ABOUT
I’m a product manager with nearly a decade of experience getting hands-on with product development, from user research to product design across big tech and start-ups, with a specialization in AI, ML, and LLMs.
HISTORY
| 2023 | Senior Product Manager Microsoft – Seattle, WA |
| 2020 | Product Manager 2 Microsoft – Seattle, WA |
| 2017 | Product Manager Microsoft – Dublin, Ireland |
| 2016 | Software Engineer Hawk-Eye – Basingstoke, UK |
| 2014 | Software Engineer EarSoft – Newcastle, UK |
PROJECTS
Helping people write prompts (LLMs in Word)
Increasing feature engagement in Microsoft Word (AI in Microsoft Word)
Brand design for personal finance blog (Finance with Mark)
Broadcast enhancement for live sports events (Hawk-Eye)
Built audio analysis library for ignite 100 start-up (EarSoft)
Helping people write prompts
Senior Product Manager – Microsoft Word
When ChatGPT released, I enthusiastically discussed with friends how we were each using it to perform research, draft reports, and even get free personal therapy. So when it came across my desk, helping Word users discover how to get the most of out Microsoft’s Copilot offering in Word was a dream project.
Understanding users
I ran user interviews, discovering how advanced users learned how to write detailed “templated” prompts to tailor the output they received from generative AI tools, using multi-turn conversations to work with the model to complete their task.
Understanding the technology
I worked closely with Applied Scientists to understand quality and latency constraints of GPT-Turbo and how we could modify the output through metaprompts.
A familiar UX
The fun came in coupling our understanding of users and state-of-the-art technology to build a familiar yet innovative UX to provide users with suggestions that didn’t interrupt their workflow but enabled them to save time and discover how to get better results from Copilot in Word.
Increasing feature engagement in Word
Senior Product Manager – Microsoft Word
Shipping 12 experiments to drive feature engagement for Editor (Microsoft’s version of Grammarly).
Note: Details ommitted or replaced to protect NDA.
OKRs
In a large product like Microsoft Word, metrics like MAU are slow to move, limiting fast experimentation.
I partnered with Data Science to identify leading indicators that correlated to the metrics we ultimately wanted to move.
We added leading indicators to experiment scorecards to allow us to see movement within weeks instead of months, increasing experimentation velocity.
Research
I led multiple rounds of research to understand which product areas to improve to increase engagement:
- Literature review
- User interviews (qualitative, 10+ sessions)
- Surveys (quantitative, 100+ responses)
We enumerated key areas of the product, such as providing basic spelling suggestions, advanced writing suggestions, and formatting suggestions, and asked two questions for each area:
- What is the importance of this?
- What is your satisfaction with this?
Combining these answers into a formula gave us an opportunity score for each feature area:
These feature areas were then prioritized when brainstorming experiments.
Brainstorming
I ran a set of brainstorming exercises across Engineering, Design, and PM, using Design Thinking exercises including Crazy 8’s, How Might We, and SCAMPER to generate 80+ experiment ideas.
Prioritization
I led prioritization using an adapted RICE framework to narrow in on 20 experiment ideas that we could spec.
Experiments
I led speccing these experiments, iterating based on feedback, and collaborating with engineers to ship to a portion of our production audience.
I analyzed the experiment scorecards, shared learnings with others, shipped successful experiments, and updated our hypotheses for future experiments.
Next steps
Updated the roadmap based on learnings from experiments and shared with leadership.
Illustrations sourced from Figma templates.
Brand design for personal finance blog
Founder – Finance with Mark
I’ve been interested in personal finances for a long-time. I’ve helped friends sell Bitcoin and get into index funds, use tax-advantaged accounts, and generally feel calmer about their money situation.
I took this passion to the next level by starting to write about living a stress-free financial live at FinanceWithMark.com.
It’s been a joy building a product hands on, here’s some of the skills I’m most proud to have developed so far.
Creating assets
My favorite asset has got to be “Wallie the Wallet”, the site mascot who embodies an optimistic approach to the journey to wealth. I’m also a big fan of “Mr Money” and his magic ability to keep his dollar bill feet firmly planted.
I’ve also learned how to creating assets from scratch in Adobe Illustrator. Draw Your Words was an instrumental course for improving my confidence.
Writing
I’ve written posts that blend personal finance with personal truths, and in sharing my journey, have helped others discover the deeper why to their relationship with money too. Discovering my voice to represent myself authentically is paying dividends beyond posts on the site.
Broadcast enhancement for live sports events
Software Engineer – Hawk-Eye
I leveraged Hawk-Eye’s industry-leading calibration and broadcast systems and added a foreground-background detection process to add player’s national flags to the TV broadcast. The solution ran across 7 courts at the Australian Open 2016, airing to 1.2M viewers for the final.
It was my first product launch, where I trained system operators on-site, experienced the pain points that exist between developing software in an office and running it in-person on-site 10,000 miles away, and felt the draw to sit in the sweet spot between tech and customer, ultimately paving the way to my path to product management.
Built audio analysis library for ignite 100 start-up
Software Engineer – EarSoft (Ignite 100 start-up)
EarSoft was my first taste of building commercial product, a three-month internship where I built an audio analysis library in C, with a C++ front-end to demonstrate the impact of EarSoft’s patented audio normalization to solve the age-old problem of spoken scenes being too quiet when you’re watching a movie at night, but then the action scenes blowing your ears off.
Ultimately the product wasn’t successful because getting TV manufactures to adopt a new volume control that operates with a range instead of a simple volume up/down control was too challenging, but I cherish my experience working with an excellent team of kind and smart engineers.