Stuart Colman is the MD Europe AudienceScience. Here he discusses the DMP opportunity and what it means for both publishers and advertisers.

For those not in know of the industry's latest TLA, how would define a DMP?

A Data Management Platform (DMP) is a technology solution that enables you to bring together your data into a single place (be this first party, third party or, increasingly, offline data), control it, make sense of it and then use it to build very defined audiences to target online.

Today there’s such a wealth of data available to marketers, be it online behaviours (sites and sections visited, content being read), in-site searches and key words used, registration data, third party partner data, transactional data, information from data exchanges, offline data etc. Much of this is being constantly generated and added to, so to really take control of it and use it to develop meaningful, actionable audience groups you need a DMP. Bring into this mix the key parameters of recency and frequency and you are then able to define an audience that is showing the traits of being in-market for a specific offering – the very time you want to reach them.

While the concept is not a new one, the execution varies drastically. If we look back to the development of offline direct marketing, there were products such as Viper which were developed to provide rapid access to multiple data sources, the ability to visualise the data and then use the insight to build targeting selections and segments for direct marketing campaigns. This is a very similar concept, although managing the huge volumes of real time online data that exist makes this much more complex.

What benefits does it have for both a publisher and advertiser?

For publishers and advertisers alike, it’s all about using effectively that most of valuable of assets – their audiences.

From the publisher perspective it’s about moving away from the traditional approach of selling sites and sections to using data to bring their audience ‘to life’ and allow their advertisers to target people and not simply placements. This is much more valuable to an advertiser, so higher rates can be charged and they can also service larger campaigns.

For advertisers, it’s about being able to make much more sense of the digital touch points they have with their consumers and use this insight to drive appropriate, relevant targeting and campaign messaging, over and above the simplistic world of retargeting.

Using this insight they can really take control of their targeting opportunities, identify which publishers are really important to them and build deeper, strategic partnerships with them and ensure they deliver their marketing messages to people who have a real interest in their products.

Is AudienceScience now providing the market a DMP solution - or has its segment building and warehouse solution also been a DMP-type offering?

We have been offering the leading Data Management solution since 2003. Although it may have had different names - Data Management Platform being the in-favour one currently - we’ve been offering a solution to manage and use data for audience targeting for over eight years

While originally our focus was on publishers, where our offering is ‘new’ is that agencies and advertisers are also recognising the value in using a DMP to help them properly control their data assets and action them effectively online.

How does a DMP make sense of the huge volumes of data are generated every day that can drive targeting?

Firstly, it must be able to handle these huge volumes of first party data that are constantly being generated – for example sites and sections visited, key words used for searches terms, content read, on-site tools used, how a user arrived at the site. This does require a sophisticated platform and solutions that focus on working more with modelled data often lack the ability to handle such volumes so need alternative ways to work with data.

Our business has a heritage in data analytics where it’s all about managing, interpreting and making sense of huge volumes of data and this ability needs also to be central to any DMP. Data collection is one thing but then the platform needs to be able to make sense of this information in a secure environment.

From this wealth of data a DMP should enable you to transparently build an audience. By selecting and combining the relevant data needed from the large volumes of information on the platform (based on the type of audience that needs to be targeted), a precise segment can be defined, built and then reached. This audience definition should be logical and it should be transparent how it has been built so it’s easy to understand the exact make-up of the audience.

Finally, and equally critical, the solution needs to allow recency and frequency criteria to be included in the audience building definition. This ensures that consumers are targeted when they are showing frequent behaviours and actions (but non-personally identifiable) that imply they are in-market for a specific product. In terms of using recency, the time frame specified will obviously be dependent on product type and value. For example, while an in-market car buyer audience may have a six week recency period, for travel insurance it may well be as little as 24 hours.

By offering such tools, a true DMP allows data to be transformed into actionable, relevant and valuable target audiences.

Are publishers in UK and Europe using DMPs to make sense of all this data? Will they need to partner with a third-party provider, like AudienceScience to help manage and leverage that data?

They are and it’s very much an increasing trend. Although the use of such solutions is more established here in the UK, we’re finding more businesses across Europe looking to adopt these technologies. In the last three months we’ve had new clients in Italy, Spain, Germany, and Poland, as well as here in the UK, adopt our solution.

As already said, the complexities of trying to gather, make sense and use data effectively means technology is essential to allow them to do this.

How does this type of technology help drive better targeting and user relevancy for publishers?

Ultimately, it’s individual people who buy products and if you make this central to what you do, then you shouldn’t go wrong. However, I think this simple fact is something which is often lost sight of in all the noise around technologies, fads and the latest three letter acronyms.

Although there seems a trend towards using modelled data to create ‘lookalike’ audiences – often because these businesses lack first party data at scale to build ‘true’ audiences - there is no substitute for real data as the basis to define an audience accurately. Modelled data, by its very nature, is less accurate, assumptive, (it assumes that the lookalikes will behave the same as the sample audience) and based on the best guesstimate.

A DMP allows first party data to be understood and used to define more precise audiences, delivering advertising that is more relevant and timely. In other words, not only does it help you build an audience, but based on a consumer’s activities and actions across a specified timeframe, it enables recency and frequency criteria to indicate when someone is actually in the buying window – the critical time to market to them. For the modelled data approach, however, it’s impossible to identify if the lookalikes are actually in-market for a product.

Building targeting via a DMP should allow you to create an ‘inclusive’ audience. Conversely, if an audience for a product is defined simply around a specific profile - for example male, 35-45, well educated, home owners – this then excludes all those consumers who may well be in the market for that product but don’t fit into this definition, resulting in lost opportunities. It’s much more valuable to be building audiences around any consumer who is in the market to actually to buy what you offer.

While broad user profiles can add some insight into what a target group may look like and help in developing creative concepts, its real first party data that drives effective targeting.

Can advertisers get the type of transparency in audience building by using a DMP?

Absolutely. Our belief is that the definition of an audience should be very apparent. By viewing one of our segments you can quickly see how it is constructed and what data was used to define it, which is not possible if it has been built using black box technology. For example if an advertiser wanted to reach an in-market family car buyer group, using our solution they could define this around consumers who are regularly engaging in content around family car brands (e.g. Ford, Vauxhall, Honda, Volkswagen); are reading reviews on specific models; are carrying out on-site searches on specific car brands; are seeking out content on insurance classes and have carried these activities out six times in the last two weeks. This audience could be quickly built and the publisher can share the definition with the advertiser to highlight exactly who they will be reaching

This transparency gives the advertiser confidence that the audience they want to get in front of is, indeed, the audience they will be reaching.

Can a DMP help advertisers use offline data in online targeting? Will the DMP become the hub for managing offline and online for both the publisher and advertiser?

This is very much a growth area. It’s more commonplace in the US but is becoming more important here.

Brands want to deliver multi-channel marketing campaigns. Based on their offline activity, they have already defined their key audiences, but have often had trouble in replicating these online in a reliable way. By having the ability to bring their offline definitions online, this gives them the confidence that they are reaching the exact audience they want to, rather than relying on a third party’s definition of their target audience.

Offline data can also help in adding extra granularity when building precise audiences. While offline demographic and lifestyle data is very useful in understanding the characteristics of an audience - for example defining age, income, gender, family composition, interests and product ownership - combining this with timely online behavioural and intent data adds extra depth to further hone targeting and relevancy.