to engage with audiences on mobile devices (where cookies are notoriously limited), how will such a targeting method work?

Bertaut says that although the audience profiles contained within the data pool are essentially based on cookie data, AdRoll’s IntentMap is valid across devices as it targets audiences’ behavioural patterns, not just static demographic data.

The prospecting function still works on mobile devices as it uses signed-in profile data, for instances on mobile apps or on a mobile browser such as Google’s Chrome, says Bertaut.

“Essentially what we’re doing is capitalising on this first-party data set that we’ve been able to put together, which is effectively across all the Googles, the Facebooks, and all the other inventory sources that we work with in order to be able to bring them this product,” he adds.

Open Vs. Closed

This method of targeting offers a variation on the existing ad targeting models posed by companies such as Facebook and Google whose respective Social Graph and Device Graph use logged-in user data to target users across screens, a methodology commonly referred to as deterministic modelling. This method of targeting users across devices is widely reported to have accuracy rates north of 90 per cent.

Speaking earlier with The Drum, Stephen Webb, head of sales for Facebook’s Atlas in EMEA, said: “We are able to leverage information insights from the 1.49 billion people that use Facebook to provide the most accurate conversion measurement across devices and browsers – it’s about reaching real people, not cookies.”

However, detractors of deterministic modelling are often keen to point out that Facebook and Google’s models, plus terms and conditions, mean that advertisers will have to use different targeting methodologies within different environments.

For instance, an advertiser will have to use one model within Facebook, another within Google, plus a completely separate on elsewhere. This fragmentation can make the execution of campaigns more complicated, not to mention lead to an increase in the required manpower.

Such an approach is often called a ‘closed’ or ‘walled garden’ approach, with Facebook and Google rivals keen to advocate a method of ad targeting that uses probablistic modelling. This often involves using third-party data sets to match audiences to ads.

Advocates of this model, who favour a more ‘open approach’ maintain it involves less ‘first party data leakage’ for advertisers, i.e. giving away their audience insight to third parties, plus it also means they won’t have to employ different methodologies, as required with walled gardens.

Paul Gubbins, head of programmatic at Millennial Media, previously told The Drum: “For example, platforms that operate within a walled garden are happy to ingest first-party data but rarely enable the extraction of post-impression insights that allow CMO’s to leverage these learnings, and inform other parts of their marketing strategies.

"Consequently, brands that want to access that data should consider open platforms that don't lock them into running campaigns in one place, but support them to use their data in the ways that suits them best.”

However, Facebook and Google maintain their methodologies have a higher accuracy rate, and Facebook’s Webb is keen to downplay the ‘closed’ accusations of its detractors. He points out that Facebook’s Atlas measurement tag can also be run in conjunction with most existing ad-serving tags, giving critical reach, frequency and conversion insights across devices and publishers.

Meanwhile, AdRoll’s Bertaut maintains his company’s methodology, which fuses deterministic data with probablisitic modelling, offers advertisers the benefits of both approaches. “Essentially what we’re doing is capitalising on this first-party data set that we’ve been able to put together [through working with advertisers using its retargeting offering], which is effective across all the Googles, the Facebooks, and all the other inventory sources that we work with,” he adds.

“We’re a strong, strong proponent of being non-siloed, and to not having to work exclusively with a Facebook, or a Google for an advertiser. It’s far more beneficiary to the industry as a whole if there is data-sharing across the different inventory sources."

Find more digital advertising content at The Drum's dedicate news hub for the sector.

Ad Retargeting Prospecting

More from Ad Retargeting

View all

Trending

Industry insights

View all
Add your own content +