You scanned the card

Here's the work.

Fractional CMO for the AI era. The role used to stop at strategy — brief the agency, approve the campaign, read the report back. It does not anymore. The same person now builds the websites, ships the campaigns, and stands up the data pipelines and MCP servers that make the system measurable. Strategy includes execution. Execution includes building.

Everything on this page is proof of that. Three brands rebuilt cold, before any conversation. And the AI reporting stack that gets deployed underneath the campaigns I run.

Joe Kim. Fifteen years in DTC and ecommerce, now consulting independently.

How the work happens

Show, don't tell.

Most fractional CMOs pitch with a deck of advice and a roadmap that lives on someone else's calendar. The work is different. Four moves, none of them a slide.

Built cold

Three companies. Three gaps. Three things I built.

I picked these brands, named the gap, and built the fix — rebuilt product pages, quizzes, diagnostic tools, live and clickable. This is what the campaign side of the engagement looks like in practice. The work you click through below is the work that ships when we work together, just done without an invitation.

Open any one.

Running in production

The reporting stack that ships underneath the campaigns.

The other half of the engagement. Strategy and campaigns are visible above. The system that measures them is below — deployed inside the org, not just talked about.

Running today across ~40 client accounts for an agency partner. Each month it reads a master Google Sheet, plans every data pull, executes them against AgencyAnalytics, GA4, and Search Console, then a deterministic script does the math and renders each branded report — filed where the team reviews, adds judgment, and delivers. A team-week becomes a scheduled batch on the 1st.

Inside one brand it works the same way, scoped to one company instead of forty — built and operated by the same person making the strategic calls.

AI has escaped the chatbox. Marketing leadership now includes building the systems underneath it.

ConfigMaster Google Sheetdrives the batchPhase 1 · scriptPlanlists every call,pins each to a filenamePhase 2 · LLMExecutecalls MCP, writes rawresponses verbatimraw filesPhase 3 · scriptsAssemble + rendermath · cross-checks→ branded DOCXSources · via MCP connectorsAgencyAnalytics — paid + socialGA4 · Search Console · Meta Adslive pullsOutputBranded DOCXsyncHosts in cloudGoogle DriveReview & deliverAccount managerMissing or mismatched data renders an explicit fallback or halts the run — numbers are never fabricated.

The hard part

Multi-source orchestration.

AgencyAnalytics, GA4, and Search Console each feed in through MCP connectors. A script plans every call and pins each response to a file; the model only runs the pulls; a second script does all the math and cross-checks before it renders. The LLM never holds the numbers, so it can't fabricate them. Scheduled batch, not a microservice.

The real unlock

"What specific product or value proposition is generating the most revenue across all channels?"

The deck is the visible artifact. The asset is every data source — GA4, Search Console, and every paid channel — wired into one workspace an LLM can read. Questions that used to mean a half-day of pulling exports are now a sentence. The monthly deck is one output of that capability.

Your turn

Curious what I'd find in yours?

Every brand above got this cold. No brief, no invitation. If you want to know what I would find in your funnel, what your customers see that you cannot, what the campaigns should be, and what the reporting stack underneath them would look like — that is the call. Thirty minutes. I will come with a point of view, not a questionnaire.

Or write me directly: hello@joekim.info