Cross-channel marketing optimization requires allocating marketing investment for the biggest bang for the buck. In a world of isolated channels and passive media consumption, determining the return on investment for marketing dollars is relatively easy. However, in today’s cross-channel marketing environment, response attribution and budget allocation are more complex because of online marketing, channel interaction and active media choice empowered by mobile and Internet technology.
Today’s consumers are typically exposed to a series of marketing touches from multiple interdependent online/offline channels before making their economic choices. Before purchasing a car, for example, a buyer may have seen a TV ad on a particular vehicle several times, visited a dealer Website referred in the TV ad, searched online to find more details, clicked on a banner ad for special APR offer and received a $1,000 manufacturer rebate that finally tipped scale in favor of buying the car. The purchase is clearly a culmination of the cumulative influence of exposures to all these marketing efforts and promotions assisting one another in generating the revenue. It would not be fair to attribute this revenue to one effort alone, whether it is the TV ad (the first touch), the rebate (the last touch) or any one effort in between. It is also arbitrary to attribute the revenue equally across all efforts because different efforts across multiple channels are expected to exert different impact on customers and their choice. In this more complex setting, attribution still needs to reflect the impact of a channel on sales. That impact, however, is difficult to quantify and estimate through statistical models because of interacting channels, interdependent touch points and active media consumption.
An analogy could be drawn using team sports. In basketball, for example, attributing points solely to the player that scored the field goal doesn’t really give the coach true understanding of the plays and players that led up to and contributed to those specific points. Indeed, any basketball team would not be successful if the coach decided to use players who can only score. He’d be forgetting players that pose a strong defense, players that can dribble the ball and players that can pass. These lower-scoring players may all contribute significantly to the overall success of a team. As “coaches” of our own marketing programs, we must look for ways in which we can utilize every team member (touch points) to achieve optimal team performance.
With limited data availability, outdated campaign tracking platforms and lack of an accepted methodology, marketers often try to solve the challenge of cross-channel attribution by adopting basic rules, such as last-touch attribution, in which 100% of the credit for sales is attributed to the final touch in the sequence. But this approach doesn’t give the marketer the complete picture.
There are several different approaches that marketers are using to go beyond last-touch attribution and obtain fractional allocation across multiple touch points:
First interaction, last interaction and position based assign attribution based on the position of a touch point along a typical touch path.
Linear treats all touch points as equally effective and attributes conversions evenly.
Customized allows an arbitrary attribution scheme that reflects managerial judgment on the position of the touch point, touch point type or traffic source.
Time decay systematically assigns a higher attribution weight to a touch point that is closer to the final conversion. These time-varying weights decay over the duration between the time of a touch point to final conversion, and the speed of decay can be controlled.
Note that the above attribution models allow a user to compare results across different models with different parameter settings. They provide a multi-faceted view depending on the position of a touch-point and judgmental input from marketing managers. In addition,
Attribution for multiple touch points can also be derived from proportionally adjusting last-touch attribution weight. The proportional adjustments are made to account for the influence of factors including frequency, recency, reach, ad size, creative content, channel interactions, among others.
Marketers can also use fixed percentages derived from simple rules or sequencing of the touch points.
At Experian Marketing Services, we take cross-channel attribution another step further by integrating established approaches with our online/offline data assets, offering a menu of progressively customized attribution solutions including the following:
Direct Attribution is applied when responses and actions can be uniquely linked to a channel.
Inferred Attribution takes advantages of Experian data matching technology to link campaign histories with responses at an individual level and infers attribution rules from the estimated relative impact of each channel on customers.
Fractional Allocation allows a marketer to assign and adjust prior channel importance measures, data matching quality indices and time-decaying factors in deriving attribution weights.
Touch-Point Attribution is a completely customized solution in which a marketer’s individual-level online/offline campaign data are merged with Experian data to obtain attribution rules that reflect the effect of campaign frequency, recency and channel interaction.
Despite these recent advances, marketers have yet to reach a consensus on preferred attribution methods or approaches. It remains to be seen whether standards and data technology will emerge to guide the industry toward the ultimate goal of developing reliable and consistent attribution practices that help optimize marketing budget allocation.
The available options, whether based on managerial rules or derived from data, provide an estimated range of attribution weights. Given this current state of affairs, especially for marketers who do not have the resources to seek customized and advanced solutions, one sensible solution is to start simple and add sophistication depending on needs and budget. The last-touch and first-touch attribution can be applied first to establish a baseline in which the contributions from intermediate touches are ignored. Then these baseline results can be compared with those from alternative methods that give attribution credit to intermediary touch points and channels. If attribution results are insensitive to alternative rules and the range of variation in attribution weights is small, a marketer can be confident in adopting any of the attribution rules attempted. On the other hand, if the range of variation is large and it is difficult to pin down a reasonable set of attribution weights, more sophisticated and customized methods such as those offered through Experian’s cross-channel attribution platform, can be pursued to derive attribution weights from customers’ transactional and behavior data and their actual media preference.
Remember, a winning team is never built on scorers alone!
View Original Article