When to Hire a Virtual Assistant: A Letter to Myself

On April 3, 2024, I paid a ₹1,200 GST portal late fee for the third month in a row because I could not find the OTP email in my inbox. The OTP had arrived. It had been buried under 47 unread emails, half of which were vendor newsletters I kept meaning to unsubscribe from. I sat at my kitchen table in Basavanagudi and looked at the ₹3,600 in wasted late fees over a quarter and thought, calmly and clearly, that I had a problem no amount of Pomodoro discipline was going to fix.

I am Murali, co-founder of mursa.me. I resisted hiring a virtual assistant for eight months before that April afternoon. The identity story I told myself was that founders bootstrap alone, that delegation was a later-stage cheat code I had not earned yet, that my situation was different because I was building an ADHD productivity workspace and it would look bad if I could not manage my own inbox. Every one of those reasons was rationalization. What broke me was not a productivity insight. It was a spreadsheet.

The week of April 8, 2024, I logged every recurring task I touched. Every calendar juggle, every invoice send, every support reply, every founder-adjacent context switch. At end of week the total was 12.4 hours. Twelve and a half hours out of a 40-hour week doing work that a $15-an-hour virtual assistant could do better than I was doing it, while our product roadmap slipped and I told my co-founder I would ship the auth flow "next sprint" for the third sprint running.

This piece is what I wish someone had handed me in August 2023, when the resistance started, or at least in December 2023, when the resistance became expensive. The 8 specific signals I use now to decide when to hire, current 2026 pricing so you know what to budget, the tasks to delegate first, the person I actually hired and how, the mistake I made trying to scale to a second VA at month four, and a Google Sheet you can copy that scores your readiness. No fake productivity percentages. Just the criteria I use now.

The 8 Signs It Is Time to Hire a Virtual Assistant

Any single sign is not enough. If you hit three or more in the same month, you are past the point where hiring a virtual assistant pays back inside a quarter. The Google Sheet above scores this automatically. Here they are in order of severity, each with the specific version of that sign I hit myself before I finally hired.

Sign 1: Admin work eats more than 30 percent of your week. Log a full week honestly. If email, scheduling, invoicing, data entry, and calendar cleanup take more than 12 hours out of a 40-hour week, you are paying yourself founder rates to do assistant work. Asana's [Anatomy of Work Index](https://asana.com/resources/anatomy-of-work) found that around 60 percent of the average knowledge worker's time is spent on "work about work" (coordination, chasing status, duplicating information) rather than the skilled work they were hired for. My April 2024 audit hit 12.4 hours out of 42 worked. Thirty percent on the nose. That was the number that finally moved me.

Sign 2: You are missing deadlines that are visible to customers. Not internal "we will ship it Friday" slips. Actual customer-facing misses. A demo that went out late, an invoice that got flagged, a support reply that took five days. My personal breaking point was a churn incident in March 2024. A pilot customer had emailed asking why their onboarding email hadn't fired. I saw the email, meant to reply that afternoon, got pulled into a call, and did not reply for six days. They cancelled. Losing a paying pilot because I could not process email cost more than a year of VA fees would have.

Sign 3: You are trading founder hours for admin work. Your effective hourly value is whatever an investor would pay for an hour of your strategic work, usually well over $100 if you are building a real business. If you spend that hour scheduling meetings a $15-an-hour VA could book, the math is losing you money every day you refuse to hire. I did the arithmetic once. At 12.4 admin hours per week times 48 weeks times a $75 effective hourly rate (a conservative estimate for founder work), I was losing roughly $44,600 per year to admin I could delegate for $12,000.

Sign 4: Your inbox has a permanent unread count above 50. Not a Monday morning spike. A permanent state. That means you are triaging, not processing. A VA can do first-pass triage, draft replies to routine threads, and surface only the emails that actually need your judgment. My permanent unread was 87 the week I logged tasks. It had been between 60 and 100 for eight months. I had also unsubscribed from things three times and re-subscribed by accident twice, which is its own diagnostic.

Sign 5: You defer strategic work to nights and weekends. The tell is not "I work late." It is "I only get real work done after 9pm because days are meetings and admin." That is a delegation problem masquerading as a discipline problem. In February and March 2024 I did all my meaningful mursa.me building between 9:30pm and 1am. My days were meetings, email, invoice chasing, and calendar dance. I told my wife I was "just in the crunch phase." I was not. It was structural.

Sign 6: Repeat tasks have no documented process. If the same 10 tasks recur weekly and only live in your head, you are the bottleneck by design. Documenting them for a VA forces the process to exist, which alone makes the business more valuable. This is the sign I resisted the longest, because writing the process felt like extra work when I was already drowning. The counter-intuitive truth: the week I finally wrote them for Rina was the week the drowning eased. Before she started.

Sign 7: You have said "I will get to that" about the same task for three or more weeks. That task will not get done by you. It will either be delegated, deleted, or haunt your list forever. Pick one honestly. My personal example: for six months I had been meaning to write a proper vendor-list template for our GST filings. It was on my Monday list for 24 consecutive Mondays. Rina wrote it on day 4 with me on a screen-share for 40 minutes. Six months versus 40 minutes because I insisted on doing it myself.

Sign 8: You have the cash to pay a VA for 90 days without stress. This is the readiness gate. Part-time VAs cost roughly $400 to $2,500 a month in 2026 (more on that below). If paying that for three months would meaningfully hurt your runway, the answer is not "don't hire." It is "hire for fewer hours first." The [SBA's guide to hiring your first employee](https://www.sba.gov/business-guide/manage-your-business/hire-manage-employees) covers the classification and payment basics if you are hiring in the US. If you are hiring internationally, the classification is simpler (independent contractor) but the pricing and time-zone math is more important.

How Much Does a Virtual Assistant Cost in 2026?

Pricing splits by geography and specialization. These ranges are based on Upwork listings, OnlineJobs.ph rate cards, and specialty agency rates I sanity-checked in July 2026. They are what founders actually pay in 2026, not what the agencies quote on their marketing pages.

The mistake most founders make is starting with the cheapest option and then quitting when quality disappoints. A better rule: pay the rate the outcome demands on your highest-leverage task, not the average of everything. If a bookkeeping error would cost you $5,000, a $12-an-hour generalist is not the answer for that specific task. Rina is a $18-an-hour generalist for inbox and calendar. For a payroll error I would hire a $60-an-hour specialist, not ask her to figure it out from a YouTube video.

What Tasks Should You Delegate to a Virtual Assistant First?

First-90-days delegation should be boring on purpose. Give your VA tasks that are low-stakes, high-frequency, and have a clear "done" state. Save the ambiguous work for month two, once trust is built and workflows are documented. This is the exact sequence I used with Rina and it is the sequence I would use again.

How to Hire a Virtual Assistant Without Regret

The hire itself takes about two weeks if you do not over-engineer it. The pattern that has worked for me and every founder I have compared notes with follows five steps.

1. Write the job post around outcomes, not tasks. "Own my calendar so no meeting takes more than one back-and-forth" beats "schedule meetings." Outcome-based posts filter for VAs who think, not just execute. My April 2024 job post was two paragraphs long: what a good week looks like for me, what a bad week looks like, and the outcome I wanted the role to own. I got 34 applications in 48 hours. The first draft of that post is in the Google Sheet linked above.

2. Do a paid trial task before the interview. Two to four hours of real work, paid, on something from your Week 1 list. Résumés lie. A paid trial does not. Skip anyone who will not do a paid trial. The good ones expect it. I paid four candidates $50 each for a 3-hour inbox-triage trial on a sandbox account. Two disappeared after the payment (learning experience). One did competent work with no judgment calls. Rina flagged a subscription renewal I had been meaning to cancel for two months and drafted the cancellation email without being asked. That single move was the whole interview.

3. Interview for judgment, not tool experience. Ask: "A client emails me at 11pm angry about a bug. What do you do?" You are hiring someone who will make dozens of small judgment calls a day. Experience with your specific stack (Notion, Superhuman, Cal.com, whatever) is trainable in a week. Judgment is not. In Rina's interview she said she would respond within an hour, acknowledge the frustration, promise a real response by morning, and would not commit to a fix or root cause without my input. That was the right answer, and she has answered dozens of real 11pm emails since with roughly the same script.

4. Start with 10 hours a week, not 40. Both sides learn the working relationship without either party being stuck. If it works, scale hours in month two. If not, you have lost $200, not $2,000. I started Rina at 10 hours a week for the first 21 days. By end of month one we moved to 20. By month three we were at 30. Ramping slowly gave both of us the exit ramp we would have needed if the fit was wrong.

5. Set a 30-day review on the calendar before you start. Two questions: what has been reclaimed, and what still is not working. Founders skip this and then let bad fits drift for months. Mine and Rina's first 30-day review was on May 3, 2024. Two agenda items. Twenty-five minutes. She surfaced two workflows where my instructions had been vague. I raised one where I felt I was still checking her work more than I wanted to. We fixed both by month end.

Who I Actually Hired (Rina, in Cebu, $18 an Hour)

Rina is a 32-year-old operations professional based in Cebu, Philippines. Before working with mursa.me she had spent four years as an executive assistant to a US-based agency owner, and one year as a founder's operations lead at a Bangalore fintech. I found her through OnlineJobs.ph in April 2024 by filtering for "executive assistant" plus "5+ years experience" plus "available part-time." I paid her $50 for a 3-hour trial. She started April 22, 2024 at 10 hours a week, $18 an hour, invoiced twice monthly, paid via Wise.

The first week was rocky in an instructive way. She sent me a Slack message at 9am IST on day 3 saying she could not find the shared password vault I had mentioned in onboarding. It did not exist yet. I had assumed 1Password was set up because I had been meaning to set it up. That morning I spent an hour setting up 1Password, a shared vault, and access controls I should have had in place a year earlier. Rina's presence forced infrastructure I had been avoiding, which is one of the underrated benefits of hiring a VA. You cannot hand off what you have not first codified.

The moment it clicked for me was Tuesday May 7, 2024, roughly two weeks in. A pilot customer had raised a support ticket that involved a billing dispute plus a technical question. Rina wrote me a Slack message at 11:04am IST that said, in essence: the technical part needs you, but the billing part I can handle if you approve the refund amount, and here is the draft reply for both. It was the first time I felt the reclaimed founder hour in real time. Not "I have more free time." Actual: the thing I would have context-switched to had already been triaged with judgment. That was the whole point.

The specific rate matters. $18 an hour is above the Philippines median of $8 to $12 I quoted in the pricing table. I paid the premium because Rina had domain experience (SaaS ops, founder-adjacent) and because I wanted judgment more than I wanted volume. Ten judgment-hours a week at $18 was better for me than 20 execution-hours at $9. That match is different for every founder, but the principle holds: pay for the layer you actually need.

As of August 2026, Rina and I have been working together for 28 months. She now runs at 30 hours a week, handles our support-tier-one queue, our founder inbox, calendar, invoicing, and month-end reconciliation with the CA. She has taken exactly one week off unannounced (family emergency, week of December 4, 2024), and one planned two-week Christmas break in 2025. She has a full second client on Fridays. When I say hiring her was the single highest-ROI decision I made as a founder, I am not exaggerating.

What Actually Changed After Hiring (Month by Month)

I am skeptical of blog posts that quote precise productivity gains, so I will only claim what I can honestly attribute month by month.

Month 1 (May 2024): reclaimed roughly 6 hours a week. Not the full 12.4 the audit had shown, because I still checked her work more than I needed to and because ramp-up on any hire is bumpy. Mursa product velocity: no meaningful change yet. Emotional state: relieved but suspicious of the relief.

Month 2 (June 2024): reclaimed roughly 10 hours a week as I stopped double-checking inbox and calendar. Product velocity: we shipped the auth flow I had been promising for four sprints. My co-founder specifically noticed and said "you seem to have unblocked yourself." I did not tell him it was because I had hired someone. That is on me.

Month 3 (July 2024): full 12 hours a week reclaimed. Rina now owned invoicing and expense reconciliation. We shipped two more features. I stopped starting my day inside Gmail, which is the single most durable change from the whole experience. Walking to the desk and opening the codebase before the inbox is the reason we shipped three meaningful mursa.me features in the quarter after hiring, versus one in the quarter before. I cannot cleanly separate Rina's contribution from other variables (fresher sleep, better routines, market fit signals), but the correlation is strong enough that I have never considered going back.

Months 4 through 12 (August 2024 to April 2025): the compounding phase. As Rina took over more of the operational surface, my calendar increasingly reflected only work I could not delegate. Investor updates were fully drafted by her from Notion templates I approved. Pilot customer onboarding calls had prep docs before I joined. My inbox response time went from 4 to 5 days to under 24 hours because Rina replied first-pass to 70 percent of it and only surfaced the 30 percent that needed my judgment. That was the whole point.

What Did Not Change

The amount I worked. I still work long hours. The hours are just aimed at building product and talking to users instead of routing emails and reconciling calendar invites. If your goal is fewer hours total, a VA alone will not get you there. That requires cutting scope, not just delegating. A VA cannot want less for you. That decision stays yours.

The Month-Four Scaling Trap

In September 2024, five months into working with Rina, I made a mistake I want to save you from. I decided the answer to "Rina at 20 hours is great, so 40 hours would be twice as great" was to hire a second VA. I posted a second job listing, ran a similar trial, hired a competent generalist named Marco based in Manila, and started dividing tasks between them.

It broke by week three. Not because Marco was bad (he was not). Because I had not thought hard about the delegation ceiling of a solo founder. The bottleneck was no longer "admin volume Rina and I could not handle." It was "context I could hand off cleanly." Two VAs needed twice the context, twice the training, twice the review cycles from me, and I did not have twice the context to give. What worked at 20 hours a week with one person did not become 40 hours across two people. It became one person doing the work well and a second person waiting for me to explain things properly.

I let Marco go after 30 days, paid the agreed month plus a two-week transition bonus, and asked Rina to expand from 20 to 30 hours. That combination worked immediately. The lesson: a first VA scales your hours. A second VA scales your team, and requires operational maturity most solo founders do not yet have. Hire the second when you have a clear operations layer, not because you have more admin than one VA can absorb. If you cannot articulate what your first VA's shift-management is, you are not ready for a second.

A Note to the Founder Who Will Not Hire

If you are reading this and you are the version of me from August 2023, telling yourself you will hire once you have more revenue or once the current sprint calms down or once you finish documenting the process, I want to be direct. The current sprint will not calm down. It never does. Revenue does not grow faster because you are personally scheduling calls. And the process will not get documented until someone else needs it, which is a delightful chicken and egg you can break by hiring the person who forces the documentation.

The version of your future that resists hiring longest is the one that stays smallest longest. That is not a productivity claim. It is a math claim. The hours you spend on delegable admin are hours not spent on the work only you can do, and every quarter you delay that switch is a quarter of compounding you will not get back. Rina cost me $180 in month one. She saved me from a $3,600 in wasted GST late fees. She saved me one churn incident that would have paid for her for a year. She helped me ship features that closed a Seed round in October 2024.

Copy the Google Sheet above. Log one honest week. Read the score. If it says "hire," post the job the same day. If it says "not yet," it will tell you exactly what to fix first (usually documentation, not more effort). Either way, you will have moved from "maybe someday" to a decision with a next step. The version of you who avoids the decision keeps paying the ₹1,200 GST portal late fee every month, quietly, in different currencies.