Patrick Reinhart

Most of HubSpot’s 200+ updates won’t move the revenue needle. Two will: Data Hub and Breeze Agents. • Data Hub finally fixes the attribution nightmare CMOs face every LT meeting. • Breeze Agents cut CAC and accelerate pipeline — but only if your playbooks are solid. Here’s the mistake I already see coming: 👉 Companies will flip these on with dirty data and broken processes. 👉 And instead of clarity, they’ll get chaos — faster and more expensive. The best operators will use this moment differently. They’ll clean their data, re-architect workflows, and point AI at the right accounts. HubSpot just drew a line… You either adapt your revenue engine for hybrid human + AI teams… or you watch your competitors pull ahead. (Full breakdown in the video 👇) #INBOUND25 #HubSpot #CMO INBOUND 🚀

26

1 Comment

Open-source validation frameworks are entering the MMM conversation. R², MAPE, stability tests, predictive accuracy benchmarks. Vendors that used to ask marketers to trust their numbers are being asked to prove them. 📊 I welcome this. Genuinely. For too long, MMM results have been passed around like they came from a crystal ball. Marketers nodded at coefficient charts they had no way to challenge. The shift towards transparent, reproducible testing is overdue. 🔍 But there is something I want to say carefully. Fit testing is the floor of validation. Not the ceiling. 🛑 A model can pass every statistical diagnostic and still produce a contribution figure that contradicts three years of market experience. A model with perfect R² can still attribute negative contribution to a media channel with USD 40 million in annual spend. A model with t-statistics above 2 on every variable can still produce a positive price coefficient in a category where higher price consistently suppresses volume. These are not edge cases. They are the kinds of results that emerge when statistical validation is treated as the end of the validation conversation rather than the beginning. Statistical validation tells you whether the model explains the data. It does not tell you whether the model is telling the truth about the market. Those are different questions. 🤔 I am all for testing. I want more of it. I just want us to test the right things, at the right depth, for what we are actually asking the model to do. The goal is not a model that passes diagnostics. The goal is a model the CFO does not dismiss.

47

17 Comments