Subscribe to DSC Newsletter

Marketing measurement has long been an arcane field - companies interested in understanding how their marketing programs impacted revenue (or brand value) would hire expensive consultants who labored long and hard to deliver complex models at great cost to help their clients set high level marketing strategies and advertising budgets.

 

This worked well until the internet came along and changed the game - new digital channels and online marketing techniques were embraced by both marketers and consumers and resulted in data streams that could be used to attribute the impact of those programs on revenue.  Further, this could be done at a very granular level - a clear trail of clicks enabled marketers to determine how effective those digital marketing programs were in driving revenue down to the individual consumer.  And so — the discipline of revenue attribution was born.

 

These two disciplines have happily co-existed up until now - largely reflecting the organizational make-up of most marketing groups where the online team is quite separate from the traditional media marketing group.  But things are changing in the brave new world of marketing measurement.  What worked well in the past does not work so well in an environment where online and offline channels are converging.  Marketers have begun to see the inherent limitations of running two separate measurement silos.  For example, digital channels:

  •  attribute all marketing impact to the digital team and not giving credit to the impact of offline media to an online purchase, and
  •  a last click or first click approach, while neat, may not capture the impact of other consumer touch point.

 

On the other hand, traditional marketing mix models:

  • generate outputs that are generally limited to tables and charts of one or two media variables at a time; important detailed, granular ROI findings and opportunities can be missed,
  • may not highlight Important interactions between media and media specific sub-attributes (e.g. radio network, radio spot length, air time),
  • can lack flexibility when coupled with their optimizer function, 
  • can often take a long time to run optimization scenarios, limiting the resultant creativity and desire to experiment that users may have,
  • result in optimization scenarios that are hard to implement because the solutions require complex simultaneous reallocations over many media variables. At best the ROI results are aspirations or upper limits to what can be achieved. Many clients are left confused about what to specifically do or try. 

 

Today’s CMO needs a measurement solution that is more holistic, more highly automated, incorporates both online and offline data and provides actionable ROI results quickly.  In short, measurement systems should better reflect the way consumers behave and the way marketers manage their business.  Jim Nail at Forrester recognized this evolution and coined the term “Unified Marketing Impact Analytics” or UMIA to reflect this emerging need.  What exactly is UMIA?  Forrester describes it as a blend of marketing analytics approaches to gain a full view of marketing performance and assigns business value to each element of the marketing mix at both the tactical and strategic levels.  As a result, UMIA can be used more broadly to provide guidance on budget allocation, measurement of cross-channel effects and deeper path-to-purchase insights.

 

Marketing Decision Science’s (www.mdscience.com) SMX system using TreeOpt Optimizer addresses these issues.

  • TreeOpt is attribution model agnostic and flexible. It can accommodate for example, linear, non linear, logit, Adbudg attribution models. It can accommodate both mix models and attribution models because attribution and optimization are decoupled from one another.
  • TreeOpt can drive to any level of granularity supported by the data. Multi-way interactions between media attributes and ad scale non-linearities are routinely captured
  • TreeOpt solves quickly, even with large complex models, typically in 1 minute or less. Solve time can be dialed up or down within SMX by adjusting optimization parameters which adjust solution granularity and search depth. 
  • TreeOpt can simplify solutions to simple, budget neutral reallocations over a small number of test cells; e.g. reallocate across 6 well defined clusters of radio networks so that the total radio budget is constant and large ROI gains are still achieved. These automatically generated simplified reallocation strategies are easy to implement, understand and are completely explicit and spelled out. These can be used as tests or roll out plans.

 

Additionally, todays marketers need to be able to control their modeling environment and not have to rely on external consultants to run different scenarios.  As todays marketing teams become more analytically sophisticated, they want to develop their own hypotheses and run their own experiments.  UMIA is moving towards software platform based solutions rather than consulting services.  

 

SMX is a cloud based software platform that provides a more simplified, automated and easy to implement approach to the new world of UMIA.  What’s more, because of this approach, SMX is a very cost competitive solution.  No longer is marketing ROI analysis limited to large companies with big budgets - it’s an economical solution no matter the size of your marketing budget.

Views: 1112

Tags: CART, attribution, marketing, mix, modeling, optimization, revenue

Comment

You need to be a member of Data Science Central to add comments!

Join Data Science Central

Videos

  • Add Videos
  • View All

© 2019   Data Science Central ®   Powered by

Badges  |  Report an Issue  |  Privacy Policy  |  Terms of Service