Programmatic is lauded as the future of digital media, enabling marketers to deliver personalised messages at scale to their prospective customers.

Global programmatic spend was forecasted to hit $46 billion last year according to e-marketer; and it will continue to rise as programmatic TV, OOH and audio gather momentum. The technology continues to evolve, with platforms now offering omnichannel targeting solutions to help advertisers reach their customers across multiple touchpoints in real time.

As ever, the success of new technology relies on our ability to prove its return on investment. And right now, the way most performance is assessed, in-particular direct response, is holding back programmatic advertising from its true potential.

The culprit is last touch attribution.  Campaign effectiveness for direct response is still predominantly assessed through last touch attribution, a methodology which decrees that the most recent piece of advertising served to a consumer be given 100 percent of the credit for a subsequent transaction.

Last touch attribution is a counter-productive attribution methodology, incentivising bad behaviour. In order to win the last touch attribution game, all you need to do is record having served the last ad to a cookie or device ID that can be tied to a purchase.

It’s a race to the bottom.

This leads to advertising budgets being inefficiently spent, with high-frequency volumes creating wastage, ads being shown against low-quality inventory and media budgets being inadvertently arbitraged as multiple buyers target the same users. Worse, there’s virtually no emphasis on identifying potential prospects and guiding them through to purchase.

The challenge with last touch attribution is that, in reality, no single channel is responsible for a conversion. Every channel, from TV and search to display and beyond needs to work together to guide a consumer through the purchase funnel.

Now, if you apply that principle to programmatic direct response activity, evaluating a buyer based on who saw the last impression removes the buying efficiencies in identifying new prospects and guiding them to purchase at optimal contact frequency.

Last touch is so prevalent in New Zealand because it’s easy to set up, use and interpret and is the default attribution model of most ad servers. Probably most importantly, it’s also cheap to use and this really matters in a smaller market like New Zealand. So, for advertisers and agencies that haven’t given much focus to attribution modelling, it will likely be used by default.

As programmatic continues to evolve, so too should the ways in which we assess its effectiveness. There are several attribution approaches that measure the impact of multi touch attribution.

Linear regression modelling distributes credit evenly to every single touch in the user’s journey. The W-Shaped model highlights the three key funnel transitions (first touch, lead conversion and opportunity creation) that marketing impacts in the customer journey. The Z Shaped, takes this one step further by adding in a fourth stage, the customer close.

It is worth noting that whilst these models give credit to all touch points, they don’t assess the relative influence that each channel had on performance that an algorithmic model can provide. These models assign a weighting to each channel and touch point which represents its perceived importance.

The two most commonly used approaches and indeed ones that I feel are the best methodologies are the Shapely value approach and the Markov chain concept, both of which are derived from game theory. The reason for this is that both models take into account each touchpoints influence as part of the customer’s total journey and provides insight into the perceived most influential channels. When adopted within programmatic activity, buyers can optimise spend to channels that have a greater influence on the outcome advertisers are trying to achieve.

I would challenge advertisers and agencies alike to work with programmatic partners who can demonstrate efficient spend of budgets through the use of algorithmic models that show a calculated, methodological approach to not only identifying users at the top of the funnel, but utilising a cross channel targeting approach that guides a user to purchase.

To help advertisers achieve this, the industry needs to make attribution an easier topic to interpret, understand and implement and most importantly, in a market of New Zealand’s size, make it cost effective. By helping brands move past last touch attribution, there is the potential to completely re-orientate programmatic around what the focus should be on – delivering real outcomes and value for clients. Only then will last touch be consigned to the history books.