Six Ways to Use Attribution Modeling in the Google Digital Marketing Stack
Ashton Howe  /  
August 22, 2017

The Google Digital Marketing stack gives you a matrix of options to meet your attribution needs


As we’ve discussed before, a strong marketing program requires the ability to clearly measure ROI. It requires quantifiable efficiency, or a way to understand how your marketing dollars are impacting conversions at a level granular enough to guide optimization.


The ability to tie your media performance through to conversion, and understand with concrete metrics where to shift budget for optimal return, is the fundamental definition of marketing attribution. A more thorough definition comes from the Interactive Advertising Bureau:

Marketing Attribution is the process of identifying a set of user actions (‘events’) that contribute in some manner to a desired outcome, and then assigning value to each of these events.


Attribution is a hot topic – as an analytical instrument, value can be tied directly back to your bottom line – and its relevancy will continue to grow as more accessible modeling options are released.


But building out an attribution program can be intimidating; internally, there will be technical considerations, resourcing considerations, budgetary considerations, and even cultural considerations (How will your paid search team feel vs. your programmatic display team? What does your analytics team think? Will your direct mail department even acknowledge the tool?)  And most importantly, which attribution solution is best for your needs as a company?


A rule of thumb at Delve is that the best solution is the one flexible enough to adapt to your business goals. Out of box solutions can be a good starting point – they can get you most of the way there, but to truly excel in analytics you need technology agile enough to account for nuance. One size does not fit all in data.


And it’s for this reason that we’re so head over heels for the Google Digital Marketing (GDM) stack: powerful flexibility. Specifically, GDM provides native integration across a diverse set of products, which gives you enough dexterity to customize measurement to your very specific goals. This is especially important within the context of attribution, and GDM does a great job of providing you with something most competitors will not: options.


Let’s delve deeper into some of these opportunities:


Baseline Attribution Modeling in Google Analytics (Free)

Google Analytics (Free) offers seven attribution models that can be quickly applied across your repository of Analytics data. This lets you generate and compare different attribution views on the fly, so you can adapt the model to fit diversified tactics. Baseline Attribution Modeling in Google Analytics (Free) is a great way to test the waters of attribution with your current Analytics data. You can explore which models work best for your business use cases, and experiment with reporting and optimization based on this exploration. This provides you with a very clear opportunity to scope for best use cases internally, prior to committing to a larger scale solution.


The seven available models in Google Analytics (Free) are:


  •  Last Touch Attribution Model: Assigns 100% of conversion credit to the last touch point in a conversion path. This is the default model applied in Google Analytics multi-channel funnel reporting. It is a beneficial model for advertisers who want to measure performance for closing engagement, and is the legacy standard for attribution modeling.


  •  First Touch Attribution Model: Assigns 100% of conversion credit to the first touch point in a conversion path. This is a beneficial model for advertisers who want to measure performance for top and mid funnel initiatives (e.g. Prospecting or Brand Awareness campaigns).


  •  Linear Attribution Model: Assigns equal conversion credit to all touch points in a conversion path. This is a beneficial model for advertisers who view all marketing initiatives as equal contributors, and is a great introduction to multi touch attribution.


  •  Time Decay Attribution Model: Assigns incrementally more conversion credit to the touchpoints that are closest in time to the conversion. This is a beneficial model for advertisers who want to measure performance for bottom funnel-heavy initiatives.


  •  Position Based Attribution Model: Assigns 40% of conversion credit to the first touchpoint, 40% of conversion credit to the last touchpoint, and distributes the remaining 20% of conversion credit across all middle touchpoints. This is a beneficial model for advertisers who want to measure the synergistic impact of top and bottom funnel initiatives.


  •  Last Non Direct Click Model: Assigns 100% of conversion credit to the last non direct touchpoint in a conversion path. This is a beneficial model for advertisers who want to measure bottom funnel paid and organic campaigns, while deprioritizing Direct as a channel.


  •  Last AdWords Click: Assigns 100% of conversion credit to the last Google AdWords click in a conversion path. This is a beneficial model for advertisers who want to measure performance for bottom funnel AdWords campaigns.


Custom Attribution Modeling in Google Analytics (Free)

If the above options don’t quite fit your needs for modeling against your Google Analytics (Free) data, there’s a supplementary Custom Attribution Modeling functionality that allows you to create and apply your own personalized model instead.


This feature gives you the ability to build a rules-based model from the ground up, to assign conversion credit to touchpoints in a way that makes the most sense for your business. Weight different campaigns and engagements in a way that reflects your individualized digital initiatives; deprioritize or filter out touchpoints based on goal; favor certain channels or creatives based on seasonality. The options are extensive, but there is one limitation to remember: Custom Attribution Modeling in Google Analytics (Free) is click based only – i.e. the conversion path against which your models can be applied will only include clicks, and will not offer insight into impression value.



Display Attribution Modeling in DoubleClick Campaign Manager

But what if understanding impression value is your primary concern, especially within the context of your full paid marketing mix?


Centralizing all view and click based media in a single ad server like DoubleClick Campaign Manager will provide you with a consistent data repository for media measurement. This process deduplicates conversions across your multiple advertising platforms, generates a single source of truth for metric volume, and – in addition to allowing you to apply many of the same attribution models from Google Analytics (Last Touch, First Touch, Linear, Time Decay, Position Based) – provides additional modeling capabilities with:


  •  Floodlight Based Attribution Model: Assigns 100% of conversion credit to the last click in the conversion path. If there was no click in the conversion path, it assigns 100% of conversion credit to the last impression.


  •  Social Based Attribution Model: Assigns partial conversion credit across all touchpoints, with impressions weighted to account for social interactions.

This is a great solution if you’re looking for basic, pre-built attribution modeling for view based channels.


Data-Driven Attribution Modeling in Google Analytics 360

As the intricacy of your marketing initiatives grows, the attribution offerings in Google Analytics (Free) and DoubleClick Campaign Manager alone may no longer fit your needs. At that point, you’ll likely require a stronger model offering, more robust in its logic, to help you more confidently optimize against marketing mix complexity.


The Data-Driven Attribution Modeling available in Google Analytics 360 provides just that: an algorithm that assigns partial conversion credit to touchpoints based on the Shapley Value solution from cooperative game theory, with user level measurement that accounts for cross device browsing. Quite a mouthful, right?


What you need to know is this: with Data-Driven Attribution Modeling in Google Analytics 360, Google does the heavy lifting for you. It provides you with recommendations based on a self-adjusting algorithm to help you take ideal action against your media buys.


And most importantly, because Google Analytics 360 plugs directly into DoubleClick Campaign Manager, AdWords, and Google Display Network, you’re able to apply this intelligence against a full view of your digital events – clicks, impressions, and site landings – to understand not only how click-based channels are performing, but how view-through is contributing as well.



Custom Attribution Modeling in Google Analytics 360 + DoubleClick Suite + BigQuery

But what if you’d like to take customized action against the full view of marketing events? What if you need to bridge the gap between paid and organic, owned and earned media, and need to ensure your solution accounts aptly for undulating, cross channel, multi device customer journey? What if you have onboarding needs for CRM, email, DMP, or offline transactional data?


The integration between Google Analytics 360, the DoubleClick Suite, and Google BigQuery provides you with a toolset to merge event-level site, media, conversion, audience, and offline data into a centralized cloud-based warehouse. Additionally, you can natively integrate event- level data from your other Google products directly (AdWords, YouTube, etc.), and layer in peripheral metadata at the user level to complete a unified view of both marketing and audience.


Once data is centralized, build a (or multiple) custom algorithm(s) to deploy programmatically. The largest benefit: you can build a customized, cutting edge attribution program you’re able to fine tune, while retaining the analytical elasticity to pivot toward new perspectives as they emerge. And with the upcoming 2018 release of Google’s Ads Data Hub, you’ll be able to incorporate cross device measurement natively across all Google media impressions, providing you with the cookieless measurement you need to optimize to the user-centric experience.



In Beta: Attribution Modeling in Attribution 360

If you’re ready for cutting edge attribution but aren’t resourced to bring the above customized solution build in house, you can reach out to a Google Analytics Certified Partner** to help you execute against your goals.


Alternatively, in 2018 Google will be releasing a new enterprise-level attribution solution to provide full scale, data-driven modeling against all your marketing drivers.


Attribution 360 promises to simplify data collection (for online, in store, television, etc.), streamline the view of all performance metrics, provide clear cut and accurate media recommendations based on data-driven algorithms, and integrate directly with your ad servers, DSPs, and search management platforms to pass attributed metrics back into the media buying tools for immediate data activation.


While modeling will mirror the current Google Analytics 360 offering, Attribution 360 as a standalone product will offer a more comprehensive tool set in terms of plug-and-play data onboarding capabilities, customized reporting, and – most importantly – the ability to optimize media in your own tools against attributed metrics. This offering will of course come with a price tag, but will provide a very competitive option for advertisers looking explicitly for closed-loop attribution automation.


What This Means for You

At the end of the day, attribution should be pursued in a very personalized manner. A one-size-fits-all modality that doesn’t cater to your business nuances is at best frustrating, and at worst dangerously misleading; you need a set of tools that allows you to customize functionality to your individual marketing mix. As attribution continues to expand into a fundamental element of the digital marketing docket, it’s important for attribution providers to keep up with the ever-evolving needs of their clients. The flexibility that the Google Digital Marketing stack provides is unprecedented in its capacity for customization, its ability to fine tune measurement, and its efficiency in pivoting seamlessly between views. After all, an attribution tool’s job isn’t to sell you on what it sees, it’s to empower you to see for yourself.


**Delve is a Google Analytics Certified Partner


Ready to learn how Delve can maximize GA 360 for your business?

Delve is a top-rated Google Analytics Certified Partner with experience managing hundreds of web analytics implementations world wide.

SEE EXAMPLES of our experience and reviews from our clients .

Shoot us an email or give us a call. We’d love to discuss how Delve can help your business win smarter.