The Past & Present of Marketing Attribution

 

Gather round, fellow SaaS devotees and metric-obsessives, because we’ve got a story of sorts—and even if you don’t fit into either of those categories (yet), go ahead and listen anyway, because this stuff’s interesting. We swear.  

 
The Past
 
Long ago, in a land not so far away—that is, starting in the ‘50s and continuing into the early ‘90s—marketing attribution went by another name. Marketing Mix Models, MMM for short, were the snail mail equivalent to today’s attribution models. They were painfully slow to offer results, which rendered them an untenable solution for today’s hyper-connected, instant gratification world. 
 
Imagine, as a marketer, having to wait for weeks to get those valuable insights into which things are working and which aren’t? Honestly, bone-chilling. 
 
Still, they (and their influence) lingered. Enter: The marketing attribution models of today. 
 
The Present
 
Now, first things first: There are a lot of attribution models. Dare we even say too many—each one built on a towering foundation of assumptions that need to be made in order to justify the weight assigned to each channel or touchpoint within your funnel. 
 
You’ve got single-touch models and multi-touch models, U-shapes and W-shapes. If you’re new to this, let us break those down real quick: 
 
First Touch: Which channel was your customer’s first point of contact? With this model, they get all the credit for the conversion. 
 
Last Touch: You can probably deduce for yourself who gets the credit in this model, but we’ll lay it out anyway: It’s whichever channel made contact last prior to the conversion. 
 
Linear: This multi-touch model gives equal credit to any channel that interacted with the customer along their journey. Sharing is caring. 
 
Time Decay: Here, timing is everything. The closer the touch point was to the conversion, the more credit is given to the channel that made that contact. 
 
U-Shaped: Moving on to the shapes! With a U-shaped model, the highest percentage of credit is given to the first and last touches (about 40% each, usually), with the rest being evenly distributed amongst the middle touch points. 
 
W-Shaped: This model is not too different from its U-shaped counterpart, but does require more division. Here, 30% each is given to the first and last touch. Another 30% is given to the touch point that coincides with lead creation. The remaining 10% is divvied up between all other touch points.
 
With all these models to choose from, and each channel scrambling for its fair share of credit, it’s easy to lose sight of why attribution is important in the first place: insight. As with any other type of data, the really good stuff is in the story being told and the decisions that story leads to. Marketing attribution is more than just a vanity project, and if you can't use it to make important decisions, it all becomes, well, worthless. 
 
For example: Those with short, sweet sales cycles might gravitate towards single touch models. If you see your typical customer journey as pretty much “one and done,” these models will reflect just that. However, there are cautionary tales to be told about organizations that are too quick to write off the earlier touches in their funnel. Even if you do have a relatively short sales cycle, the chances of a single touchpoint deserving 100% of the credit for a conversion are pretty slim. As such, you run a serious risk of overspending too early or late in your funnel, and neglecting valid touch points that, in reality, are significantly contributing to conversions. 
 
For our friends with long sales cycles, things are often even less straightforward. Deciding how much weight to give to each channel gets harder the more touch points there are along the way. When multiple touch points exist per channel, data becomes even more critical to cut through the added layer of haze. Be mindful that you're not falling into common traps--for example, giving a touch point too much credit simply because you spent a lot on it and want it to be worth it. Give touch point analysis the time it deserves; understanding why something worked is just important, if not more so, than just the knowledge that it did (or, k'now, didn't). 
 
Recently, our CSO, Jen, took part in a panel organized by People.ai. Though the panel was about marketing challenges in general, Jen and panelists Kevin Liu and Beau Vasquez—from Mongo DB and DocuSign, respectively—chatted a bit about how their companies approach attribution. 
 
Kevin stressed that not all channels can or should be treated the same, which is why Mongo DB looks at attribution in two different ways. When it comes to social media channels, which generally have a lower level of engagement than other, more influential content, Mongo DB uses a W-shaped model. For more engagement-heavy, influential channels (webinars, whitepapers, etc.) they go with even spread attribution. This means that each campaign that touched a given opportunity gets an even amount of credit once that opportunity converts. 
 
Since DocuSign sells to customers than span from single-user to enterprise, Beau called attribution in their case "tricky", saying there is "no one silver bullet." For them, the key is to utilize tools that can help with insight into DocuSign's wealth of data, so they always understand what their teams are doing, and what's working. 
 
So, which model should you put your faith in? The answer to this question depends on your organization and the nature of your customers’ journey. However, there are a few questions to consider that will help set you on the right track:  
 
1. What is the length of your sales cycle?
2. On average, how many touches do you have per lead?
3. What are the different interactions points in each of your channels, and how to they measure against one another? (e.g. What is a content download worth when compared with a view? How would you weigh an event registration versus actual attendance? Do different types of content contribute the same amount towards conversion?)
 
 
 
Previous Article
Data Stewardship: A Recipe for Actionable Insights in SaaS
Data Stewardship: A Recipe for Actionable Insights in SaaS

Next Article
SaaS's New Structure: Building Your RevOps Team
SaaS's New Structure: Building Your RevOps Team