DATA ANALYTICS · DEMAND INTELLIGENCE · MEASUREMENT
Conclusions follow the findings.
I decide from evidence, and I build the systems that let a team do the same. That's the throughline of how I work: start from the decision the data needs to change, reconcile the numbers before trusting them, and follow the findings wherever they go. Right now I'm putting it to work — leading the initiative that turns a company which rarely decides from data into one that does.
Client anonymized under NDA. Figures are shown as indexed shares and conversion rates from a representative discovery pull; the method and reasoning are shown in full.
ROLE
Initiative lead
CLIENT
Global B2B technology company
OUTPUT
Funnel model · Diagnosis · Dashboard
HOW I THINK
How I make a data-informed call
This is the method before it's any one project. It leans as much on editorial judgment as on analytics: the discipline is knowing what the numbers are allowed to tell you, and when they've actually earned a decision.
01
Start from the decision, not the dashboard
A metric earns its place only when it changes a choice. I begin with the call that needs making, then find the data that moves it.
02
Reconcile before you trust
Align definitions and denominators before reading a single rate. Half the disagreements in a funnel are vocabulary, not performance.
03
Let the questions lead
Name what the data can't yet explain, and resist the tidy conclusion. The open question is often the real finding.
04
Connect it to the story it changes
Analysis matters when it points to a different way of working. I close the loop from number to narrative to decision.
The thinking in action
One initiative I'm leading, to make data the default input to decisions.
What follows is that method applied: a demand-intelligence discovery for a global B2B technology company, walked through as it happened.
IN PRACTICE — START FROM THE DECISION
Numbers, but no decisions made from them
The deeper problem was never a lack of numbers. It was that almost no decision actually started with them. Reporting existed, but it lived in disconnected systems that didn't agree: two sources described the funnel with different stages, so a single conversion rate read 53% in one source and 31% in the other. Asset-level content data existed only in fragments, and nothing joined content to pipeline at all.
The temptation in that situation is to jump to a fix. As lead, I set the opposite rule: discovery first, and agree the real gaps before anyone builds anything — so that data, not instinct, becomes the default input to the decision.
IN PRACTICE — READ IT HONESTLY
A structured model, leaking at the ends
The first pull showed a genuine five-stage funnel, gated well past an email capture. Sales accepts roughly two of every three MQLs, so the qualification handoff holds. The losses sit at the top and the bottom.
INQUIRIES
100%
↓ 39%
AQL
39%
↓ 65%
MQL
25%
↓ 53%
SQL
13%
↓ 14%
WIN
2%
Stage volumes indexed to inquiries = 100%. Step-to-step conversion shown between stages.
64%
Sales accepts about two in three MQLs. The handoff is not the leak.
39/14
61% of inquiries never reach AQL; 86% of SQLs never close. The middle holds; the ends leak.
5
A gated, multi-stage model — qualification sits well past an email capture, not at it.
IN PRACTICE — LET THE QUESTIONS LEAD
Where a good MQL and a real win diverge
Splitting conversion by channel exposed the real story. Some channels manufacture MQLs that never become revenue. The lever is channel-weighted scoring and the owned content layer — the MQL definition alone can't explain this spread.
CHANNEL
INQ → MQL
SQL → WIN
Content Syndication
99%
1%
Online Advertising
58%
1%
Event
39%
4%
Website
19%
1%
Email
5%
2%
The questions the data raised — before any answer
Why does syndication convert 99% to MQL but ~1% to win?
Likely a model weighting volume over fit. Denominators needed before reading too far.
If email → MQL is just 5%, what defines an MQL today?
The data shows the distribution, not the trigger. One to confirm with governance.
Why do website leads convert best mid-funnel?
Lowest volume, highest quality. Worth understanding, and doing more of, if it holds.
IN PRACTICE — RECONCILE BEFORE YOU TRUST
Four things discovery had to confirm
A conversion rate is only as honest as its denominator and its definition. Before drawing a single conclusion, I documented what still had to be nailed down.
1
Reconcile the taxonomy
Two sources use different stages, so MQL→SQL reads 53% versus 31%. Align before trusting any cross-source number.
2
Get the denominators
A 99% conversion on a small base is not a 99% on a large one. Weight every rate by absolute volume.
3
Document the MQL trigger
We had the distribution of MQLs, not the definition. Confirm what action actually flags one today.
4
Close the content gap
No layer joined content to pipeline. How do we measure top-of-funnel content and tie it to the journey?
IN PRACTICE — CONNECT IT TO THE STORY
From digital brochure to demand engine
The website was the one owned surface where content and conversion could connect, and it was being run as a brochure. Same surface, different definition — the analysis pointed to a shift in what the site is for.
DIGITAL BROCHURE · TODAY
DEMAND ENGINE · THE SHIFT
BUILT TOInform
BUILT TOConvert
ORGANIZED BYThe org chart — products, solutions
ORGANIZED BYThe buyer's journey — intent and stage
MEASURED BYWhether it's current and on-brand
MEASURED BYPipeline influenced
THE RESULTNothing compounds
THE RESULTAn owned audience compounds
THE EVIDENCE — THE SITE, IN THE NUMBERS
The fragmentation wasn't a hypothesis
A customer-journey-analytics pull confirmed it. Traffic piles onto the home page, the best theme content sits behind gated forms, and the channels that will decide the next few years barely register.
PAGEVIEW CONCENTRATION
Home22%
Events & webinars3.2%
Top theme pages (gated)~2.5%
The home page draws about 6.9× the next page. Demand for the themes is real, but it's gated and scattered.
SESSIONS BY CHANNEL
Direct32%
Referring23%
Paid Search18%
Organic Search16%
Organic / Paid Social11%
AI Engines0.04%
Paid search outruns organic, and answer engines are a rounding error — the exact surface buyers are moving to.
THE MARKET SIGNAL — THIRD-PARTY RESEARCH
94%FORRESTER, 2026
of B2B buyers used AI in their last purchase. Answer engines now outrank websites and reps as the top research source.
88%MOZ, 2026
of AI answer citations don't appear in the organic top 10. Being the clearest, best-structured source beats ranking.
69%G2, 2026
chose a different vendor than planned after AI chatbot guidance. A known leader can be excluded, not just an unknown lifted.
43%HUBSPOT
average organic-traffic lift from a topic-cluster model — authority that compounds over 6–12 months.
WHERE IT'S HEADED
Four objectives, in order
01 · CONNECT THE DATA
One view across content, demand gen, field, and social — the layer the work had been missing.
02 · FIND THE GAPS
Where measurement and definitions break down across the funnel, documented rather than assumed.
03 · LIFT LEAD-TO-MQL
Raise the quality of demand, not just the count of leads, by leaning on the channels that actually convert.
04 · BUILD THE OPTIMIZER
Use the connected view to steer spend and effort toward what converts to pipeline.
Foundation, then diagnosis, then conversion, then automation. Each objective earns the next — and none of them get to skip the evidence.