Four titles, one decision

Context

I moved to Dubai and started looking for work in operations. The first thing a job search asks you to do is name yourself, and I named myself three times before I understood what the question was actually asking.

AI Operations Specialist. The first one. Accurate to what I believed I was building at the time, and it read as current.
AI Operations Lead. Same conviction, new words. It felt like a step up in seniority, which is the only thing it changed.
Operations & Process Transformation Lead. The most honest description of the work I want to do, and the least searchable string I have ever written.

Each one felt right when I wrote it. None of them produced a single interview.

The constraint

Three rewrites and no interviews leaves you two explanations, and the tempting thing about both is that they let you stop thinking. Either I wasn’t good enough, or the market was closed. Neither is falsifiable, and neither tells you what to change on Monday.

The real constraint sat underneath both of them: I had no way of finding out which was true. I kept changing an input and never read the output — three edits with no success criterion attached, so not one of them could fail informatively. That isn’t a job search. That’s a person rewriting a sentence.

The self-audit

So I stopped rewriting and audited the thing itself. The finding was uncomfortable and simple: the market does not search for titles you invent for yourself.

Recruiters and matching algorithms work on language that already exists — the vocabulary already sitting in job postings, in search filters, in the saved searches a recruiter runs every morning. “Transformation Lead” describes an intent. It isn’t a role anyone types into a search bar, so it never enters a result set, so it never gets evaluated at all. A title that doesn’t enter the funnel can’t be rejected by it — which is exactly why three rounds of rejection-free silence felt like nothing was happening. Nothing was.

The mechanism is the part that mattered. I had been optimizing for how the title felt to me, when the only thing a title is load-bearing for is being found.

What I did

The fourth change was not another identity-driven flip. It was a deliberate empirical test, and the difference is that I set the criteria before I wrote the words. Three of them, all of which had to hold at once:

It has to have real search volume. Not a title I liked — a title that already exists in postings and in the searches recruiters actually run. That constraint on its own eliminated all three of my previous headlines.
It has to match the profile I actually have. Analyst, not Lead. I’m a candidate with a limited UAE track record, and a seniority claim this market can’t yet verify costs more than it earns. The honest level is the one that gets read.
I have to be able to defend it end to end. Every claim behind the title backed by concrete process-mapping and automation work I can walk through in an interview, in detail, without reaching. A title that wins the search and loses the call isn’t a win.

Business Process Analyst was the only string that cleared all three. I rewrote the profile around it, published it, and then did the part I’d skipped three times already: left it alone long enough to read the result.

The transferable part

The title isn’t the thing that transfers to the job. The method is: form a hypothesis, put it in front of the system, read what comes back, adjust. That’s the same discipline I bring to a workflow that keeps breaking or a cost calculation that stopped reconciling — including the one where I rebuilt an ERP’s cost logic outside the ERP specifically so I could read what it was doing. A change you can’t measure isn’t a test. It’s a preference.

Results

The first call came shortly after the profile went live — the first one in three months of trying.
Two more recruiters reached out inside the first couple of weeks, taking it to three conversations where the previous three titles had produced none.
Each conversation returned data, not just an outcome. What this market actually hires for, which words in a posting are load-bearing and which are decoration, where the profile still reads thin. That’s the input to the next iteration, and it’s the part a callback count doesn’t capture.

What I’d do differently

I’d have treated the first title as a test, not a statement. Pivots one through three had no success criterion attached, which is why three months of them taught me nothing. The same three edits with a defined signal to read would have found the answer by June.
I’d have set a fixed test window and an application count for each title. Every rewrite ran until I lost patience with it, which is not a stopping rule. A committed period per title, with a defined number of applications inside it, does two things I didn’t have: it forces me to retire a title that isn’t landing instead of letting it drift, and it produces an applied-to-interviewed rate per version — a real number to compare, rather than a general sense that nothing was working.
I’d have read the market’s language before writing my own. The search-volume check that picked the fourth title was available before I wrote the first one. Nothing about it required three failures first — it only required asking how the title would be found instead of how it sounded.

Looking back

Four rewrites felt like drift while they were happening. Looked back on, only the first three were drift. The fourth was the moment I stopped asking what sounded right and started asking what was true.