How to learn to love data and not fear it

Possibly the best part of my job is that I get to interact with leaders in the online publishing and agency spaces every day. Recently, Revenue Science had a European publisher summit, where our key publishers from across Europe met in London to share best practices on behavioural targeting and to continue driving the industry forward in this space. As part of the event, Justin White, Business Director at Neo@Ogilvy, gave a presentation on behavioural targeting from the agency perspective. One point he emphasised over and over again was that data is king. He said that media owners and publishers are sitting on a goldmine of data and audience analysis, and should use that insight to sell in the smartest way. Another major point he emphasised is that publishers should use technology to defend their turf and innovate. This resounds perfectly with the recent 'Marketing and Media Ecosystem 2010' study presented by Booz Allen Hamilton. Their study looks at what media owners and agencies will need to do to be industry leaders in the future. I recommend it as required reading for any media owner and it can be downloaded from IAB.net. The study shows very clearly that consumer insights and behavioural targeting top the list of 'must have' capabilities. As we all know now, the industry is at an inflection point, and has moved (and will move even further) from experimental into mainstream. And while most major media spenders have not shifted 'significant' budgets to online, it is widely recognized that this is in the process of happening as we speak. This is clearly demonstrated by the many recent reports that online will overtake TV ad revenue in 2009. One very tangible consequence of the rapid growth of online advertising, especially as more offline branding campaigns come online, is that premium inventory is more constrained than ever. Many of the publishers we work with were totally sold out of the premium inventory months in advance in late 2007. And this trend will become even more pronounced as more revenue pours into online -- especially in the U.K. Effective audience analysis and targeting will become more important than ever. So as the online advertising industry becomes even more mainstream, rigorous data analysis and strategy will take a front seat with both media owners and agencies. This of course can be a frightening proposition, especially to many ad sales teams who have not traditionally spent time pouring over data and audience analysis. But instead of being an area of concern, media owners should embrace this move. Here is a very simple example of how analytics can be helpful in targeted advertising. Imagine you want to target a message to travellers going to Asia. If your analytics show you that the average user has 6.5 page views in the Asia section within Travel over a month, then your 'Asia traveller' segment should be defined as all users with at least seven page views in the Asia travel section in the last month. Likewise, if users in your car insurance section only view an average of 2.5 page views per month, then that segment should be defined as all users with at least three page views in car insurance in the last month. This very simple exercise shows that taking a closer look at your analytics can produce the most targeted audiences, which will be of the greatest benefit to your advertisers. Another way to use data to your benefit is to do a deep analysis of the search terms being used on a publisher's site and the words that are often read in the articles. Selecting one word as a starting point, 'hybrid car' for example, can lead to other relevant words which you never considered like 'carbon offset', 'carbon footprint', 'bio fuels', etc. And then go even deeper and look at interaction with each phrase. If the visitors are reading a lot of articles with one of the phrases, and very few articles with another phrase, then that should impact your audience definition. It is especially true in behavioural targeting -- when you take a closer look at the specific actions your audiences perform, it will help you build the most relevant audience segments. Time and again, our publisher partners have found innovative ways to use the data and analytics like this to build effective target audiences, and agencies and advertisers always appreciate the deeper look. One very encouraging sign is that many publishers are taking a close look at doing deep analysis on audience composition, time spent on page, branding metrics, A/B testing and other non-click related metrics that will further validate the online advertising model. FT.com plans to be very aggressive in this space. As Stuart Colman, their Head of Digital Sales Europe, recently told me, it will help bridge the gap between CTR metrics on one hand (a metric which is becoming increasingly outdated and less used) and full branding studies on the other extreme, which is a pretty big gap. The publishers that take these steps and integrate those findings into their full set of online capabilities, including behavioural targeting, will find the most success today and will move forward. Publishers: start using your data and analytics to provide the best possible advertising opportunities to your clients -- whether for behavioural targeting or for non-targeted campaigns. Ask your data and targeting partners to work closely with you to produce cutting edge insights. This is one of they key steps that will take your sales effort to the next level and set you apart as a world class media owner. Agencies: demand that your publisher partners and media owners provide you with cutting edge data insights, and data to support what they are selling you. But at the same time, share all data points with the publishers that you can. They are just as interested in improving back-end performance numbers, and data is the key! Data analysis and advanced behavioural targeting is the way forward -- and it is not scary. To the contrary, it is exciting! Jeremy Mason is managing director, European operations, Revenue Science.