Advanced media attribution; A should-be-standard practice

For the last few year Ad Server providers have been working to improve their ability to attribute the contribution media placements make to an acquisition in a more detailed and accurate way.

For those not in the media world, at current the standard way to measure online media is through a ‘last cookie wins’ model which means that the last ad unit to be seen or clicked get the acquisition. The problem with this method is that a consumer will generally neither interact nor respond to an advertisement on the first interaction.

In traditional media we deal with this by booking media that follows our target demographic as they watch TV in order to reach the maximum average exposure volume. The problem we face with online media is a self imposed one, we have made ourselves too accountable. With TV (for example) it is perfectly acceptable for us to say that we reached the audience an average of (x) times stretched across 2 stations because we know that they have very similar audiences. However (in the majority of cases) when it comes to digital we are treated in an overly DR way, which in term impedes our ability to truly achieve maximum ROI. For example:

Lets say a large portion of users read the review section of then research on IMDB then either click on an ad or proceed directly to QuickFlix. Now in a standard (and most commonly set up) ad server the acquisition would belong entirely to IMDB and if we were assessing the value of the media we would see only this final figure. The joy that advanced attribution models bring is that we are able (within the tracked cookie window) to trace the path the user took, in terms of their ad exposure. We can then dissect this data by applying different values to different media placements (e.g. a large OTP is worth 5 while a 468×60 is only worth 1 due to clutter) and we can then break out the value of the acquisition accordingly. In this way we may find that (in the above example) actually more value than IMDB because people also go to and then convert, but in this instance the acquisition goes to SMH because we are not running media there.

The primary challenge faced in making advanced attribution models the new standard is a lack of simplification, that is, that all ad servers deliver different data when reporting in this manner and the data is often extremely complicated to understand when in its raw form (indeed most ad servers charge you for the service then deliver an edited report). Once we have a more precise consensus as an industry around what exactly we expect from this tool we can expect to see significant progress.

If I were to make one honest prediction for a trend to watch out for in 2010 it would be for Australian agencies to start making a concious effort to increase the usage of attribution modelling across clients.

As technology currently sits advanced attribution reporting is the best way we have to understand the immediate impact and uplift that brand media has on performance/acquisition media placements.

I am interested to know, if you work in digital media (of the strategy/planning/buying/implementation variety) are you using advanced attribution reports at all? and if so who do you use and how do you find their offering?


4 thoughts on “Advanced media attribution; A should-be-standard practice

  1. interesting debate – in my minds attribution is a nightmare

    the digital solutions around this (engagement mapping etc) look at digital in almost complete isolation – so try and work out all the different exposure points and add a ‘value’ to each of them to work out the role each component played.

    this is as flawed as last cookie attribution as it treats digital as a stand alone medium when we know that other mediums contribute to an action.

    So let’s say in a day I drive to work – listen to the radio and hear an ad for ANZ – then I logon to x news site and see a banner. Later that day I am watching the news at home and an ANZ ad comes on … and then I search for ANZ and click on a paid link.

    This scenario cannot be attributed correctly through technology … but we know TV and radio have played a role. I could have even had a chat with someone around finance products. Even with multi attribution techniques it’s misleading as it’ll only look at internet based contact points.

    Now – add another curve ball … different computers used. hence cookies go out the window.

    Digital’s obsession with attribution is not a positive thing … it often shows a lack of understanding across the industry around what makes a consumer ‘do something’. It’s not as simple a journey as we wish it was.

  2. Absolutely, I completely agree that viewing any medium as a silo is not helpful and that we must look holistically at the entire media plan (ATL/Content/Etc) in order to assess real performance.

    However what I feel engagement mapping allows us to do is gain more understanding about what IS recordable. Unfortunately, whether we like it or not digital has hamstrung itself by pushing its ability to track metrics as its number one selling point.

    The primary improvement I see from engagement mapping over last cookie wins is simply that (to a degree) we are able to see these touch points and that in certain cases it can help us as agencies explain to marketers why an investment in something they are not seeing DR from is still worth investing in. After all, as you stated, they already do this with TV, Print and Radio. Unfortunately because of our own over-reliance on data the burden is on us to provide data that supports this.

    Once again, this is not relevant to everyone. A car company, or a FMCG will likely be happy to go off branding metrics and may not be evaluating every placement down to a conversion level. But for those clients that are conversion focussed it provides us with evidence of the incremental gain possible through online “branding” investment.

    Just my thoughts on it, obviously. 🙂

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s