I just found this old video from 2015 when reviewing my old youtube account. It's about CPQ analytics.
We had no fancy AI back then. No auto-generated dashboards, no embedded data models in your CRM or CPQ UI, no slick predictive prompts. Just raw curiosity, SQL, and a desire to understand what was really going on beneath the surface of sales. That’s what drove the original presentation I gave on CPQ analytics ten years ago. And after rewatching it this week, I’m left with one surprising feeling:
It’s still ahead of its time.
Let’s unpack why.
Back in that 2015–2016-era presentation, I made the argument that CPQ data is structurally better than CRM data or survey-based insights. Why? Because it reflects actual buyer behavior, not opinions or filtered reporting. Every configuration is complete—even if a customer only answers one question, the rules fill in the rest. It's an automatic goldmine.
And yet, a decade later, most companies are still just scratching the surface:
They track hit rate—but not why they’re winning or losing.
They see top-selling configurations—but don’t analyze passive interest.
They look at what’s sold—but not what was offered and lost.
This is a problem of mindset, not technology.
The part I still stand by most strongly is the use of win/loss clustering. That slide with red dots (lost), blue dots (won), and network lines connecting them? It’s more than a pretty picture—it’s a map of opportunity.
“This is where you’re leaving money on the table. This is where you need to innovate. This is where sales tactics might swing the deal.”
That simple three-part segmentation—undervalued wins, competitive battlegrounds, unmet needs—still isn’t being adopted widely, even though most Tacton CPQ users have the data structure to support it.
In that talk, we demoed how to extract from Tacton’s TCsite using a plugin and an ETL pipeline. Today, we could do the same thing faster and cleaner—especially with Tacton’s evolving API landscape—but the key idea hasn’t changed:
The barrier is not access to data. It’s knowing what questions to ask.
One of my core messages was: don’t start by analyzing—start by asking. What do you want to understand? Which feature should we kill? Which product line needs a redesign? Is our pricing inconsistent across configurations?
If you can’t answer that, no dashboard will save you.
Even then, a big part of the Q&A turned into a discussion about pricing targets and effective pricing per configuration. The issue? Different teams place profit margins differently—some in the chassis, some in the gearbox, some split across features. The result is that the perceived value per configuration varies wildly and often makes no sense to the customer—or even to the sales team.
Today, some of our customers have taken steps to solve this using CPQ-integrated pricing analytics. But many are still pricing in silos. Ten years later, the question remains:
Are we managing prices—or just reacting to market pressure?
We closed that presentation by reminding everyone that analytics are only useful when tied to real action. KPIs should be few and purposeful (remember the car dashboard metaphor: speed, fuel, RPM), and every chart should lead to a decision.
That’s what we now formalize in our CPQ Analysis Workshop at cpq.se. We help manufacturing companies go from data extraction to actionable insight. Often in just five online sessions.
What we realized then—and what we’ve proven repeatedly since—is that most CPQ users don’t need more tools. They need better questions, cleaner models, and a way to prioritize.
Watching this 10-year-old video today feels a bit like watching a startup pitch before the market was ready. The vision was clear, the tools were there, but the broader CPQ community hadn't caught on yet.
Well, now it has.
AI is here.
BI tools are smarter.
Your CPQ setup is probably more mature than it was then.
So if you’re still wondering where the next step is, here’s your answer: start using the data you already have - and let it tell you where to go next.