The Power of Multichannel Attribution at TD Bank [#CZLNY]

Multichannel attribution is the key to making good marketing optimization decisions, yet too many marketers still rely on non-attribution reports, potentially up to 75 percent. But Nick Necsulescu of TD Bank thinks this is the wrong approach.

“Last-click attribution does not acknowledge the other channels or tactics that have an influence on the customer journey,” during his session at ClickZ Live New York. “It’s unbelievable that we still rely on these reports when our ecosystem has become so multichannel and integrated. Multichannel attribution gives credit where credit is due.”

There are many baseline models for multichannel attribution, which defines how credit should be directed to interactions or touch points in a conversion path. Using them as strict models can be a start, but TD Bank customizes the use of several together in a custom model.

Baseline models include:

  • Last click (100 percent credit given to the last channel)
  • First click (100 percent credit given to the first channel)
  • Linear (Equal credit given to each interaction)
  • Time delay (Assigns more credit for those channels closest to the conversion)
  • Position based (Assigns credit based on 40 percent to the channel at the beginning, 20 percent to the channel in the middle, and 40 percent to the channel at the end)

“Your mix of these models should be custom to your business and will drive how your results are determined,” Necsulescu said. TD Bank has created a technical framework to track the customer journey and link to the database, based on a custom position-based model.

But there are a few gaps: Some mobile activity is off the grid, and call center data is not fully incorporated.

“Using a centralized CRM, we can track offline purchase of people who created an account online – however, if they do not log in, we don’t have a way to track those people,” he said.

Necsulescu uses a hockey analogy, where points are given both for the goal and for the assist. However, if the goal is reached without any assists, then 100 percent of the credit is given to that channel. “It’s important to ensure that single-channel conversions are pulled out and measured as such,” he advised.

But most conversions are multichannel, Necsulescu explained. “We found that 60 percent of last-click attributions applications were assisted by another channel. Knowing this, how can you not measure everything, rather than allow last-click to drive the understanding?”

TD Bank looks for audience intent at the start of the interaction, as well as conversion to an application (e.g.: a mortgage). The bank found that when social media was part of the conversion path, completion rates soared to 33 percent.

“We also found that when someone is ready/committed, they move quickly through the application process, Necsulescu said. “We look at how decisive the consumer is. When someone is using multiple channels, they are less committed. Those in a single channel, moving forward with more purpose, convert at higher rates,” he said.

Metrics are tracked for conversions and acquisition and engagement across channels. This is not about mapping a journey, but understanding the impact of each channel on the ultimate conversion. “It was an eye-opening to see how much of an impact the digital channels have.”

How can the rest of us mimic TD Bank’s success with multichannel attribution? Required for success is a real commitment to this approach and to marketing optimization, Necsulescu said. That requires alignment of corporate culture and good team communication. In addition you need:

  • A budget
  • Big data
  • Analytics tools
  • Analytical skills and staff experience
  • Analytics innovation – the insights and willingness to move beyond reports
  • Time and a commitment to stick with it

Related reading