Understanding the Customer Journey: A Primer on Multi-Touch Attribution
There is an old adage among marketers that “half of my advertising spend works, I’m just not sure what half”. Many in the online marketing community consider this a truism only for traditional marketing, such as print. Online marketing, it is argued, is extremely measureable – allowing marketers to accurately assess a return-on-investment of their efforts. This assumption is being challenged by multi-touch attribution (MTA).
In the most general of terms, MTA basically means assigning a value to each interaction in the customer journey. Sounds pretty simple, eh? It’s not. Here are a list of considerations that must be tackled for companies to build an MTA approach:
- How should we value the first click, middle clicks and last clicks of a customer journey?
- How do you track customer journeys in a multi-screen world (desktop, mobile, tablet)?
- How should time-decay (or latency) of the interaction factor into the MTA model?
These are difficult questions, with answers that vary depending on industry and even between businesses within the same industry.
The way to address these complex questions is two-fold: data and technology. Firstly, you need to have the data from all of your online (and offline, if possible) ad campaigns. Secondly, you need an appropriate technology in order to turn this data into useful information. This will allow you to build out and implement an MTA model for your company.
Where can you find the technology to do this? Analytics solutions, such as Google Analytics, now offer multi-channel views (see screenshot below). However, the information is somewhat limited due to the fact that only ad spend from Google can be easily brought into the platform. Most importantly, these platforms only allow you to view multi-channel trends and do not allow you to automatically ‘action’ the information.
There is also a small but growing number of specialized vendors, such as MarketShare, which focus-solely on enterprise-level MTA. In order to do this, they have developed complex, proprietary platforms that use sophisticated algorithms that take into account everything from display advertising’s impact on search volume demand to weather patterns. The end-goal of this exercise is to develop actionable recommendations (and even bidding algorithms) based on causality, not just simple correlation.
As you can imagine, the topic of multichannel analytics and MTA can be pretty daunting. Here are five easy steps to follow in order to get started:
- Identify and Implement Goals: Your online marketing spend needs to be evaluated against a meaningful metric. This can include revenue, transactions, lead generation form completion etc.
- Identify Your Online Marketing Channels: This includes both paid (example: PPC) and free (example: organic) channels.
- Check Your Analytics Solution: See what kind of multi-channel viewing abilities you have already in your analytics solution (as mentioned above, Google Analytics has a number of visualizations).
- Look at How the Channels Play Together: Are there complementary channels that work well together? Are there channels that drive conversions all by themselves?
- Test & Interpret: Try changing the level of spend for a marketing channel that has a strong complimentary channel (for example, if PPC and display ads share a large percentage of conversions, change the level of spend for one of these), observe the results (for example, see if raising spend on display leads to more conversions on PPC).
Need help identifying how well your marketing channels play together? Vovia can help determine where marketing dollars overlap and recommend how to be as efficient as possible with your online advertising bucks.