B2B attribution was built for one buyer. Forrester puts your average buying committee at thirteen stakeholders. Here's what breaks, and what replaces it.
This is a lovely piece, especially to someone that has been saying attribution is broken for the better past of 15 years. It is also correlation. You also can't measure what you can't track.
Buyer journeys also don't start when they engage. It stars way before then in the dark funnel, closed communities, exclusive industry events.
When the buying committee comes to the table, they come ready with a day one consideration list.
That now depends more on trust which is earned- buyers are super risk averse at the moment. That trust also happens with consistent brand promise which requires penetration and reach, not assuming their problem (we see this most with the companies we work in your space - you lost that deal, because you don't), and talking about value.
For measuring all of those, you need causality. Because B2B is non linear, an open system, highly probabilistic and time lagged. Attribution, incrementality, econometrics don't cut it anymore.
Causal AI does - with counterfactuals in real time. Not based on past data that is old and outdated for you to base any activity on.
Agreed on the correlation point, that's the gap most of this debate skips entirely. Most of the journey is permanently dark now.
We've been saying attribution is broken since 2012, and the data backs up your dark funnel point too. Across the client sites we track (billions of page views now), roughly 80% of the buying journey happens before anyone fills in a form from what we can see. Closed communities, events, whatever earns trust before a company shows up in a CRM at all.
A committee arriving with a day one shortlist is exactly what most attribution models are blind to, because none of that ever left a trail to measure.
I love the idea of causal AI, my understanding is that solving the counterfactual problem in a non-linear, time-lagged, probabilistic system, in real time is still a gap. Are you aware of anyone who's doing that credibly right now? I haven't seen a vendor manage it without hand waving over the hard part!
This is a lovely piece, especially to someone that has been saying attribution is broken for the better past of 15 years. It is also correlation. You also can't measure what you can't track.
Buyer journeys also don't start when they engage. It stars way before then in the dark funnel, closed communities, exclusive industry events.
When the buying committee comes to the table, they come ready with a day one consideration list.
That now depends more on trust which is earned- buyers are super risk averse at the moment. That trust also happens with consistent brand promise which requires penetration and reach, not assuming their problem (we see this most with the companies we work in your space - you lost that deal, because you don't), and talking about value.
For measuring all of those, you need causality. Because B2B is non linear, an open system, highly probabilistic and time lagged. Attribution, incrementality, econometrics don't cut it anymore.
Causal AI does - with counterfactuals in real time. Not based on past data that is old and outdated for you to base any activity on.
Thanks Ayse :)
Agreed on the correlation point, that's the gap most of this debate skips entirely. Most of the journey is permanently dark now.
We've been saying attribution is broken since 2012, and the data backs up your dark funnel point too. Across the client sites we track (billions of page views now), roughly 80% of the buying journey happens before anyone fills in a form from what we can see. Closed communities, events, whatever earns trust before a company shows up in a CRM at all.
A committee arriving with a day one shortlist is exactly what most attribution models are blind to, because none of that ever left a trail to measure.
I love the idea of causal AI, my understanding is that solving the counterfactual problem in a non-linear, time-lagged, probabilistic system, in real time is still a gap. Are you aware of anyone who's doing that credibly right now? I haven't seen a vendor manage it without hand waving over the hard part!
You need to talk to @markstouse about CausalAI. He has a book coming out in Jan on this too. He is a dear friend and brilliant!!
Look at Proof Analytics- he works directly with the C suite on GTM effectiveness, not even the marketers😉
Sounds great Ayse - I always appreciate your input. Thanks once again for some great tips!
My pleasure. It is nice this back and forth. I really enjoy your articles!