Five Case Studies, Five Gaps Closed

I find what’s missing, then build the system to close the gap.

A magazine publisher redesigning its value proposition? An AI startup scaling faster than its systems could keep up? A retailer surging back into cultural relevance, needing a briefing framework fast enough to keep pace? I’ve transformed systems for all three. Nobody handed me a mandate, a precedent, or a plan. I found what was missing and built it.

Download résumé LinkedIn

Previous work Condé Nast · Gap Inc. · Mobian

Solutions used by L’Oréal · Estée Lauder Companies · LVMH · Hermès · P&G · PepsiCo · Anthropologie

How I work
I find the real problem It’s usually not the one written on the brief. I start by figuring out what stakeholders actually need.
I build the smallest thing that works A one-pager people use beats a system nobody opens. I keep it focused, and everything on it has to earn its place.
I make it run without me If it breaks the day I leave, I didn’t finish. I write it down, hand it over, automate what I can.
I don’t wait for the “right time” No team, no budget, no precedent, no time - that’s just the starting point. The work still has to happen.
What’s inside
01
No mandate
Marketing guide  ·  Gap Inc.
Gap’s cultural relevance was surging and campaigns were going viral. I built a one-page guide that gave the brand team a fast, shared read on what each season’s messaging had to do for the customer. They started pulling it every season on their own, and it spread to other markets and its sister brands (Old Navy, Banana Republic, Athleta).
02
No precedent
A number that didn’t exist  ·  Condé Nast
No one had an agreed way to value the press and buzz the company’s brands generated. I forecasted that publishers would need it to hold their own against influencer marketing, then built the method, picked the vendor, and got executives behind a number they’d never used. The measure reached the rooms where ad budgets get set.
03
No team
A 60-hour job, automated  ·  Mobian
Reviewing “Where My Ads Ran” reports by hand to flag off-brand creators took up to 60 hours a campaign, with no engineer to help. I built the first version of the pipeline myself, then scaled it with one. Sixty hours became three minutes, and one Fortune 100 client brought 5× more business to us in a quarter.
04
No time
A five-day report  ·  Mobian
The same report ate about a week of hand-work every cycle. I mapped the whole process and, working with Convey, turned it into an agent. Now it’s about fifteen minutes, delivered to inboxes.
05
No plan
2,000 survey answers  ·  Condé Nast × Anthropologie
A big custom survey landed with no content plan attached. I read the data for the through-line, picked the angle, and shaped a Vogue series that went on to outperform.
01 · No Mandate

A one-pager the marketing team started using on their own

What was in the way Gap’s cultural relevance was surging and campaigns were going viral. Brand teams had to move just as fast and stay centered on the customer, and there was no shared system to support them.

Cross-team trust Built the brief Made it stick
01 / 05
The situation

Every season, the brand team planned campaigns off a stack of reports: social listening, search, audience research, market trends. Each one described what customers were doing. None of them gave the team a direction to build the season’s campaign around.

What the marketers and their agencies needed was one page they could brief from. Not more data. A point of view.

They had a dashboard for every angle and no answer. I gave them one page.
What I did

I read the signals and identified four real customer needs that kept showing up, then wrote them as plain statements anyone could brief from. I called them Need States.

A stack of signals Social chatter, search, audience affinities, market trends
Read together What keeps coming up across all of them?
Four needs the work has to hit Written plainly enough to brief a creative or media team from

Here’s an example: four need states

Help me dream Put the product inside the version of the season they’re picturing - the trip, the debut.
Support my identity Back something they already care about: their sport, their music, their taste.
Keep me connected Show the product with other people in it, not on a lone model.
Surrender to the moment Reward being present. Surprise and “you had to be there” do the work.

And the one-pager itself - this example is built on illustrative signals, not the real season’s data.

An example Need States one-pager
The Need States one-pager.
Illustrative An example of the kind of guide the framework produces. Wording adapted from one season’s output, not a live brief.
What came of it

Adopted without anyone being told to.

Every seasonMarketers pulled it themselves, for their own teams and agencies
More markets and sister brandsRolled out to Gap’s international teams; the method also picked up across its sister brands (Old Navy, Banana Republic, Athleta)

The one-pager became integral to how the CMO and marketing leadership talked about the season - one read built from scattered, disparate inputs.

02 · No Precedent

There was no way to value publisher buzz vs. influencer, so I built one

What was in the way The measurement didn’t exist inside the company. I had to design it, decide who we’d buy data from, and get executives to trust a number they’d never seen before.

Vendor pick Wrote the method Sold it up
02 / 05
The situation

The company’s brands and events generated an enormous amount of press and social conversation, and no consistent way to say what it was worth. Sales wanted to put it in front of advertisers. Editors wanted it for event recaps. Everyone wanted a number they could defend.

I forecasted this before anyone was asking for it: publishers would have to hold their own against influencer marketing, and earned media value was how they’d do it.

How do you set a standard when nobody hands you one?
What I did

I ran a bake-off between three data vendors, tested their methods against our own brands and events, picked an approach, and wrote the rules for how we’d apply it, so the answer came out the same way every time. Then I got the measure into the sales and executive materials where it would matter.

Forecasted the needPublishers would need this to hold their own against influencer marketing
Vendor bake-offDMR, Launchmetrics, CreatorIQ - tested against our brands
Wrote our methodConsistent definitions and benchmarks
Made it repeatableSame report, same shape, every event
Into the decksSales pitches, RFP responses, executive materials

Where it ended up

Spring and fall sales pitches, our earned value against competitors’
Renewal talks, where sales could show a brand the earned value it got against what it spent
Quarterly business reviews with Estée Lauder Companies, L’Oréal Paris, Gucci, Hermès, Louis Vuitton
Joint-business plans, where earned value was the added-value case for a year-over-year spend increase
Recaps for big events like Vogue World and the Met Gala

How I framed the comparison

Illustrative Shape only, numbers invented - to show how the comparison was laid out for a sales conversation. Not real client or competitor figures.

How I built the report

Earned media value tracked over a reporting period
Earned media value tracked over time, from the repeatable report. Figures redacted.
What came of it

A number that shapes budget decisions.

$90M+In event and sponsorship revenue the measure helped shape
Kept goingPicked up by other markets after leadership backed it

The measure became part of the pitch for events like Vogue World and the Met Gala, and part of how the company answered advertisers.

03 · No Team

A 60-hour manual job, no engineer, so I built the pipeline myself

What was in the way Deciding which creators a Fortune 100 brand shouldn’t run ads against meant working through “Where My Ads Ran” reports from YouTube and Google by hand, up to 60 hours a campaign. There was no engineer to build a faster way.

Built the pipeline Scaled with an engineer 5× the client’s business
03 / 05
The situation

Before every campaign, someone worked through “Where My Ads Ran” reports from YouTube and Google by hand, flagging which creators were a bad fit for the brand: high-risk, wrong audience, low-attention filler. Names in, names out, no reasoning kept. Up to 60 hours.

One campaign had most of its budget running against kids’ cartoons and videos made to keep dogs calm. Technically safe. Nothing to do with the brand.

No engineer, no headcount. I built the pipeline, and it had to hold up on live Fortune 100 campaigns.

Safe, but not a fit

Budget in kids’ content~half
Budget in brand-aligned contentunder 1%
Illustrative A picture of what I was finding. Exact percentages, spend, and client identity withheld.
What I did

I built the pipeline that turned those “Where My Ads Ran” reports into exclusion lists. A separate engine classified each creator, by risk, attention, audience, and dozens of other categories, and I connected that output to the campaign config, so the exclusion list fed straight into how Fortune 100 campaigns ran on YouTube. Sixty hours of manual review became about three minutes. Once the pipeline proved out, I hardened it with an engineer, and the list scaled from roughly 50,000 creators to more than 600,000.

Turnaround
By hand~60 hrs
Pipeline~3 min
List size, after scaling
Start~50K
Scaled600K+

After that, steering a campaign away from the wrong creators and toward brand-aligned ones was a quick job, not a week’s.

What came of it

A prototype that became how campaigns got configured.

~60h → ~3mTime to build a campaign’s exclusion list
50K → 600K+Creators classified and screened, after scaling
5×More of one Fortune 100 client’s business onboarded, in a quarter

Built solo as a prototype, then made production-ready with one engineer once the prototype had buy-in.

04 · No Time

A five-day report, rebuilt to run in fifteen minutes

What was in the way Every cycle, the same report ate about a week of hand-work before anyone could act on it. That doesn’t scale, and it burned people out on the boring part.

Mapped the process Built with Convey Now it’s automatic
04 / 05
The situation

The go-to-market team kicked off a report by email, then waited five days while someone pulled the data, ran the analysis, and built the deck by hand.

The thinking wasn’t the hold-up. The assembly was.

The slow part was putting it together, not working it out.
What I did

Working with Convey, I turned the whole process into an agent. It takes the email, needs only the brand name and what that brand counts as wasted spend, runs the analysis, and builds a finished deck in Google Slides - sent straight back to whoever asked. About fifteen minutes.

Before
Request by email
Pull & analyze by hand
Build the deck by hand
~5 days
After
Same emailBrand name + what counts as wasted spend
The agent runs itAnalysis, then a deck in Google Slides
Deck lands in the inbox
~15 min
Before~5 days
After~15 min
What came of it

A weekly slog that now runs itself.

~5 daysHand-built, before
~15 minSame deck, agent-run, after
~480×Faster, give or take

Built with Convey. My part was turning a fuzzy manual routine into steps an agent could run, and deciding what “good” looked like at each one.

05 · No Plan

2,000 survey answers and a deadline - I found the story

What was in the way A big custom survey landed with no content plan attached. Editors needed to know what in it was worth writing about, and soon.

Read the data Picked the angle It outperformed
05 / 05
The situation

Anthropologie and Vogue surveyed more than 2,000 brides about how they shop for their weddings. Rich data, no direction.

Someone had to decide what it was actually about before it could become anything.

A pile of answers isn’t a story until someone decides what it’s about.
What I did

I read it for the thread that held it together and found one: the wedding isn’t a single day anymore. It’s a string of events - engagement shoot, parties, rehearsal, ceremony, after-party - and each one needs its own outfit. That’s what the coverage was built on.

Engagement shootCalled the first big moment by ~70%
Engagement party
BacheloretteYounger brides drove an uptick in gifting the crew
Rehearsal dinner
Ceremony to receptionWhere the two-dress finding lives
After-party
From the analysis The bachelorette gifting trend among younger brides is directional from the survey read; no figures shown.

The findings that carried it - all publicly reported

Planning 4+ events beyond the wedding day: engagement photos, rehearsal dinner, honeymoon, bachelorette~2 in 3
Call their wedding theme “romantic,” with “bohemian” rising35%+
Have two or more dresses for the ceremony and reception~half
Say the after-party calls for an outfit change~66%
Plan a bridal party of 5+, and of those, 22% larger than 1063%
PR Newswire, 13 Mar 2023 · Vogue.com, Anthropologie Weddings series
Vogue.com · sponsored Wedding Wardrobe 101: How to Dress for Everything, From the Engagement Party to the Goodbye Brunch Three-part series on Vogue.com through May 2023, built event by event PR Newswire Anthropologie Weddings Releases Results of First Comprehensive Bridal Survey, Produced by Vogue with Anthropologie 13 Mar 2023 · “more than 2,000 people” · findings released as a three-part sponsored series
Wedding Wardrobe 101 as published on Vogue.com
“Wedding Wardrobe 101” as published on Vogue.com.
What came of it

A pile of data that set the coverage.

3-partVogue.com series built on the survey read
Top performer“Wedding Wardrobe 101” among Vogue Weddings’ best articles of the period

Raw data in, a clear angle out, and a story the audience actually read. I ran the analysis and told editors and the client which parts mattered.