Sr. Account Ops Manager · Jun 2019 – Mar 2025

Causal IQ

I built 4 Python tools and a custom CTV measurement framework that became the company-wide standard. Made operations 60x faster. Adopted by every Account Manager and Data Analyst, with the CTV codebase handed off to Causal IQ at my departure.

Faster Reporting

From hours to seconds

Company-Wide

Standard Tooling

Adopted across the org

Profitability Rank

Out of 22 managers

$900K+

Monthly Ad Spend

Peaking at $1.2M/mo · 10+ DSPs

The Story

The only operator writing code

At Causal IQ, a programmatic advertising platform with ~125 employees, I was the only operator writing code to solve workflow bottlenecks. Account Managers spent hours each day on manual tasks: pulling reports, adjusting bids, building insight decks, and managing site lists across 8-10 DSPs (DV360, The Trade Desk, Xandr, Yahoo, Zeta, Basis, Quantcast, Vistar, YouTube/TrueView, and more).

I built four Python/Pandas tools plus a custom CTV measurement framework to automate that work. Leadership adopted them as the company-wide standard. They are the reason I ranked #1 or #2 in profitability out of 22 managers. I managed higher volumes of spend with zero errors because the tools did the repetitive work for me. The book of business included CDC, Hertz, Grammarly, Orangetheory, Thomas J. Henry, Primo Water, AAA Auto, PMI Colombia, Floor & Decor, Party City, and national political campaign teams.

Built with AI-assisted engineering - Claude Code, Codex, Cursor, and Copilot are daily-driver tools in my workflow. The CTV codebase was handed off to Causal IQ at my departure for continued use across the broader client book.

</>

AI-Assisted Engineering

The Causal IQ Python automation suite and CTV measurement framework were built with Claude Code, Codex, Cursor, and Copilot as daily drivers - the same workflow that produced my Stanford BIOE 230 platform in 24 hours.

Notable Clients

Who the tools actually served

The book of business at Causal IQ. Each name represents campaigns I personally optimized using the automation suite - across industries from public health to legal to political to CPG.

CDC

Public Health

Direct programmatic activation

Hertz

Travel · Mobility

Grammarly

SaaS

Orangetheory

Fitness

Primo Water

CPG · Beverage

AAA Auto · Direct Auto

Insurance

PMI Colombia

Regulated CPG

Philip Morris International

Thomas J. Henry

Legal · Anchor Client

CTV measurement framework built for this account

Party City

Retail

Floor & Decor

Home Improvement

Grew from ~$10K to $250K/mo (25x)

National Political Campaigns

Political

Multi-team programmatic activation

The Scale

Ten DSPs. One automation suite.

The tools automated workflows across 8-10 DSPs where $900K+/month in spend was managed, peaking at $1.2M/month during high-volume periods. DV360, The Trade Desk, Xandr, Yahoo, Zeta, Basis, Quantcast, Vistar, YouTube/TrueView, and others.

Before & After

The transformation, visualized

Click any bar to jump to the full tool breakdown below.

Bid Optimization

-98.9%

Before: 4 hours / day After: 5 minutes

4 hrs → 5 min

Audience Insights

-99%+

Before: 2 full days After: minutes

2 days → minutes

Site List Management

0 incidents

Before: error-prone manual process After: zero brand safety incidents

Errors → 0 incidents

Campaign Reporting

60x faster

Before: hours of manual compilation After: real-time dashboards

hours → 60x faster

hours saved per month across the organization.
If every Account Manager and Data Analyst saves 4 hours per day, that is 96 hours saved daily - or 2,016 hours per month redirected from spreadsheets to strategy.

Tool Showcase

5 tools. All with video walkthroughs.

Each tool has a real video walkthrough. Select a tool below to see the full breakdown and demo.

Tool 1 of 5

Automated Optimization Engine

Problem

Managing bid and budget adjustments across 100+ line items was a 4-hour manual process every day, prone to human error. A trader had to pull data from each DSP, compare it against KPIs, decide what to change, then manually enter adjustments one by one.

Solution

Script calls the TTD API to pull real-time performance data.
Pandas merges this with historical benchmark data.
The engine flags underperformers and outperformers based on 15+ KPI variables (e.g., if CPA exceeds the threshold and conversions are below 1, it decreases the bid by 20%).
It auto-generates a bulk-upload CSV ready for direct import back into the DSP.

The 15+ KPIs the Engine Tracks

CPA

Cost per Acquisition

click to flip

If CPA > threshold AND conversions < 1 → decrease bid 20%

ROAS

Return on Ad Spend

click to flip

If ROAS < target for 3+ days → pause line item, reallocate budget

CTR

Click-Through Rate

click to flip

If CTR < 0.08% → flag creative for swap, reduce bid 15%

CVR

Conversion Rate

click to flip

If CVR > 2x avg → increase bid 25%, expand audience

IMP

Impressions

click to flip

If delivery < 70% of pacing goal → increase bid, broaden targeting

PACE

Spend Pacing

click to flip

If spend > 110% of daily budget → cap bid, alert AM

VIEW

Viewability

click to flip

If viewability < 60% → block low-viewability domains

FREQ

Frequency

click to flip

If freq > 8/user/week → apply frequency cap, expand reach

WIN%

Win Rate

click to flip

If win rate < 15% → increase bid 10%, check floor prices

CPM

Cost per Thousand

click to flip

If CPM > 2x category avg → shift to open exchange, adjust targeting

VCR

Video Completion Rate

click to flip

If VCR < 65% → swap to shorter creative, boost CTV allocation

REACH

Unique Reach

click to flip

If reach plateau for 5+ days → expand geo, add new audience segments

CPC

Cost per Click

click to flip

If CPC > target AND CTR rising → maintain, watch conversion lag

BSS

Brand Safety Score

click to flip

If BSS < 90 → add domain to blocklist, verify ads.txt

ENG

Engagement Rate

click to flip

If ENG > 3x avg → increase budget 30%, replicate creative across campaigns

Impact

4 hours reduced to 5 minutes Zero manual entry errors 100+ line items per run

Video Walkthrough

Tool 2 of 5

Audience Insights Generator

Problem

Building data-driven insight decks for clients took 2 full days of manual data pulling, manipulation, and chart creation. Account managers were spending their time in spreadsheets instead of on strategy.

Solution

Ingests raw log-level data from the DSP export.
Runs Temporal Analysis (time of day, day of week performance patterns) and Contextual Analysis (top performing domains, content categories).
Generates demographic breakdowns and audience composition charts.
Outputs a formatted PowerPoint deck automatically, ready for the client.

Impact

2 days reduced to minutes 0 hours on chart-building 100% of time on strategy

Video Walkthrough

Tool 3 of 5

Dynamic Site List Management

Problem

Managing blocklists and allowlists across 10 different DSPs was inconsistent and error-prone. Each platform had its own format requirements. Without verification, client spend could end up on fraudulent or non-brand-safe domains.

Solution

A centralized tool that cross-references campaign site lists against the IAB ads.txt standard and internal transparency scores. It verifies domain legitimacy, flags non-brand-safe inventory, and auto-generates upload-ready CSVs formatted for TTD and DV360 - each in the exact format that platform expects.

Impact

8-10 DSPs, one source of truth IAB ads.txt verified Zero brand safety incidents

Video Walkthrough

Tool 4 of 5

Interactive Optimization Reports

Problem

Campaign optimization reports were static spreadsheets that took hours to compile. Traders and account managers had no quick way to see performance trends, compare line items, or drill into the data that mattered most for their next decision.

Solution

An automated reporting pipeline that pulls campaign data, processes it through Pandas, and generates interactive dashboards. Instead of a flat spreadsheet, the output is a filterable, sortable view of campaign health that updates with each new data pull - giving traders a real-time picture of what is working and what needs attention.

Impact

60x faster reporting Company-wide adoption Real-time campaign visibility

Video Walkthrough

Tool 5 of 5 · Built for Thomas J. Henry

CTV Measurement Framework

Problem

Thomas J. Henry - the legal anchor client - needed proof that layering CTV on top of Display was actually moving the needle. Last-click reporting credited Display because CTV is unclickable. The real question was: does CTV exposure shorten the time between first ad served and a signup? No off-the-shelf attribution tool answered that for a programmatic-direct law firm campaign.

Solution

Built a Python framework that ingested log-level data from DV360, The Trade Desk, and CTV-specific exchanges.
Joined exposure logs against client-side signup events using device graph identifiers and time windows.
Bucketed users into three campaign-structure cohorts: Display-only, CTV-only, and CTV + Display layered.
Computed time-to-signup (first exposure -> conversion) for each cohort. Surfaced uplift, recency curves, and incremental cost per signup.
Generated weekly cohort reports the AM team could hand the client without a data scientist in the loop.

The finding

Layering CTV with Display reduced time-to-signup by 5-7 days compared to Display-only structures. That number justified a budget expansion from $700K to $1.2M monthly on the account - and gave the AM team a defensible CTV measurement story for the rest of the book.

Impact

5-7 days faster time-to-signup $700K → $1.2M monthly budget Codebase handed off to Causal IQ at departure Reusable across the client book

Read the deep dive

The Pipeline

How the 5 tools work together

Four Python tools run the daily ad ops pipeline. The CTV measurement framework sits alongside it, scoring incremental lift across campaign structures.

The Adoption

From one person to the entire organization

The tools were not mandated. They spread because they worked. When you make someone's day 60x faster, adoption takes care of itself.

The Numbers

Impact at a glance

60x

Faster Reporting

Hours to seconds

Company-Wide

Standard Tooling

Adopted across the org

#1-2

Profitability Rank

Out of 22 managers

$900K+/mo

Ad Spend Managed

Peaking at $1.2M/mo

8-10

DSPs Simultaneously

DV360, TTD, Xandr, Yahoo, Zeta, Basis, Quantcast, Vistar...

5-7 days

CTV Lift on Time-to-Signup

Thomas J. Henry · $700K→$1.2M budget

Tech Stack

Built with

Languages

Python Pandas NumPy

DSP APIs

DV360 The Trade Desk Xandr Yahoo Zeta Basis Quantcast Vistar YouTube / TrueView

AI-Assisted Engineering

Claude Code Codex Cursor Copilot

Output

PowerPoint CSV Bulk Upload Interactive Dashboards

Standards

IAB ads.txt Brand Safety

Let's talk about what I can do for you.

I'm exploring leadership roles where building the technology and running the business aren't treated as separate jobs.

Start a Conversation → View on LinkedIn