A year ago almost to the week, I was a marketing analyst who'd started to feel the walls of the job closing in — brilliant at reading the numbers, locked out of the machinery that produced them. Twelve months on, I've spent the year up to my elbows in that machinery: SSIS, BIML, PowerShell, SQL Server, the whole enterprise ETL stack I could barely name last December. It seems like the right moment to write the honest ledger — what the switch actually took, and what it gave back — while it's still fresh enough to be true.

What it cost

Let me not romanticise it. Changing fields mid-career has a real bill, and pretending otherwise doesn't help anyone standing where I was.

The expert-to-beginner drop. I went from being the person others asked to being the person who asks. That's a genuine ego cost, and if you're honest with yourself, it stings more than you expect. There were weeks where I watched colleagues do in ten minutes what would've taken me an afternoon, and I had to sit with that rather than pretend it wasn't happening.

The evenings. The on-the-job learning wasn't enough on its own, so a lot of this year happened after hours — certifications, tutorials, breaking things in a sandbox to understand why they broke. That time came from somewhere. It always does.

The temptation to fake it. The hardest discipline all year was resisting the urge to nod as if I understood, to protect my pride. Every time I gave in to that even slightly, it cost me later. The learning only worked when I was willing to be visibly, repeatedly the person who didn't know yet.

What it gave back

And the other side of the ledger, which is why I'd do it again without hesitating.

Depth I actually wanted. I no longer stand at a locked door. When a number is wrong now, I can follow it back through the pipeline to where it went wrong and fix it at the source. That's the thing I moved for, and having it feels exactly as satisfying as I hoped.

A rare combination. Here's what I didn't fully see coming. Most people in enterprise ETL came up through engineering and never spent years thinking about how to communicate a number to a non-technical human. Most people from my old world can read data but can't build the plumbing. I'm becoming someone who can do both — build the pipeline and explain what comes out of it to someone who'll make a decision on it. A year ago I worried my communications background was baggage I was leaving behind. It turns out to be the most transferable thing I own.

Proof to myself. There's a quieter gift, too. I now know I can walk into a field where I'm a beginner and become genuinely competent in a year through nothing but stubbornness and evenings. That's a thing you can only learn by doing, and it changes how you look at the next intimidating thing.

The skills ledger, concretely

Let me put actual names to "learned a lot," because vague growth stories are cheap and a specific one might be useful to someone plotting the same jump.

Twelve months ago I could not have told you what any of these were. Now they're tools I reach for without thinking:

  • BIML and BimlScript — generating whole families of SSIS packages from metadata instead of building them by hand, which quietly reframed how I think about all repetitive work.
  • PowerShell for deployments — turning "click through the deploy and pray" into a script that runs the same every time and tells you afterwards what it did.
  • Real error handling — quarantine tables, OnError handlers, transactions that roll a failed load back to a clean state instead of leaving it half-done.
  • Incremental loading done safely — watermarks that only move inside the transaction, so an interrupted run can never lose a row.
  • Data Vault — and, more usefully, a feel for when its complexity earns its keep and when it doesn't.

But the through-line matters more than the list. Almost everything I learned this year is a variation on one theme: make it consistent, make it repeatable, make it fail loudly rather than silently. That's not a set of tools. It's a way of thinking about building things other people depend on — and it's what I'll carry into whatever comes next, long after these specific technologies have moved on.

What I'd tell last-December me

If I could send one note back down the year, it'd be short.

The humbling is the price of admission, not a sign you made a mistake. Pay it cheerfully. And stop worrying that your old skills are wasted — they're the thing that'll set you apart once the new ones catch up.

I don't know exactly what next year holds — more depth, cloud, probably things I can't name yet the way I couldn't name BIML twelve months ago. If there's one bet I'd place, it's that the pull towards the cloud only gets stronger, and that the skills worth having will be less about any single tool and more about the judgement of when to use which. But the direction is set and the doubt is gone. The dials were a fine place to start a career. The machine is a much better place to build one.

Thanks for reading along this year. Here's to whatever the next one turns out to be built from.