“Attribution” has become one of the most cringe-inducing terms in B2B marketing. People have been burned or distracted by the influencer camp bashing all things data. Unfortunately, it’s to our detriment!
While attribution was intended as a campaign optimization tool, board members and CEOs pressure marketers to report how much pipeline and revenue they bring to the business. Single-touch models proved less than ideal in B2B because of buyer committees and circuitous buyer journeys. To date, multi-touch attribution is the best estimate to figure out how much pipeline and bookings marketing influenced versus other departments.
And it’s flawed for a variety of reasons.
There are three very vocal influencer camps shouting from the rooftops about attribution:
- The group that built their brand bashing multi-touch attribution
- The group that advocates for measuring ALL THE THINGS
- The group advocating for a middle ground
Unfortunately, the most reasonable category is not the loudest. And as someone sitting in camp number three, I’ve spent a lot of time thinking about how we got to where people poo-pooing data for marketers are celebrated. Yet, these same individuals advocate for practices we abandoned for a good reason and advise people to go against what works in the C-Suite.
If we understand why something went so wrong, we can make something better. So let’s explore how we got here and what we can do to fix it.
1. Vendors Oversold Multi-Touch Attribution as THE Silver Bullet
I remember when multi-touch attribution was first adapted to the digital world, and, boy howdy did vendors oversell its capabilities.
We all are susceptible to marketing. Everyone wants something easy to fix the tension with the sales team and our CEO. Unfortunately, the old saying, “if it seems too good to be true, it probably is,” hits home here.
It’s important to recognize three foundational truths when it comes to multi-touch attribution:
- Executives think of sales as a linear, one-person-one-touch journey. It’s not.
- Just because it’s digital doesn’t mean it’s trackable.
- We can’t (and should not try to) read minds.
A perfect storm developed after attribution became widespread, and now the knee-jerk over-correction is happening all over social media.
We took a tool meant to optimize campaigns and tried to prove marketing’s worth (more on that later). We deal with executives stuck in a linear buyer journey mindset. Some of our most influential selling tools (I’m looking at you, “word of mouth”) can’t be tracked.
Those of us who understand the magic of brand marketing also know we’ll continue to suffer from a lack of visibility into what matters. But that doesn’t mean we can’t use multi-touch attribution as a good-faith estimate that makes our CEO, board, and sales leader happy. To do this, we must level up our data game and succinctly communicate what is and is not possible.
2. It’s Used to Defend Marketing
Some brilliant marketing and revenue operations professionals will die on the hill that reads, “we should only use attribution for campaign optimization.”
I respect these people and see where they’re coming from, but I definitely disagree.
As a marketing leader, I have felt the tremendous pressure marketing leaders are under to develop a metric that the rest of the organization can believe in that shows how much pipeline and bookings their department brought to the table.
It’s hard for people with accounting or software engineering backgrounds to understand how complicated B2B buyer journeys are. Those people are used to balancing accounts to the penny or line-by-line code reviews. As a result, it’s hard to communicate the degree of subjectivity and ambiguity that will always exist in marketing.
Unfortunately, marketers who ditch data and embrace ambiguity fail when forced to prove how their team has impacted the business.
Having an estimate is better than coming to the table with a shoulder shrug. Marketers must learn how to make the most out of attribution, representing other departments’ efforts fairly in their models. For example, if you present a chart that shows all marketing tactics leading to pipeline, your sales leader will call “b*llsh*t. Collaborate cross-functionally when deciding which touchpoints you consider “influential.”
Too many marketers have been burned by attribution because it wasn’t implemented to support their primary objective.
3. “What I Have is Good Enough”
I get it. It’s tough to choose an analytics tool over a revenue-generating campaign when allocating budget. Instead of gaining access to the metrics your boss is demanding, you’re focused on building pipeline your boss is demanding.
However, anyone with variable compensation based on a single-touch model knows the pure hell you’ll experience when defending your bonus. I’ve had marketers beg, plead, and steal to get opportunities created against the “correct” contact record in Salesforce or throw a tantrum for a “primary campaign” on the opportunity to be changed.
Instead of focusing on what is best for the business, everyone is obsessed with a single influential point in time.
Getting attribution right is challenging, as is selling it to the rest of the organization. So it’s essential to find a vendor or consulting firm that can help you instrument attribution to support your primary objective, teach your team which model to use when, and help you socialize the concept with the rest of the organization in a way that fosters adoption.
This brings to mind one of our favorite buyer journey diagrams:
4. It’s Hard to Explain
When I’m on calls with prospective clients, and I ask why they’re ripping out their existing attribution tool, more than half the time, it’s because “it’s a black box.”
They don’t understand how it works. Which means they can’t defend it or explain it.
The rest of the time, I hear, “We aren’t using it.” But, when we dig into the “why,” it’s because they didn’t implement the system to help them defend marketing’s contribution versus other departments.
The most important part of setting up attribution is deciding which touchpoints are influential across departments. The model should include sales calls, form fills, and deals registration from partners. It shouldn’t have low-barrier or passive touches like web visits, email clicks, and email opens. Those low-value touches dilute return on ad spend.
Get agreement on the data and philosophy behind your multi-touch attribution model. Then pick the most straightforward model to explain.
I use even-weighted attribution over any model that has complex logic. As long as the sales leader agrees that we’re looking at the right signals, the rest doesn’t matter.
5. It’s Easy to Over-Engineer (And Then Abandon)
Two scenarios lead to over-engineering attribution models:
- The business intelligence team heard marketing wanted to do attribution, and it sounded like a fun project
- Everyone thought the attribution tool was wrong (because it wasn’t implemented to support how the business wanted to use it), so they’re bringing it in house
When I talk to prospective clients who decide to build their own attribution tool, I wish them luck and make a note to follow up in three to six months. The amount of work and thought that needs to go into a project like that is staggering. Chances are high that attribution reports won’t be available for a year.
I get concerned when I talk to a prospective client who wants a “custom model” and control the weight of each touch. This is usually a symptom that the sales team is blowing up attribution as “wrong” and “useless.”
I’ve said it before, and I’ll say it again. The key to getting adoption is agreeing on which touchpoints to include. After that, the rest is easy with the proper support.
Before you build your model in-house and get fancy with which personas and touch points get more points than others, ask yourself, “Am I doing this because the sales team or other executives don’t trust the model I have today?”
If the answer is “Yes,” a fancy model will only worsen matters. Trust can only be formed through understanding, which means accepting that the major hurdle to adoption is that the sales, product, and channel teams don’t feel their considerable efforts are fairly represented.
6. It’s Difficult to Maintain
We touched on this briefly in “What I Have Is Good Enough,” but I’ve felt this pain and need to communicate it.
When you use your database and logic, you have to anticipate a great deal of time will be dedicated to “fixing” the model whenever it “breaks.”
What breaks a model?
- Tool purchases
- Tool deprecation
- Customization of your CRM or marketing automation platform
It’s easy to assume that your current tech stack will always be as-is, but most operations professionals know better. Where our target buyers congregate (digitally speaking) changes quickly and without warning. For example, the somewhat temporary mass exodus of Twitter. This means your marketing team will need to change technology to adjust.
Every time a new tool is incorporated into your stack, it will take weeks to get that tool’s data integrated, then mapped, and then weeks to update every calculation that feeds your attribution model.
How to Fix It
I’ve seen attribution work for marketers. I’ve used it in board rooms and can attest to how much better my conversations with the board went when everything I reported was grounded in our business objectives rather than framed as “close enough.” It’s magic.
But it’s not easy.
To be able to rely on marketing to represent our efforts, we need to implement it with that goal in mind. This means we must represent other departments in our model and not cave to the temptation to look good by bloating our numbers with low-value touches.
Then we have to be able to socialize and sell the model. The only way marketing leaders can do this is to sit with operations and understand the model’s shortcomings.
This means admitting attribution is an estimate, and it’s the best we can do with today’s technology when it comes to representing marketing’s share of pipeline and bookings.