It's 2025, and I'm still having this conversation.
I pulled out my old Pedowitz Group shirt from 2010 for this week's Revenue Marketing Raw because here's the uncomfortable truth: most marketers are still using lead scoring models from that era. Some haven't touched their scoring logic in 15 years. Others—and this blows my mind—haven't even turned it on.
30 to 40 percent of marketing organizations with marketing automation don't have lead scoring enabled.
Every platform on the planet has this capability. HubSpot, Marketo, Eloqua, Salesforce Marketing Cloud—all of them. And marketers are spending tens of thousands, sometimes hundreds of thousands of dollars on these systems and can't be bothered to flip the switch.
What the hell?
Back in the early Eloqua days (I was employee #32, Debbie was customer #12), I helped create one of the first lead scoring frameworks. The A1/B1/B2/C model—demographic fit plus behavioral score. At the time, it was revolutionary. It was the best we had because there was no other intelligence available.
That was 2004.
Here's what's insane: that exact model is still being used today, unchanged. Marketers are sitting in rooms with pieces of paper, guessing. "Let's give this activity 5 points. Let's give that activity 8 points."
Based on what? Intuition. Hope. A dart thrown at a board.
Meanwhile, they're drowning in data that could actually tell them what predicts revenue. They just won't use it.
I hear it constantly: "We can't do lead scoring because our data is so bad."
Bullshit.
You have pockets of clean, usable data right now. Your website activity data, for starters. Instead of blindly assigning 5 points for a page view and 8 points for a form submit, try this radical idea: look at what your closed-won customers actually did.
Pull your last two years of opportunity and lead history. Analyze the activity patterns of the people attached to deals that closed. What pages did they visit? What content did they consume? What was the progression?
Now you have objective data instead of guesses. That's not rocket science. That's basic analysis that nobody's doing.
Here's another problem nobody wants to talk about.
If you're in B2B with a considered purchase, you're not selling to one person. There are 10 to 15 people on that buying committee. But your scoring model? It's obsessing over one individual's score while ignoring the other decision-makers entirely.
Where's the account score? Where's the intelligence about the buying committee? You're basing your pipeline and forecast on one person when the actual purchase decision involves a dozen.
This is why your "qualified" leads aren't converting. You're measuring the wrong thing.
Let me ask you a question: What percentage of your B2B revenue comes from existing customers?
For most companies, it's 50 to 60 percent. Sometimes more.
Now look at your lead scoring model. How much of it accounts for install base activity? Cross-sell opportunities? Upsell signals?
If you're like most organizations, the answer is zero. Your entire scoring model is built around net-new acquisition, which represents maybe 40 percent of your revenue opportunity. You've already lost before you started.
One of our life sciences clients had this figured out. They sold expensive scientific equipment with disposables. They knew the run rate on those disposables for every customer. When consumption started dropping, alarms went off. Emails went out. Sales followed up. That's using data intelligently.
Customer survey data. Net promoter scores. Login rates. Product usage. Renewal patterns. Conference participation. Advisory board engagement.
You have mountains of customer data that should be feeding a completely different scoring model than what you use for prospects. But it's not. Because you haven't built it.
Here's the uncomfortable truth about why sales ignores your MQLs.
When sales closes a contract, there's a black-and-white number. Real revenue. Real money.
When marketing says "here's a qualified lead," it's all fuzzy. Where did you use objective data? Where did you guess? Sales followed up on your "qualified" leads before, and they weren't qualified. Nobody wanted to talk to them.
So now sales doesn't trust you. They built their own qualification process. Your lead scoring has become theater—activity that looks like work but produces nothing.
The only way to fix this is to sit down with sales and build the model together. Use actual closed-won data. Let predictive modeling do its job. Stop guessing.
Every major marketing automation platform now has AI and predictive modeling built in. You can literally turn it on and let the system analyze your data to predict which leads will convert.
Are you using it? Probably not.
At minimum, run a test. Turn on the AI scoring alongside your existing model. Track both through your pipeline. See which one actually predicts revenue.
You might discover that your carefully crafted manual scoring model—the one you spent weeks building—performs worse than the algorithm you could've activated with a single click.
Twenty-five years Debbie and I have been doing this. We used to wear lab coats at marketing conferences trying to get people to think more scientifically about their data.
We get that being a marketer today is hard. You're overloaded with tasks and execution imperatives. There's no time for analysis.
But at some point, you have to stop the crazy.
If you're not using lead scoring at all, here's your action plan:
Your lead scoring model is lying to you. Or worse, you don't have one at all.
Either way, it's time to fix it.
This post is based on a recent episode of Revenue Marketing Raw with Jeff Pedowitz and Dr. Debbie Qaqish. Watch the full episode at pedowitzgroup.com/revenue-marketing-raw.