Prepared for Dropbox
Tony Lee × Dropbox
A working view of how brand and performance media could be planned together. Based on the role description, with questions I’d want to explore with the team.
01 · What I’d want to understand
Start with the decisions the team needs to make.
- Brand and performance. How are they planned today, and where do their goals or measurement disagree?
- The media portfolio. Which channels are earning more investment, and where is the evidence less clear?
- Business outcomes. How does media contribute across self-serve signups, product-led growth and sales-assisted pipeline?
- AI in the workflow. Which recurring tasks are suitable for automation, and what would need human review?
- Measurement. What can the existing data, experiments and Marketing Data Science team already tell us?
02 · A first-ninety-days outline
Learn enough to make the first changes count.
This sequence would depend on the team’s priorities, existing commitments and available data.
-
Days 1–30
Understand the current plan
- Bring spend, objectives and measurement into one view.
- Meet the channel leads, agency partners and Marketing Data Science team.
- Identify the decisions that are currently hardest to make.
-
Days 31–60
Agree on the next decisions
- Establish how evidence will inform investment changes.
- Bring my 70/20/10 allocation approach into the discussion, including how it fits existing commitments.
- Scope a measurement test and a small automation pilot with clear checks.
-
Days 61–90
Act on what is ready
- Make changes supported by the evidence available.
- Start or review the first test according to its required timeline.
- Share the reasoning with the team, and clarify ownership of what happens next.
03 · One experiment to consider
What does upper-funnel media contribute?
A matched-market holdout could help estimate the contribution of video or social activity. The first step would be agreeing on the outcome, selecting a defensible comparison and allowing enough time for the effect to appear.
Self-serve signups and sales-assisted pipeline may need different observation windows. I’d want to understand those before proposing a duration or promising a result.
04 · Automation and responsibility
Give automation a bounded job.
Pacing checks, reporting drafts and anomaly flags are candidates to evaluate. For each, I’d want to know the inputs, the cost of a wrong answer, how the output is checked and who makes the final call.
Building this site with AI shows my willingness to work directly with the tools. A production media workflow would need its own evidence of reliability.
05 · Relevant experience
Where my experience connects.
| Area | Evidence |
|---|---|
| Media and growth | 13+ years across in-house and agency work, including Shopify, Yamibuy and Starz. |
| Brand and performance | Experience across entertainment launches and e-commerce acquisition. |
| Acquisition and measurement | Yamibuy acquisition results: CAC 15% below baseline, with volume up 25% year over year. |
| Conversion decisions | The Shopify interactive experience that achieved the highest session-to-lead conversion rate on record. |
| Building an operation | First U.S. employee; a book of twelve accounts in eighteen months. |
| People | Recommendations from former direct reports describing their development and experience working with me. |
| Building with AI | This interactive site: a concrete building example, with its limits made explicit. |
Contact
Worth a conversation?
I’d like to hear which of these questions matters most to the team, and where this first view misses something.
DECLÉOR
Soffe
Party City
Sam's Club
sweetgreen
Nintendo
20th Century Fox
Starz
FX Networks
Shopify
Fandango
Searchlight
A24
Google