Don’t you wish B2B marketing had something as effective as Facebook to help us target our ideal customers?
I’ve been running Facebook ads for a long time. In fact, I’ve pretty much been on the platform since it was a thing. My expertise with the platform is predominantly because of my background as a B2C demand generation specialist. Simply put, Facebook offers some competitive advantages that most other platforms couldn’t dream of. These advantages are mostly because of the amount of data that they possess about those of us who participate in the ecosystem (which is to say nearly everyone).
For instance, every page you like, every post you click ‘like’ on, every brand you follow, all of the information you input about your job, your title, your education, and even your day to day browsing activities on platform contribute to your “graph,” which is Facebook’s way of describing the totality of your online persona. Facebook can then take that information and use it to help its advertisers target more efficiently. One way that this is done is through the use of “lookalike audiences.” Here’s how it works:
How lookalikes work, and their effectiveness.
You upload a list of, say, 500 high-value customers with whom you’ve already worked. Facebook takes that list, finds those people on the Facebook platform, and uses an Artificial Intelligence to deconstruct the ‘graph’ of each successful customer. Then, the AI looks at all of the commonalities between those successful customers and creates a ‘persona’ of sorts.
The real magic happens next.
The platform can then use predictive analytics to look at all of the people on Facebook and produce a list of the top 1% who are the most algorithmically similar, or the “lookalikes.” It’s not uncommon for those lookalike audiences to shave 50% or more off of the cost of acquisition in a B2C environment. Behold:
You want to take a guess at which adset has the lookalike audience applied? If you said the top one, you’re starting to get the point. Although it’s hardly a conclusive data set, the application of a lookalike audience to this particular advertising set resulted in something like a 53% reduction in CPL and resulted in two ‘purchases’ against none. Although we’ve not made much of an attempt at an empirical study here, this is not an isolated result. Similar results can be found everywhere with a google query.
Is B2B left out in the cold?
So it’s obvious that lookalike audiences are the future of targeting. The problem for B2B marketers, as of this writing, is that there really is no platform upon which a similar targeting structure can be built for B2B demand generation specialists. LinkedIn aside, nobody is really volunteering information about themselves as it relates to their business units, purchase behaviors, etc. More alarming, perhaps, is the difficulty of penetrating account based clients with this sort of targeting. So what happens next? Is the B2B space just out of luck?
Welcome to the nerdery.
It’s time to toot our own horn a little bit, because LeadCrunch[ai] has actually cracked this egg. We’ve kidnapped an elite team of Data Scientists from places like Stanford and locked them in a room with unlimited Redbull and Doritos until they figured it out. What they came up with are called ‘vectors.’
Vectors are similar to Facebook’s ‘graph,’ except their tentacles reach out far and wide into the B2B business space and seemingly touch everything. Number of employees. Revenue. Job titles. Sure, but that’s the easy stuff. These are simple commodity firmographics such as have been at play forever in B2B marketing.
Think deeper, more meaningful ways to slice up a company and you’re beginning to get the point. We can look at not just how many people a company has and what its industry is, but also the skills the employees possess, their educational backgrounds, their tenure at the company, their years of experience, and the profiles of the company’s leadership. The results can be stunning.
We can deliver B2B lookalike audiences.
We can deliver predictive account lists, plus the contacts in those accounts, plus human-verified accounts.
And it works. Some of our results are ridiculous. For instance, we saved Oracle+Bronto a ton of money on acquisition costs by switching to B2B lookalike modeling that doesn’t suck.
Is this the future of B2B demand generation?
Oracle+Bronto saved 67% on acquisition costs. It’s hard to imagine an industry walking past a technology capable of that level of disruption to the status quo. That being said, the data nerdery is a beehive of activity as the team continues to refine the vector approach, adding new dimensions and ways to slice and analyze the B2B space in ways that have never been attempted, and the results can be expected to cascade with the rising level of sophistication in the A.I.
The two key takeaways here are 1) you should probably buy stock in RedBull, because we’re going through ten cases a day. 2) You should consider a trial of our new B2B lookalike targeting capabilities. We can produce results with really reasonable test budgets and you can buy human-verified, A.I generated leads on a CPL basis. You’ll be the envy of the water cooler community for being the first one to find it. It’s the first step to winning the love of the sales team and being revered as the demand generation monster that you’re meant to be.