How I work, in four phases.

The shape of an engagement

I'm a marketing analyst first. That word matters: I open with measurement, not with a creative pitch. Most marketing problems aren't creative problems — they're a missing number, a leaky stage, or a channel being asked to do work it can't do. The job is to find that number first, then build around it.

A typical engagement is three days a week for ten weeks. Phase 1 takes a fortnight. Phases 2 and 3 overlap in the middle. Phase 4 is the last fortnight and the bit that survives me leaving.

Map the funnel before anything else.

Most teams know they have a problem somewhere. Few have it written down stage by stage. Phase 1 is to draw it.

I sit with whoever owns marketing and we list every step from impression to revenue — paid impression, organic impression, click, landing-page view, lead capture, qualification, sales call, close. Then we put a number next to each one. Where we have data we use it. Where we don't, we mark the gap and come back to it in Phase 4.

The output is a one-page funnel map with the largest drop-off circled in red. That's the only thing Phase 2 is allowed to work on.

• Inputs: CRM access, GA4 access, ad-account access, two hours with the founder or marketing lead.
• Output: a one-page funnel map with absolute numbers, conversion rates, and the one stage that matters most.
• Time: 10–14 days.

Score every fix on impact and effort.

With the bottleneck identified, Phase 2 is a structured audit of everything that touches it: positioning, messaging, channel mix, creative, on-page SEO, technical SEO, lead-capture mechanics. Twenty to forty findings, each scored on a 1–5 impact and 1–5 effort scale.

The deliverable is a Notion or Google Doc the team can read in 20 minutes. Each finding has: the screenshot, the recommendation, the expected lift, and the rough hours. We sort high-impact / low-effort to the top and start there.

This is also the phase where I push back on assumptions the team has stopped questioning. "We've always done LinkedIn this way." "Ebooks don't work in our market." Those sentences are usually wrong, but you can't argue with them — you can only test them. I write the test plan into the audit.

• Output: ranked audit document, 20–40 findings, top 5 selected for Phase 3 production.
• Time: 2–3 weeks, overlapping the start of Phase 3.

Build the production line, not just the campaign.

Phase 3 is the bit most marketers want to start with. It's also the bit that wastes the most money when started without Phases 1 and 2.

Depending on the audit findings, this is where I build:

• Ebook funnels — long-form lead magnet, landing page, capture, automated nurture sequence. Worked at Excelerate. Lower CPL than paid every time.
• Quiz funnels — interactive scoring + segmented follow-up. Higher engagement than ebooks, lower commitment than a sales call. Where Make.com earns its keep.
• AI-generated promo videos — HeyGen avatars + ElevenLabs voice + a tight script template. We hit +70% view-through at Excelerate after switching to this format.
• LinkedIn carousel + post engine — repurpose every long asset into 6–8 short ones. The 60K weekly impressions number was built on this loop.

I document the whole stack — tool by tool, prompt by prompt, with screenshots and run-throughs — so the team can keep producing after I leave. A funnel that only I can run isn't a funnel, it's a dependency.

• Output: 1–3 production lines live, with the SOP doc and Loom walk-throughs.
• Time: 4–6 weeks, overlapping Phases 2 and 4.

See the full tooling stack →

Hand back a dashboard the team actually opens.

Most marketing dashboards die from either too many metrics or too few. Phase 4 is to find the seven numbers the team will look at every Monday — and only those.

I stand up GA4, Search Console, and either Looker Studio or a Notion-embedded equivalent depending on the team's habit. The dashboard tracks: top-of-funnel impressions, captured leads, qualified leads, opportunities, closed-won revenue, and the two stage-conversion rates that matter most. Plus one custom metric specific to the bottleneck identified in Phase 1.

Then I run two weekly meetings with the team — same agenda each time — until they can run it without me. That's the test. If after four weeks the meeting still needs me, the dashboard is wrong, not the team.

• Output: live GA4 + dashboard, weekly review template, two-week handover.
• Time: 2 weeks.

What I won't do

Worth saying out loud, because it saves both of us a wasted call.

• Brand identity from scratch. I work with what exists or partner with a designer. I have opinions on logos but they shouldn't be paid for.
• Pure paid-media management. I'll diagnose paid and write the briefs, but day-to-day Google Ads / Meta Ads bidding is a specialist craft and I'd rather hire one.
• "Vanity metrics" reporting. If a dashboard exists to make the team look good rather than to make decisions, I'm not the right person to build it.
• Black-hat SEO of any kind. Including link-buying and AI-generated content farms. The E-E-A-T explained piece is why.