Starting a writing habit, in public
Eight years of migration notes have gone into documents about four people read. Here's the plan to fix that.
I have been a customer engineer at Google Cloud since 2018, first in Gurgaon and since the end of 2021 in London. Most of that time has gone into the same kind of problem: an enterprise has decided to move something large and awkward, and someone has to work out what the target architecture is, what it will cost, and why the last three attempts stalled.
The output of that work is almost always a document. A migration discovery framework. A TCO model. A root cause analysis explaining why a hybrid VPN route was black-holing traffic for a week. These get read by the account team, maybe the customer’s architect, and then they go into a Drive folder and nobody opens them again.
Which is a waste, because the useful part is rarely the conclusion. It’s the reasoning — the thing you found out on day three that made the original plan wrong.
So I’m going to write one post a day here.
What I expect to write about
My day job is the ISV side of the partner ecosystem: software vendors building on Google Cloud, and the GSIs and RSIs who deploy for them. That produces a steady supply of things worth writing down.
Marketplace onboarding, which almost everyone underestimates, and not for technical reasons. Agent engineering — I’ve been building internal agents with ADK, A2A and MCP for problems like TCO analysis, and the distance between a working demo and something you’d let a colleague depend on is wider than the demos suggest. Migration architecture, where the interesting failures are usually in identity and networking rather than in the workload. And the commercial side, because a migration business case is a technical artefact whether or not anyone calls it one.
Some posts will be two paragraphs and a diagram. That’s fine. The point isn’t that each one is significant.
Why daily
I’ve tried the considered-essay-once-a-month version of this before and it produced nothing, because there was always a reason to wait until I understood the topic properly. A daily cadence removes that excuse. It also means the writing happens while the detail is still in my head, rather than reconstructed from a calendar entry six weeks later.
If you work on any of this — building on Google Cloud, getting a listing through Marketplace, moving SAP somewhere it doesn’t want to go — I’d like to hear where you disagree. Contact details are on the CV page.