The Exchange Lab - Powered By Proteus

News / Events

Attribution Modelling – Cause for Concern or For the Cause?


James Aitken is CEO and co-founder of The Exchange Lab. Prior to that James founded ad network MediaBrokers in 2001 and ad serving technology company Atlas Europe, both of which were acquired by aQuantive in 2006 and Microsoft in 2007.

Attribution is seen by many as the holy grail of marketing, but getting it right in a multi-device world is not easy. The goal of attribution is to understand and apply value to different marketing touch points and assess the value in each of these events. It’s what all marketers should be striving for and it holds the key to learning and optimisation in marketing campaigns. It also sounds simpler than it is as many measurement methods don’t knit together easily. Differing technologies, measurement methods and connections online and offline can often lead to a confusing and blurred picture of a brand’s return on investment.

I founded Atlas in Europe in 2001 – it was a third party ad server and the first to drop 1×1 pixels to help marketers determine where digital media conversions/actions were coming from. The Atlas “universal action tag” was used by many marketers to help them understand attribution. Since that time there have been huge shifts in the attribution industry, with many companies being bought by the big media owners. Atlas was bought by Microsoft in 2007, and is now owned by Facebook; DoubleClick and Adometry are both now owned by Google; Converto was bought by AOL. Attribution can offer transparency and proof in technology, something that all marketing companies strive for. The problem is that in order for an attribution company to truly offer that transparency to marketers, I believe they must be independent and not owned by big media companies.

So, what does the future hold for digital attribution? What do brands have riding on getting it right? Can modern marketing methods such as programmatic advertising help with the process?

Many measurement methods

For a long time, the digital advertising industry has worked on last click/last touch methods of attribution, although we now realise that this isn’t efficient at telling the whole story. The consumer market is shifting. Consumers are not married to one device. 90% of consumers use multiple screens sequentially. Cross-device analysis paints the story of a consumer moving from one device to another throughout their day. They may research products on their mobile phone and purchase on their desktop later in the day. It doesn’t mean that the ad they saw on their mobile device was not important to the final transaction or in the path to purchase. There are still no perfect attribution models, but the industry is making steps in the right direction.

Multiple source attribution methods analyse data from across all digital marketing methods, be that search, email, video or rich media takeovers, and assess the sequence of a customer journey. Fractional attribution methods can assign and weight the importance of each of those methods, according to influence on the final outcome. Marketers aim to understand the affect each channel has on their consumers, rewarding the channel supplier accordingly and giving credit where credit is due.

Ultimately, what marketers are looking for is a complete picture of their consumer to make more informed decisions. This will only be fully realised by connecting the online and offline worlds. Still in its infancy, offline/mobile data is collected from mobile phones and in-store beacons meaning that retailers with physical stores can now start connecting activity online to in-store buying. Innovation in the area is driven by the desire for a full view of the consumer journey and will continue to accelerate this technology.

Who understands attribution?

One of the biggest challenges marketers face with attribution is truly understanding how it works, followed by achieving buy-in from across their business. Making key stakeholders aware that traditional digital measurement methods alone are not capable of explaining return on investment can be difficult. Translating results to different business leaders requires patience, understanding and ultimately, a method to bring all touch points together. Programmatic technology can help with this process by bringing all channels together, a process which will continue to develop as traditional media such as television and radio begin to be traded programmatically.

Data-driven marketing

Similar to algorithms in programmatic technologies, predictive algorithms in attribution models allow for easier testing and learning and can be used alongside human insight to assess whether the attribution solution in use is the right one for your organisation.

Programmatic technology itself allows brands to activate data on a scale never previously achieved. Data holds the key to unlock customer insights and enables brand communications to become more targeted. Programmatic has the ability to break down silos, operating mobile, video, social and display campaigns as one. This enables clear insights to be drawn from one source, making the path to a full view of the consumer easier. With attribution added into the mix, marketers gain a clear view of which interactions are connecting with consumers.

As brands move more of their budgets toward programmatic marketing accurate, attribution models will become important to ensure they get the full picture of marketing success. Independent attribution companies with no allegiance to stakeholders offer marketers greatest value. Today, the best way brands can begin to achieve one view of the consumer across all channels is by removing campaign silos and working with partners who take an agnostic approach marketing.