VP of Paid Media, elk Marketing
2026 Meta Agency AwardOur team’s Compartés holiday campaign ↓
For years, the way my team runs accounts lived in a 200-page guide and in my reviews of their work. In 2026 I built it into AdNexus, the AI system elk’s strategists now use on live client accounts. I built the department from one person, and today it’s six strategists and about $100M a year in client media spend.
~$100M
Client media spend a year
Managed by the department I lead
6
Strategists report to me
Hired, trained and reviewed against one standard
860
AdNexus pull requests merged
July 15 to October 12, 2026
01 / The work
Compartés was my client. Its 2025 holiday campaign won the 2026 Meta Agency Award for Performance Campaign of the Year by a Boutique Agency in North America. It is also the best example of how I run an account: agree the target first, spend in steps, and judge the result against the plan.
Team campaignHoliday 2025CompartésI helped lead the elk team on the campaign. I built the $200,000 five-stage Meta plan, set the weekly budget steps, decided what creative we needed and wrote the award submission. Creators, designers and strategists made the ads and ran the account day to day.
We set a Meta return floor of 4x and split $200,000 into five stages that followed the prior year’s demand. Budgets rose 10 to 20% a week, and only after the last step held.
Over the nine days around Black Friday, Meta spend nearly doubled, purchases more than doubled and cost per purchase fell 10%.
Full-season return slipped from about 5.1x to 4.6x while store sales grew 33% and first-time orders grew 42%. Growing the business inside the floor was the goal, not the highest possible ratio.
02 / The team
When I joined elk in 2017, I was its only paid-media person. The team I’ve built since works from one written standard, and I review the work against it every week.
pages
Audit, optimization & training guide
I’m VP of Paid Media, and six strategists report to me. I hire through panel interviews, test projects and scorecards, and brought on two strategists this year.
The standard is written down: the audit and optimization guide, the Ad Training Library of playbooks and exercises, a launch checklist with screenshot proof and same-day checks, and a report format that puts the answer first. Biweekly 1:1s, monthly account audits and a weekly paid-media roundtable are where it gets taught.
When the team started using AI, the rules stayed the same: plan the work, get a person’s approval, check the result against real account numbers. Writing that down was the easy part. Getting six busy people to follow it on every account, every day, is why I built AdNexus.
03 / AdNexus
The standard,
built into software.
AdNexus is elk’s AI operating system for paid media. It connects Google Ads, Meta, TikTok, Microsoft Advertising, GA4, Triple Whale, Slack, Asana and meeting notes, and runs every account through the checks my team learns by hand. A strategist asks a question in plain English and gets a sourced answer, a dashboard or a change ready for approval.
I wrote the requirements from years of reviewing the team’s work, then built it myself with Claude Code and Codex on Next.js, TypeScript and Postgres. Between July 15 and October 12, 2026, 860 of my pull requests were merged through CI on a protected main branch.
It is built for the whole agency, with dashboards for strategists, account executives, project managers, creative teams and leadership. AdNexus reads the client’s goals before it recommends anything, and nothing changes in a client account until a strategist approves it. A full audit and optimization of an unfamiliar account now takes about 30 minutes.
In production at elk on live client accounts.
Find the waste first.
Eighteen monitors run on a schedule, covering search terms, ad copy, landing pages, stock, seasonal messaging and budget pacing. They stay quiet while an account is healthy. When something breaks, the strategist gets the finding, the evidence and a fix ready to approve.
Decide what to test next.
See which ads are working and why, then turn it into the next round: a plain explanation for the client and the next concept for the designer.
Keep the history with the numbers.
Platform results sit next to the Slack threads, Asana tasks, meeting notes and playbook guidance for the same client. When results move, the What Changed investigator compares the two periods and says where the change is and what to check next.
Meta, in depth
On Meta, one return figure hides the most. An awareness ad, a lead ad and a sales ad should not be measured the same way, and Meta’s own attribution rarely agrees with the store’s. The goal is simple: know which ads are working, why, and what the team should do about it this week.
Working screens, shown with fictional account data.
04 / Career
Nine years at elk, after in-house ecommerce and agency consulting.
VP of Paid Media · Built the paid-media department from one person
Digital Marketing Consultant · Contract work, alongside elk from 2017
Marketing Manager
Education Webster University · Marketing and Digital Art
05 / Also built
The same habit outside AdNexus: notice a chore I keep repeating, build the tool, keep fixing it.
A searchable record of my projects, decisions and sources. New notes wait for a check before an assistant can rely on them.
In daily useKeeps each AI project’s chats, tasks and changes in one place, with a second model to review the work before I keep it.
PrototypeCapture, mark up, paste. My editor and renderer work cut preview render time 61% on a 4K image.
Public, v0.3.6Turns an ad planner into the file the launch tool expects, without retyping or rewriting approved copy.
Try itNext /
I’m looking for a remote VP or Head of Paid Media, Growth or Marketing Operations role: leading the team, and building the systems it works in.