3 Challenges Of Attribution Modeling: The Bad, The Bad And The Ugly

Any method of attribution has its strengths and drawbacks. Contributor Kohki Yamaguchi outlines factors every marketer should be aware of.

Chat with MarTechBot

shutterstock_181286708=devices-touchpoint-attribution

Multi-touch attribution is a modern approach to modeling marketing impact.

Back in May, three attribution technology vendors — Adometry, Convertro and DC Storm — were acquired by GoogleAOL, and Rakuten Marketing respectively. Clearly, attribution is now being seen as integral to the future of digital marketing, and is rapidly rising in priority for many marketing organizations.

Today, there are many different approaches to attribution floating around, ranging from the basic to the incredibly sophisticated. While the advancement of attribution is an exciting development, there are still many limitations on model accuracy — and each approach has its own strengths and drawbacks.

Understanding these caveats should help marketers choose the appropriate attribution solution, comprehend results, and better act on the insights and recommendations generated.

(For a broader review of different approaches to marketing impact modeling, please refer to my previous article on this topic.)

A (Very) Short Primer On Multi-Touch Attribution

Multi-touch attribution is a way to allocate credit towards marketing touchpoints that preceded a conversion within a customer journey.

Screen Shot 2014-09-24 at 12.50.00

How credit is allocated in multi-touch attribution

By looking not only at the last touchpoint before a conversion, but the entire sequence of conversions (based on some lookback window or path definition), it becomes possible to better allocate credit between channels and campaigns that tend to occur earlier in the customer journey vs. those that occur later.

3 Challenges Of Attribution Modeling

Though attribution models range from basic, rules-based models to advanced algorithmic models, they share some persistent problems that are difficult, or in certain cases impossible, to fully solve. Below are three commonly encountered issues.

(Please note: all points below are generalizations, and do not speak for any given attribution solution. The list below can be used as part of a checklist when choosing an advanced attribution method or vendor.)

1. Attributed ROI Is Not True (Incremental) ROI

The whole idea behind attribution is to divide up credit for a conversion amongst touchpoints preceding it, so you can determine what’s working and structure future investments accordingly. However, marketing in the real world does not work in such a simple manner.

For most businesses, there exists a customer base that was built up over time and feeds a good portion of recurring revenue. In addition, there is existing brand equity that influences customer decisions. There are also various offline marketing efforts that are untrackable.

The digital marketing touchpoints that led directly to an individual sale only contribute an incremental impact on top of all of these.

Screen Shot 2014-09-24 at 12.51.33

Attribution models versus reality

So in some ways, attribution is based on a rather distorted view of reality. While it is possible to still reach a recommendation on relative budget allocation, the attributed ROI does not actually represent the true ROI of the channel, campaign, or creative.

2. Attribution Does Not Account For Offline To Online Effects

When looking at the relative impact of digital channels but not taking offline media into account, determining the influence of offline marketing on digital channels becomes an issue. To illustrate this problem, I will take the example of paid search.

As everyone knows, paid search is driven by search engine queries. The volume or likelihood of search queries for a particular brand or product depends on consumer interest at any given time.

Consumer interest, in turn, is influenced by total marketing effort. This means that paid search impressions, and therefore clicks and spend, are almost impossible to disentangle from the indirect influence of other types of marketing and promotions on search queries.

This problem makes it difficult to estimate marketing impact accurately for certain marketing channels that occur later in the customer journey. With query volume data, it is possible to attempt to separate these effects using statistical or algorithmic techniques, but the results will be of varying accuracy depending on the approach: sometimes good enough to inform budget allocation, sometimes not.

3. Attribution Does Not Account For Influence Of External Factors

Similarly to numbers one and two above, most types of attribution models do not account for other external factors that influence sales and marketing effectiveness, such as pricing, promotions, seasonality, economy, and more. These may seem like peripheral issues, but the amount of influence they have over attribution results can be quite significant.

For example, according to decades of academic research, pricing with respect to competition is known to have 20-25 times greater impact on sales on average compared to the total effect of advertising across all channels.

This means that if spend, effort, or reach on certain channels are linked to discounts and promotions — which is quite often the case — then the effect of advertising can be biased or even completely washed out by the price or promotional effects, resulting in an unreliable read on advertising impact.

So Why Attribution?

Having read this column thus far, hopefully you are not disillusioned by gaps in the promise of attribution. If you are, then let me say that — despite all of these problems — I am still a firm believer in the future of multi-touch attribution. Here are the reasons:

1. It Is Still Infinitely Preferable To Last Touch

Though the problems of multi-touch attribution are difficult to overcome and sometimes impossible to fully resolve, last touch attribution has all of the same problems made 10 times worse. Almost any type of multi-touch attribution model is preferable to a last-touch model in terms of getting one step closer to understanding true marketing impact.

2. It Represents The Future Of Actionable Marketing Impact Modeling

While other approaches such as media mix modeling account for some of the above issues, attribution is still the only way to gain visibility into marketing impact within a short time frame. With more and more budget moving into digital and real-time marketing, the need for marketing agility will tip the scales in favor of models that are immediately actionable.

3. Attribution Methodologies Are Improving

There are companies that do understand the challenges listed above and more, and are working to take them into account when building their models and algorithms.

No Model Is Perfect

As industry thought leader Avinash Kaushik put it:

[blockquote]…every attribution model has built into it biases and opinions that often struggle to stand any intellectual scrutiny, or the simple laws of common sense.[/blockquote]

Even the best of algorithmic models will fail to take into account some factor that causes results to be off track. The important thing is to test, predict, compare against results, and iteratively improve the approach to get us gradually closer to the holy grail: understanding the true impact of our marketing efforts.

Marketing attribution and predictive analytics: A snapshot

What it is. Marketing attribution and predictive analytics platforms are software that employ sophisticated statistical modeling and machine learning to evaluate the impact of each marketing touch a buyer encounters along a purchase journey across all channels, with the goal of helping marketers allocate future spending. Platforms with predictive analytics capabilities also use data, statistical algorithms and machine learning to predict future outcomes based on historical data and scenario building.

Why it’s hot today. Many marketers know roughly half their media spend is wasted, but few are aware of which half that is. And with tight budgets due to the economic uncertainty brought about by the COVID-19 pandemic, companies are seeking to rid themselves of waste.

Attribution challenges. Buyers are using more channels and devices in their purchase journeys than ever before. The lack of attributive modeling and analytics makes it even more difficult to help them along the way.

Marketers continuing to use traditional channels find this challenge magnified. The advent of digital privacy regulations has also led to the disappearance of third-party cookies, one of marketers’ most useful data sources.

Marketing attribution and predictive analytics platforms can help marketers tackle these challenges. They give professionals more information about their buyers and help them get a better handle on the issue of budget waste.

Read Next: What do marketing attribution and predictive analytics tools do?


Opinions expressed in this article are those of the guest author and not necessarily MarTech. Staff authors are listed here.


About the author

Kohki Yamaguchi
Contributor
Kohki Yamaguchi leads product marketing at Origami Logic, a cross-channel marketing intelligence solution for modern marketers. With a career of 8 years in marketing and analytics spanning various functions, Kohki's focus has always been on translating data into strategy, simplifying the complex, and bridging the gap between data and organizational silos.

Get the must-read newsletter for marketers.