You've Hired Three VAs. None of Them "Worked Out." Here's Why.
Your third VA just asked you the same question your first VA asked eighteen months ago: what do you want her to do with an email that doesn't fit any category in the training doc. You've rewritten the job description twice. You've switched agencies once. You've interviewed six candidates for a role you keep describing as "someone more proactive this time." The failure pattern is identical every time — which is the part nobody's naming.
You've Diagnosed the Wrong Variable
When a VA doesn't work out, the instinct is to audit the person. Was she slow. Was she not detail-oriented enough. Did the agency send you someone who wasn't a fit for your "vibe." Did she just not care as much as the last one seemed to, before she stopped seeming that way too. So you fire, you re-post, you interview again, you pick more carefully this time — screen for "self-starter," ask behavioral questions about ambiguity, run a paid trial week — and somewhere around month three, the new hire starts doing the exact same thing the last one did: forwarding things back to you with "just wanted to double check" attached to decisions that were never actually hers to make alone.
That's not a coincidence, and it's not a worse hire than the last one. That's a role producing the same output regardless of who's plugged into it, which means the problem was never sitting inside any of the three people you hired. It's sitting in the seat itself. You can interview harder, pay more, hire through a fancier agency — none of it changes what happens the first time she hits a situation the training doc didn't anticipate, because the training doc was never the thing standing between her and a good decision.
The Loom Video Isn't the Gap
Most founders' second move, after "hire someone else," is "document more." Record another Loom. Write another page in the SOP doc. Make the training more thorough so the next person "gets it" faster. This feels like progress because it produces something — a doc, a video, a checklist — and it's also the wrong fix for what actually breaks.
SOPs are excellent at one thing: telling someone exactly what to do in a situation you already anticipated. They are structurally incapable of covering the moment your VA hits an inbox item that doesn't match any bullet point — which, in any real business, is most Tuesdays. A client asks for something slightly outside scope. A vendor invoice is $200 higher than usual. A lead replies at 11pm asking a question that's 90% covered in the FAQ and 10% not. In that moment, she has two options: guess and risk being wrong in front of you, or escalate and wait. Every VA you've ever fired picked option two, because option two is the rational choice when nobody's told her what she's allowed to decide without you. You didn't hire someone unproactive. You hired someone smart enough not to gamble with your business on your behalf — and then built a role that punished her for exactly that instinct by calling it a lack of initiative.
Why the Job Ad Keeps Attracting the Same Result
Swap the candidate, keep the role exactly as it is, and you'll get the same outcome — not because you keep losing the hiring lottery, but because the role determines the ceiling, not the person sitting in it. This is easier to see at different revenue tiers, because the same setup that works fine at one size becomes unworkable at the next, with nothing about the VA herself changing at all.
At $300K, a generalist VA absorbing overflow works fine. You're still close to nearly every decision in the business, so when something ambiguous lands in her inbox, you're one Slack message away and the delay costs you nothing real. At $800K to $1M, the business has enough surface area — enough clients, enough moving parts, enough decisions happening in parallel — that "just ask Whit" stops being a minor friction and starts being the actual bottleneck. Past $1.5M, it gets worse in a way that's easy to miss: you've usually added a second and third person by then, which means the ambiguous items aren't routing to you once anymore, they're routing to you from three different directions at once, all slightly urgent, all slightly outside what anyone was told they could touch. The VA isn't failing more at each stage. The gap between what she's equipped to decide and what the business now needs decided has just gotten wide enough to see — and wide enough to cost you actual hours every week.
What Actually Has to Exist Before the Next Hire
The fix isn't a better candidate, and it isn't a longer onboarding doc. It's mapping, in advance, three things no job description usually covers: what she can decide alone and never needs to mention, what she should decide and then loop you in on afterward, and what genuinely has to come to you before it moves. Most roles you've built only have two settings — "handle it" and "ask me" — which means anything remotely ambiguous defaults to the second one by design, not by accident. There's no middle tier, so everything that isn't explicitly pre-approved collapses into "escalate to founder," which is what's actually filling your inbox back up every week regardless of who's sitting in the seat.
That three-tier map is the actual infrastructure a VA role needs to hold weight — not a longer SOP, a different kind of document entirely, one that captures the logic behind your decisions instead of just the steps. A refund request under $50 with no prior complaint history: decide it, no need to mention it. A refund request from a client who's already flagged twice: decide it, then tell me why. A refund request over $500 or from anyone on a retainer: that comes to me first, always. That's a decision framework. "Handle customer service issues professionally" is not — it's a vibe dressed up as an instruction, and no VA, however good, can execute a vibe consistently.
Without that framework, you're not hiring an assistant. You're hiring a very expensive relay system that routes everything back to the same single point of failure it was supposed to relieve. With it, the same tasks that used to require your judgment now have somewhere else to go, because someone besides you knows what "the Whit answer" would be in most situations, and has actual standing permission to give it without checking first.
The Fourth VA Isn't the Fix
If you post the job again without building this first, you're not starting fresh — you're restarting the same ninety-day clock with a new name attached to it. The next person will train the same way, hit the same category-less email in month two, and end up back in your inbox asking permission for something you thought you'd already delegated. She'll get blamed for it in the exit interview you have with yourself. She won't be the reason.
The role has to be built before the person can succeed in it. That's not a hiring problem, no matter how much it feels like one every time you're rereading resumes at 9pm. That's an architecture problem — and it's the same one whether the seat says virtual assistant, ops manager, or client success lead.
If this is the third time you've had this exact conversation with yourself, the Founder-Proof Quiz is a good next fifteen minutes — it's built to show you exactly where the decision-making gaps are sitting before you write job description number four.