David Magee

Product Director at Boden | Views are personal | Formerly: Product @ World of Books, Energy Innovation @ Bowman Power, Aerospace @ QinetiQ | WBS MBA

[David Magee @DavidMagee]

Thanks @harleyf - we made it - Oct launch and >1M orders by early Dec.

Shopify x WOB:
7M SKUs
5M daily reprice events
4 currencies
3 languages
2 international warehouses
1 platform
Biggest BFCM, ever.
Biggest Christmas trading, ever.
2025 will be WILD.

[David Magee @DavidMagee]

I’m thrilled to be joining @Bodenclothing as Director of Product & Portfolio.

Why Boden? They’re a British heritage brand I’ve grown up with, and there’s definitely something special about joining a business that’s loved by three generations of your own family.

Boden is also at a really exciting moment: back to profitability, live on a fresh @Shopify platform, approaching £400M in revenue with a fast-growing US business, the first US store has opened in Atlanta and another shortly in Nashville.

I’ll be leading product, portfolio, UX and AI, and I’m super excited to spend time getting to know our customers, our team and the tech stack!

More soon.

[David Magee @DavidMagee]

Of course every large platform experiments. That's kind of a given.

But to make the Meta comparison is simply a category error - Meta is rented audience. The entire deal with Meta is that you don’t control the algorithm, you buy probabilistic access to attention and the black box is the product you’re renting.

Experimentation isn’t a violation there; it is the thing you bought. Shopify is the opposite. It’s your store infrastructure, the system of record for a channel you own etc.

I'm massively pro Shopify but also cautious about things like this

[David Magee @DavidMagee]

@andrewjfaris Depends on the decision and what they optimize for.

In aggregate, we (merchants) get better checkout, but not always in the moment.

[David Magee @DavidMagee]

@JamesRNail @ericries Sure, but “more customers” assumes they exist to be won. If every competitor gets the same productivity boost, nobody gains share, prices fall, and the gains leak to consumers. Cost-cutting is the rational individual move precisely because demand is fixed. That’s the trap.

[David Magee @DavidMagee]

@SBarrettBar Her Grace the Communist. Laughed out loud. Fantastic.

[David Magee @DavidMagee]

@s_capatina @tobi @jforjacob Can you read?

[David Magee @DavidMagee]

@nic_carter If you're in HR pivot to UBI

[David Magee @DavidMagee]

This feels like a wild thing to AB without merchants knowing/having control. At perhaps the most critical point in the customer journey, changing the user experience. I am so insanely in favour of how Shopify continues to adapt the platform and make it better for both merchants and customers but this feels like it is slightly too much, asp if the underlying hypothesis is one that may primarily benefit Shop.

[David Magee @DavidMagee]

Who can intro me 😂

[Polymarket @Polymarket]

NEW: AI consultant reveals a client accidentally spent $500,000,000.00 in a single month after failing to set employee limits on Claude usage.

[David Magee @DavidMagee]

@mstockton Wrote about some real-world implementation lessons here from PE-backed scale up

lmk if you think it is useful!

https://t.co/927UqRAP9U

[David Magee @DavidMagee]

https://t.co/hZZoneobid

[David Magee @DavidMagee]

The model-obsolescence point is the one most enterprises will under-build for. The fix is to make the AI 'factory' or harness the unit of investment, not the agent.

Architecture specs change when the technology changes; feature specs survive a rewrite of the system beneath them.

A proper harness (instructions, evals, tools, guardrails, observability etc) absorbs the model swap.

Enterprises that treat each agent as a deployable artefact will pay the re-do tax every time a model ships, but I think the ones that treat the factory as the unit will not.

[David Magee @DavidMagee]

Super interesting! Have been working at WoB on "Modular Dark Factory" - idea borrowed from Fanuc's lights-off factory in Japan. Idea: a spec goes in, deployable software comes out, humans are absent from the production loop but very present in defining and improving the system.

We've built (currently) six interlocking modules:
•⁠ ⁠Declarative feature specs (the input layer)
•⁠ ⁠Testing harness (machine-checkable acceptance criteria, incident-to-test loop)
•⁠ ⁠Security agent (WoB ASVS + AppSec standards, runs on every PR in CI)
•⁠ ⁠Data layer (mart-layer contracts, PII boundary at staging)
•⁠ ⁠Observability + "the Mayor" (decision authority that dispatches RCA / Dev agents)
•⁠ ⁠Deployment pipeline with auto-rollback on post-deploy contract breach

[David Magee @DavidMagee]

@eliweisss In the UK…Vervaunt are INCREDIBLE. Shout if you want an intro.

[David Magee @DavidMagee]

@rorysutherland Big fan. As a Shopify merchant, it’s also an (almost) entirely free new acquisition and retention channel.

Shopify also had some amazing paid campaign ROI guarantees on Shop App at one point too.

[David Magee @DavidMagee]

By the time our leadership team got around to a serious conversation about AI strategy, the staff of World of Books had been running one for the better part of a year.

Here's the story of AI implementation at an enterprise Shopify merchant:

https://t.co/927UqRAP9U

[David Magee @DavidMagee]

https://t.co/hZZoneobid

[David Magee @DavidMagee]

@PEoperator Some more broad organisation reflections here - not sure if useful! https://t.co/39ied2UaAq

[David Magee @DavidMagee]

https://t.co/hZZoneobid

[David Magee @DavidMagee]

https://t.co/hZZoneobid

[David Magee @DavidMagee]

@rorysutherland Hitchens will concede on the cathedrals, at least. Surely.

[David Magee @DavidMagee]

@PEoperator Pretty SWE specific - will write on org-wide shortly

[David Magee @DavidMagee]

@PEoperator I wrote a piece on what I’ve been up to in a PE portco - we’ve gone pretty deep on this over the last 6-9

https://t.co/lI4FQnSp7C

[David Magee @DavidMagee]

https://t.co/Hg1iV839dk