Experts Point To Buying Groups & Bad Data As Impediments To Attribution Modeling

Experts agree that attribution is difficult these days; so much so that Demand Gen Report’s 2020 Marketing Measurement and Attribution Survey Report revealed more than half of respondents (59%) said they are not doing attribution analysis in their marketing measurement efforts at all.

“A big part of the reason that people aren't doing [attribution] as much as you might think is because [these models] don’t tell you what you think they're telling you,” said Kerry Cunningham, VP, Principal Analyst of Marketing Operations at SiriusDecisions, in an interview with Demand Gen Report. “And it's almost always bad news for the marketing organization. If you look at the first touch or last touch — or even the fanciest of multi-touch attribution models — it really undervalues the influence marketing has.”

“If I get 500 leads this month and I look at those 500 leads as a representation of a possible selling opportunity, my conversion rate is based on how many of those reports end up as opportunities divided by 500,” said Cunningham. “But the truth is, when you look at the data, what it really looks like is those 500 leads actually represent maybe 200 different opportunities — often 150 or maybe 100 different opportunities — because many of those leads are coming from the same organization, and they're about the same thing.”

Taking Multiple Stakeholders Into Account When Measuring Marketing’s Value

As modern organizations have formal buying groups in place to make purchase decisions, it is no longer enough to cater to a single person and measure the value of your efforts based on that single lead alone.

Demand Gen Report’s latest B2B Buyer Behavior Study indicated that 61% of respondents said the number of team members involved in the purchase decision had increased over the prior year. Furthermore, 67% of buyers brought in other team members to help with the research process within the first three months, and 71% of the respondents said they had a formal buying committee in place to review all purchase decisions.

“Our data shows that 81% of the time, the B2B buyer is at least three people, and about half the time it's five people,” said Cunningham. “Each one of those people does some of their own research. So, if some of those people are showing up as leads all the time, your lead data is very lumpy. When you have only one lead from a given prospect account, it's highly unlikely that that account is actually in market. If you only get one lead from that account in 60 days, that account is not going to buy something from you, and they're not going to buy something from you until more people show up. So, we believe one of the best signals when more than one individual from an account is showing up and looking at your content.”

Cunningham shared an example of a SiriusDecisions customer that flipped the switch on its attribution efforts to better analyze and report on specific leads within an account. The company had a small number of accounts (in the hundreds) where it already had the data on people for cross-sell and upsell efforts.

“In those accounts, they have a lot of contacts in Salesforce,” he said. “So, when a new lead comes in, they check to see if they already have the contact. If they already do have the contact, they just send a notification to sales that says, ‘Hey, this contact that we already know about in your account is active.’ There's no lead that goes over to sales; it's just a notification.

“And then they've got that other situation where they've got new leads coming in from those accounts and only the first one of those leads actually could ever make its way on to an opportunity,” he added.

Now the company matches every lead to a specific account. If that lead is from an account that is attached to a selling opportunity, the company opens it up as an opportunity for sales. Then, if every lead that comes in from that account is interested in the same solution, they are also attached to that opportunity. As a result, the BDR/SDR doesn’t qualify each individual lead. Instead, they qualify opportunities when three or four people are attached to it.

“Just by making that change, they reduced the number of things that they sent to the BDR team by half,” said Cunningham. “They increased the productivity of the team by 134%. So, they've dramatically increased the output of the team. They produce the same output with about 20% fewer BDRs because it's just so much more efficient. Now marketing has reporting that says, ‘Well, it looks like we're most successful with an account where we get three leads on that opportunity or two leads on that opportunity.’ Now marketing can report on the influence it has.”

Focusing On The Data, Not The Models

Another challenge with attribution is that most organizations’ data is unorganized. The same Demand Gen Report Measurement survey cited above showed that true marketing measurement is a challenge for 36% of organizations because of incomplete and messy data. Allen Pogorzelski, VP of Marketing at Openprise, a data orchestration solution, emphasized the critical need for good data in order to effectively pursue attribution modeling.

“At the end of the day, attribution isn't about the model; it's about the underlying data, and nobody has good underlying data to make it work,” he said. “Most salespeople create an opportunity with no contacts or one contact. And the reality is, in a buying group, on average, there are three to five people making that decision. Data is already so wrong, that traditional attribution is really, really hard to do. So, a lot of people don't and then they're just kind of flying blind.”

Data is the foundation of all aspects of marketing — including measurement and attribution — and with a majority of B2B organizations pivoting their strategies for a digital-only world, it has become even more critical to have a 360-degree view of the buyer.

“You need clean data in order to do attribution,” said Mike Geller, Co-Founder and CTO of Tegrita, a marketing automation and strategy consultancy. “In general, attribution is a ‘go-forward’ thing. It’s very hard to go backward and clean the data up and try to make the connections.”