man holding phone with illustrated application symbols bursting from phone

Marketers want a clear way to decide how much value each channel or campaign adds to their companies’ success. But it can be hard to tell which of the many attribution methodologies in use will provide the most accurate answers. Meanwhile, deciding on a specific attribution approach is hardly a zero-risk proposition. After all, the wrong attribution method can give too much credit to the incorrect channels—guiding advertisers to drive dollars away from successful channels in favor of less effective ones.

To address the attribution question, MarketShare turned to big data for help. We tackled the issue with a computer simulation of a year’s worth of advertising on a group the size of the United States (316 million people). The aim of the project was to create a complete "population" in which it was clear how advertising influenced purchase decisions—and then to run various attribution approaches to see how close these approaches came to describing the simulated "reality."

The simulated individuals were exposed to online and offline advertising; and were programmed to respond to marketing as a real consumer would. To ensure that the simulation mimicked the real world, we imbued each software "consumer" with an innate degree of interest in the product, as well as saturation points beyond which an ad or a channel would decline in effectiveness. The resulting analysis was huge—generating 2 terabytes of simulated tracking data over two weeks’ time.

Once we had created the simulation, we ran attribution analyses on the ways advertising influenced the simulated population’s purchase decisions. We analyzed four of the most common digital attribution models: first click, last click, matched pairs, and discrete choice models. Since the simulation already “knew” exactly how advertising had influenced purchase decisions, we were able to match the various attribution analyses with the actual influence that ads had had on purchases.

Next, we scored the attribution methodologies to see which matched most closely with the ads’ actual influence. The result showed us which attribution models were the most—and least—effective at understanding the customer’s path to purchase.

Conclusion: The Most Effective Attribution Methodology Is...

As shown in the ranking below, there was one clear winner amongst the attribution approaches—discrete choice models (nearly 100% accurate); with first-click coming in last among the approaches (45% accurate).

The high degree of accuracy of discrete choice model can be attributed to a number of factors:

  • The model takes offline factors into account, not just online (other models can only look at online data).
  • Discrete choice models operate at the individual consumer level, providing a high level of granularity.
  • While other models take a binary view of ad effectiveness—Was the consumer exposed to an ad, or not?—discrete choice looks at recency and frequency of ad exposures (How many ads was the consumer exposed to, and how recently?). This creates a far more nuanced picture of the impact of an ad, and of a channel.

In short, discrete choice modelling provides a more complete and exact understanding of what drives individuals toward conversion. This, in turn, produces a more accurate understanding of which marketing investments work, and how.

For a deep dive into the study methodologies and findings, download the complete MarketShare brief.