Virtual assistance and back-office operations

What this actually solves

The arithmetic is uncomfortable once you write it down. A great deal of your week goes into work that is genuinely necessary and worth far less than your own hour is worth — merging duplicate CRM records, chasing an invoice, rearranging a meeting, copying order details between two systems that do not talk to each other. None of it can be dropped. All of it is being done by the most expensive person available.

The obvious remedy is to hire someone, and for a small business that is a much larger step than it looks. A desk, a contract, payroll, statutory obligations, someone to manage them, and a commitment that is awkward to reverse if the volume of work turns out to be two days a week rather than five. This service exists in the gap — the work gets done, the commitment stays proportionate, and it can grow or shrink without a restructure.

There is a change worth being direct about. A great deal of this work is now faster because drafting, summarising, cleaning and reformatting can be assisted. That does not remove the need for a person, because the judgement, the chasing, the exceptions and the relationships are the work. What it does mean is that an assistant who uses those tools well gets through substantially more than one who does not, and that is where the value is — not in buying cheap hours, but in buying hours that clear more.

How the work runs

1. A fortnight of learning the business and writing it down. Not a handover document produced at the end, but a procedure written for each task at the moment it is first done — purpose, what is needed beforehand, the steps in order, the checks, and the points where something must come back to you. By the end of the second week you own a set of documents you did not have before, whoever ends up doing the work. Two weeks.
2. Access, set up properly. A password manager, per-tool access rather than one shared login, delegated accounts wherever the tool supports them, and a written list of exactly which systems are in use so revoking everything is one deliberate action rather than an archaeology exercise. Two to three days, before any task work begins.
3. Decision boundaries, agreed in writing. Which calls the assistant makes alone — which enquiry is worth pursuing, which invoice gets chased today, which meeting moves — and which ones never leave your desk, meaning anything touching money, discounts or a promise to a customer. Ambiguity here produces either paralysis or an expensive surprise. Three to five days.
4. A daily rhythm rather than a task queue. Inbox and urgent items in the morning, the heavier data and system work in the middle of the day, follow-ups and the day's summary at the end. A twenty-minute call once a week for priorities and blockers. Working to a rhythm is what stops this becoming an interruption-driven arrangement where nothing substantial ever gets finished. Ongoing.
5. Time logged from the first day. Tracked against tasks, not as a single monthly total, so both sides can see where the hours actually went and whether the arrangement still matches reality. Ongoing, a few minutes a day.
6. A short monthly summary. What was cleared, hours against each area, what is on hold and waiting for you, and anything noticed that is worth changing — a form that keeps producing bad data, a recurring task that should be automated away rather than done faster. Monthly.

What moves the price

The range is $500–$2,500/mo, ongoing, and three things determine where you sit in it.

General admin versus domain knowledge. This is the largest single factor. Inbox management and data entry sit at the bottom of the range. Someone who already knows how e-commerce supplier and returns flows work, or property listings, or recruitment pipelines, or clinic billing, works with far less supervision and catches the errors that cost money — and is priced accordingly.

Volume and the number of systems. One inbox and one spreadsheet is a different engagement from an inbox, a CRM, an accounting package, a marketplace back-end and a supplier portal, each with its own conventions and its own ways of going wrong.

How much judgement is included. Executing defined tasks is the lower end. Triaging what matters, deciding what gets chased and closing loops without being asked is a different level of responsibility and it is what the upper end of the range buys.

What usually goes wrong

Hours are sold and never measured. The agreement fixes a number of hours a month. Small requests arrive daily, none of them individually unreasonable, and before long the same fee is buying considerably more work — at which point the conversation is difficult because you believe the scope has not changed, and from your side it never visibly did. Logging time from day one and showing it monthly makes the drift visible while it is still small.

Credentials are shared over chat or kept in a spreadsheet. It is the fastest way to start and the fastest way to end a business. Passwords in message history cannot be revoked or audited, one shared login makes every action unattributable, and a single compromise reaches everything at once. A password manager is not administrative overhead here, it is the thing that makes the arrangement safe enough to exist.

The work gets done and never written down. Everything ends up in one person's head, which feels efficient right up to the first week of leave, a resignation or a volume increase. With no procedures nothing can be checked, nothing can be covered, and nothing can be handed to a second person without starting again. Writing the document while learning the task costs almost nothing. Reconstructing it afterwards costs a great deal, and it usually happens at the worst possible moment.