I make marketing measurable, repeatable, and tied to revenue.

Pipeline intelligence. Built, not reported.

10+ years · B2B + B2C · Revenue Intelligence · ABM · Lifecycle

Pipeline Funnel · Live demo

Awareness

4.2k

MQL

1.8k

SQL

640

Closed won

180

Expansion

Pipeline Velocity

23d

avg MQL → Close

↓ 4 days vs prior

MQL → SQL

SQL → Close

15d

Details · Awareness

4.2k

Reach

+180

43%

Aware→MQL

+2%

Organic vs paid

reach

Net Revenue Retention

118%

↑ 6% vs prior · install base

First touch now starts in search and AI answers: buyers arrive pre-researched. The 43% that converts to MQL had shortlisted us before the first click. Awareness is earned upstream of the ad.

WORK

APAC Regional Market Intelligence

A uniform campaign strategy across three APAC markets was suppressing regional conversion potential. Rebuilt campaign architecture around regional psychology, delivering 20% conversion lift and 50% revenue growth.

The Measurement Stack

Ask three models what a campaign returned and you get three confident, different answers. Single-model attribution did not survive signal loss; the 2026 answer is running the models together and reconciling them into one defensible number.

How I would measure in 2026

One campaign. Three honest numbers.

The same $120k integrated campaign spanning paid, email, and events, read by the three models serious teams now run together. Each model takes the stage, answers, and hands over.

01 / 03

Model 01

Multi-Touch Attribution

What did this campaign return?

0.0x return

Built for: tactical credit. It shows which touches, sequences, and content moved deals at campaign speed.

Cannot see: anything untracked. In 2026 that is a lot: dark social, AI-assisted search that never clicks, and agent traffic polluting the click trail. The deeper flaw: it reads presence as influence. A touch is not a cause.

MTA handed retargeting the biggest share. It reaches people already on their way to buy.

Model 02

Marketing Mix Modeling

Same campaign. Same spend.

0.0x return

Built for: budget allocation. It reads channel contribution across offline, brand, and seasonality with no user tracking needed. Open-source frameworks like Robyn, Meridian, and PyMC-Marketing made it affordable; modern causal versions refresh weekly and calibrate against lift tests.

Cannot see: short flights. Integrated stacks now push MMM toward campaign grain, but the math still needs long, steady history; a six-week flight never produces the sample. Strategy in focus, tactics out of frame.

MMM cut paid search's credit nearly in half. Branded clicks harvest demand created upstream, increasingly inside AI answers.

Model 03

Incrementality

And what did it actually cause?

+0.0x incremental

Built for: causal proof. Geo and audience holdouts separate what marketing caused from what would have happened anyway, with synthetic controls now building the counterfactual on smaller budgets. Uplift modeling extends the same logic to targeting: who is worth spending on at all.

Cannot see: everything, affordably. One test buys one answer: one channel, one period. It costs reach, discipline, and enough volume to read, so you run them where the money argues loudest.

Under a third of the tracked return was truly incremental. The rest was demand wearing marketing's badge.

The verdict

0.0x · 0.0x · +0.0x. Same campaign.

None of the three is wrong. Each answers a different question. So the job is reconciliation. Multi-touch runs weekly for tactical calls, mix modeling stays current to set the budget envelopes, and holdouts run always-on where spend concentrates. Once a month it all reconciles into one number the room can defend, on first-party data with consent handled up front. On this campaign the call was to strip the spend that was harvesting demand, keep the program that remained, and retest next quarter to confirm the trimmed version clears breakeven.

Illustrative campaign, illustrative numbers. The method is the point.

The Post-Sale Engine

Every team runs a pipeline to the sale. Almost nobody runs the one after it, where renewal and expansion are actually decided. A framework for post-sale as a second pipeline: handoff, adoption, renewal, and expansion run as one motion.

Frameworks · How I would run the stage after the sale

The Post-Sale Engine. The stage most plans skip.

Every team runs a pipeline to the sale. Almost nobody runs the one after it, where renewal and expansion are actually decided. I treat post-sale as a second pipeline with four stages. Tap each one.

01Handoff

02Adoption

03Renewal

04Expansion

Handoff: the sale closes, the story continues

Most marketing goes quiet at the signature. The handoff moves everything the pre-sale motion learned into the post-sale one: what the customer bought, the problem they hired it for, and what success was supposed to look like.

Onboarding communication continues the story they were told before they bought, so week one feels like a continuation instead of a cold start with a new department. Effort scores at each onboarding step, first login, data migration, first real use, catch friction where it actually builds, and CSAT after the first support interactions checks whether the promise is holding.

Adoption: the health score watches everything

Accounts show you how they are doing before they say it. The health score blends behavioral signals like usage patterns, feature adoption, and champion activity with relationship signals like NPS, CSAT after key touchpoints, support tone, and executive engagement, weighted by what actually predicts renewal.

On top of the score sits a churn model that now reads the unstructured signals too, with an LLM layer pulling the competitor mention out of a call transcript or the budget worry out of a support thread, and survival modeling estimating not just whether an account is at risk but when. The model is not the hard part. Without a playbook behind it, flagged accounts still churn, so dips trigger plays: enablement content for the unadopted capability, champion nurture when engagement thins, and a flag to the account team when the signals stack up. The loop from signal to outreach closes in hours, not weeks.

Renewal: the window opens earlier than the invoice

A renewal is decided months before it is signed. The engine opens the window a quarter before the date, builds the value recap from outcomes captured along the way, and treats risk signals like a champion leaving, usage falling, or a detractor NPS response as escalations, not observations, with the save effort sized to the account's predicted LTV.

The executive business review is where the survey data meets the room: it is the moment to confirm the customer can state the ROI in their own words, because the ones who cannot are the ones who churn.

This stage is orchestration, not solo marketing: customer success owns the relationship, sales owns the commercial motion, and marketing arms them both with the evidence and the sequences.

Expansion: earned, not pitched

Expansion is a permission you accumulate. Healthy adoption plus captured outcomes tell you which accounts have earned the conversation, and usage patterns point to what the next natural step is.

The motion then widens the buying committee rather than re-selling the champion, because the second deal is bought by more people than the first one was.

Acquisition has owners, dashboards, and budget. The revenue after the sale usually has nobody. The second pipeline gives it the same discipline the first one gets: stages, signals, and plays, run across marketing, customer success, and sales as one motion. It is where net revenue retention is actually won.

Tools

A 90-day plan you can walk through, and a quarter of budget you get to run. Tap a tab to switch.

Thinking

The interface shift

For a decade the job was to be found, then to be cited. The next one is to be callable.

When a buyer's AI agent shortlists vendors, checks pricing, or starts a purchase, it doesn't read your landing page. It hits an interface. MCP, the standard that lets agents connect to external tools, has already won that layer, adopted across every major AI provider and now housed at the Linux Foundation. So the question shifts from "is my content citable" to "can an agent transact with my product without a human in the loop." This is still early and uneven, so the honest read is directional, not done. Forrester now predicts that in 2026 one in five B2B sellers will face AI-powered buyer agents negotiating quotes. The companies structuring their data, pricing, and offers for machine consumption now will own agent-mediated demand the way the SEO winners owned search. Marketing's next job is making the company legible to machines, not just persuasive to people.

Beyond AEO

Optimizing for AI answers is no longer the edge. The edge is owning what the AI says about you.

SEO to AEO is settled now, mainstream enough that agencies sell it and platforms measure it. The frontier moved. Your brand is no longer what you publish, it's what an answer engine says when a buyer asks, and the majority of B2B buyers now start their research inside AI rather than Google. That makes citation share a measurable discipline, and it makes accuracy a brand-safety problem: the model will describe you whether or not you've shaped what it knows. SEO tactics stop at what you publish; this discipline starts after. The ones winning are managing what the model believes about them with the same rigor they once gave their Google rankings.

The timing lever

Everything about the spend gets optimized, and nothing about the clock.

Deals are lost less often to the wrong channel than to the wrong moment, and the moment is invisible if you only measure win rates. Connect campaign first-touch data to deal outcome data and the pattern is almost always the same: the channel was fine. Marketing arrived after the buying decision had already started. Earlier content, not more content, is the lever still sitting there.

Scroll to read each position

1 / 3

CONTACT

Get in touch

I'm always open to
a good conversation.

I read everything. Roles, collaborations, and marketing questions are always welcome.

Connect on LinkedIn → See my dashboards →