Appearance
By Scot, IT support technician
Applies to: General Business, IT Support
Article Type: Informational
Last Updated: 2026-08-09
You are the IT department. That arrangement works well for longer than most people expect, and when it stops working the change is gradual enough to miss from the inside. This guide sets out the signals that doing it yourself has stopped being the cheaper option, the five shapes outside help actually comes in and what each is genuinely for, what to have in front of you before the first conversation, and how to judge a provider when you cannot judge their technical work. It also says plainly when staying DIY is the right answer, because for a great many one-person businesses it is.
In this guide
The clearest signal is a piece of arithmetic that takes two minutes, and it is the one most owners have never actually done.
Count the hours. Log them for two weeks, or scroll back through last month's calendar, messages and browser history. Count everything: the printer that stopped talking to the laptop, the password reset, the machine that would not wake, the hour spent reading forum posts about a sync error, the Saturday setting up the new starter. Count the interruption tax as well — every time IT pulled you out of real work, add the twenty minutes it took to get back into it.
Multiply by what an hour of your time earns. Use your billable rate if you have one. If you do not bill by the hour, divide last year's revenue by the hours you worked and use that.
Annualise it. Weekly hours multiplied by the weeks you actually work.
A worked example. Four hours a week is 200 hours a year. At $80 an hour that is $16,000 of capacity — and, more tellingly, five working weeks. Most owners who run this for the first time are surprised by the week count rather than the money.
Now the honest caveat, because the number lies in one specific way. Those hours only convert into money if you would genuinely have billed them. If your diary has gaps, the real cost is lower, and what you are measuring is quality of life rather than lost revenue — which is a legitimate reason to get help, but a different argument. If your diary is full and you are turning work away, the number is real and you should treat it as real.
What this test does not tell you is whether help costs less than that. That comes in section 3, and for a lot of small businesses the answer is that it does not — which is a perfectly good outcome from a two-minute calculation.
Time is one signal. Four others matter more, because each is about risk rather than cost, and risk does not show up in a spreadsheet until it has already happened.
The same problem keeps coming back. Fixing something and watching it break again next month means you found a workaround rather than a cause. That is not a failing — diagnosing a root cause takes knowledge of how the pieces fit together, which is precisely what you are paying for when you pay someone. A recurring fault is the cheapest possible reason to bring somebody in, because the scope is small and obvious.
Somebody other than you started asking. The pressure usually shows up around five people. Up to that point "ask Steve" is a workable system. Past it, you are a helpdesk with a queue, the queue has no hours, and every person waiting on you is also not working. The tell is not the volume of questions — it is the moment you start putting someone off because you are busy.
An incident showed you had no safety net. A drive that died, a ransomware email somebody nearly clicked, an account nobody could get back into. The question that matters afterwards is "how long until we are working again", and if the honest answer is "I don't know", that gap is the whole argument. Close the obvious half of it yourself first — How to Back Up Business Data When You Don't Have IT covers what a working backup looks like, and a tested restore is worth more than any retainer.
A client sent you a security questionnaire. Once customers start asking how you handle data, who has access, and what happens when something goes wrong, the answer "I handle it myself" costs you contracts regardless of whether it is true. This is the signal that most often forces the decision, and it arrives with a deadline attached. It is also the one where a one-off specialist beats a monthly retainer, because what you need is documentation and evidence rather than ongoing support.
People say "get IT support" as though it were one product. It is five, they cost wildly different amounts, and picking the wrong shape is how small businesses end up paying a lot for something that does not fit.
| Shape | What it is genuinely for | Typically charged | Failure mode |
|---|---|---|---|
| Break/fix contractor | Something is broken now and you want it fixed | By the hour, often with a minimum callout | Nobody is preventing anything — you only ever buy emergencies |
| Managed service provider (MSP) | Ongoing monitoring, updates, and a helpdesk your staff can call | Monthly retainer, usually per user or per device | Paying a per-seat monthly fee for work that amounts to two hours a quarter |
| Fractional or part-time IT person | Real ongoing needs that do not fill a job | Day rate, or a fixed number of days a month | The work grows past their days and nobody notices until it breaks |
| Specialist for one project | A migration, a security review, a network installation | Fixed price against a defined scope, or a day rate | Scope creep, if "defined" was never written down |
| First internal hire | IT is now part of how the business runs, not an overhead | Salary | Hiring at the wrong seniority — a junior with nobody to learn from |
A break/fix contractor is the right starting point for most small businesses and the one people skip past too quickly. You call when something breaks, they fix it, you pay for the time. There is no such thing as engaging one too early. The failure mode is on the other end: businesses that stay on break/fix for years buy nothing but emergencies, at emergency rates, and never get the boring preventive work that would stop the emergencies happening.
A managed service provider puts your business on a monthly retainer, typically priced per user or per device, and in exchange monitors your machines, applies updates, runs a helpdesk your people can call directly, and often handles hardware procurement. Engage one too early and you are paying a recurring per-seat fee for a workload that does not exist yet. Engage one too late and you have spent two years as an unpaid helpdesk for a dozen people. The crossover is usually somewhere between five and fifteen staff, and it depends far more on how technical your team is than on how many of them there are.
A fractional or part-time IT person is a day a week, or a day a month, from someone who knows your setup. It suits a business with genuine ongoing needs that do not add up to a job — the awkward middle where an MSP's per-seat pricing looks poor value but the work is real. The risk is drift: the work quietly outgrows the days, and because your person is competent they absorb it until something falls over.
A specialist hired for one project is the shape most under-used by small businesses. A migration to Microsoft 365 or Google Workspace, a security review to answer a client questionnaire, a proper network installation in a new office — these are finite jobs with a defined end. Buy them as finite jobs, at a fixed price, against a written scope. Doing your own email migration to save a few thousand is the single most common expensive mistake in this whole area, because the failure mode is lost mail and there is no undo.
A first internal hire rarely makes sense below twenty or thirty people, unless technology is your product, in which case it happens far earlier and for different reasons. The trap is hiring cheap: a junior with nobody to learn from and no senior to escalate to will be unhappy and ineffective, and you will conclude that internal IT does not work when what did not work was the structure.
Any quote you receive is only as good as what you told the person, and four short documents do most of that work. Build them before you speak to anybody.
An inventory of what you run. Every laptop, desktop, phone and tablet used for work, with rough age. Every piece of software you pay for. Your internet connection and who provides it. Your printers. It does not need to be elegant — a spreadsheet or a page of notes is enough. What it needs to be is complete, because the thing nobody mentioned is always the thing that breaks.
A list of who has access to what. Which people can sign into which systems, and who holds the administrator account for each one. Include the ex-employee nobody removed, the shared login three people use, and the account that belongs to a person who left in 2023. Writing this down is uncomfortable and it is the most valuable of the four.
A note of where the data actually lives. Not where you think it lives. Client files, accounts, contracts, photos, the customer list — for each, name the actual place: OneDrive, a Google Drive folder, a NAS in the corner, one particular laptop's desktop, an accountant's portal. Anything that exists in exactly one place is a risk you can see now rather than later.
An estimate of what "down" costs per hour. If nobody can work, what does an hour cost? For a solo consultant it is your hourly rate. For a shop it is takings. For a business with staff it is wages plus lost revenue. This number decides everything about what level of support is worth buying, and without it every provider will sell you the tier they prefer to sell.
Build these four and something useful happens: a fair number of owners find the exercise itself answers the question. Either the picture is small and tidy and you keep doing it yourself, or you find three things that frightened you and you know exactly what to ask for.
You cannot assess whether someone is technically good. You can assess how they answer questions, and that turns out to be a strong enough signal on its own.
Six questions worth asking, and what the answers tell you:
On response times, what reasonable looks like: commitments tiered by severity, so "everyone is down" and "one person's printer" carry different promises; business hours defined in writing, with out-of-hours priced separately if it exists at all; and a named route for reaching them that is not one person's mobile number. Small providers offering four-hour response inside business hours for anything urgent is a normal, honest shape. Be more suspicious of unusually strong promises than of moderate ones.
One more test, and it costs nothing: send a mildly technical question by email before you engage anybody, and see how the reply reads. Somebody who explains their answer in language you understand will keep doing that for the next three years. Somebody who buries you in acronyms at the sales stage will not improve.
Both traps are about who holds the keys to your business. Both cost nothing to avoid at the start and a great deal to fix afterwards.
Trap one: you must own your own domain, accounts and data. Your domain name must be registered in your business's name, in an account you can sign into. Your Microsoft 365 or Google Workspace subscription must be billed to you, or at minimum transferable to you. You must hold an administrator account for it yourself. This happens far more often than it should: a provider sets everything up efficiently under their own umbrella, everything works fine for years, and then you want to leave and discover that your email, your domain and your files sit inside somebody else's account. Understanding the Microsoft 365 Admin Center for Non-IT Owners covers keeping your own Global Administrator account and why a second one matters — hold that account whoever else has one.
Ask one question before signing anything: "If we stopped working together tomorrow, what do I already own outright?" Write the answer down.
Trap two: you must be able to leave. Read four specific things in any agreement — the notice period, whether it renews automatically and how long before renewal you have to give notice, whether there is an offboarding or handover fee, and what exactly they hand over when you go. A twelve-month auto-renewing term with a ninety-day notice window is a contract that is difficult to exit on purpose. Handover should include documentation, account details and passwords in a usable form, and it should be described in the contract rather than negotiated at the moment relations have soured.
Neither of these is a reason to distrust providers as a category. Most are straightforward, and the ones who answer both of these in plain terms and without irritation are telling you something good about how they operate.
For a great many one-person businesses, doing it yourself remains correct, and it is worth saying so directly rather than treating hiring as the mature end state.
If your setup is a laptop, a phone, a Microsoft 365 or Google Workspace subscription, and cloud storage, there is nothing there that needs a professional on a retainer. That stack is designed to be run by its owner. Paying a monthly fee for somebody to monitor it buys you very little, and the money is better spent on better hardware or a second machine you can fall back on.
What DIY does require is that four things are genuinely handled rather than intended:
That last one is the safety net, and it is the thing running solo genuinely lacks. It is not a technical problem and it does not need a provider. It needs one document and one person who knows it exists.
The realistic middle path, and the one most small businesses land on: run it yourself, keep the name of a contractor you have used once for something small, and buy specialists by the project when something finite and important comes up. That gives you a number to call without a monthly bill, and it is a perfectly respectable end state rather than a stepping stone to something else.
WARNING
The one thing to check before you sign anything: your domain name and your Microsoft 365 or Google Workspace subscription must be registered to your business, with an administrator account you hold yourself. A provider who owns them owns your email, and recovering that later means a transfer dispute with somebody you have already fallen out with.
| Symptom / Error | Potential Cause | Solution |
|---|---|---|
| IT eats a day a week but hiring looks unaffordable | Comparing a full retainer against DIY, rather than the smallest useful option | Start with a break/fix contractor by the hour, or buy one specialist project. Neither carries a monthly commitment, and both cut the recurring problems that consume most of the day. |
| The same fault keeps coming back | A workaround was applied, not a fix | This is the cheapest possible reason to bring somebody in. Give a contractor the specific symptom and what you have already tried — a defined, small job. |
| A client's security questionnaire asks questions you cannot answer | No documented policies or evidence | Hire a specialist for a one-off review with the questionnaire as the scope. This is a finite project with a deliverable, not a reason for a retainer. |
| A provider's quote is far higher than expected | They are pricing the unknown | Give them the four documents from section 4. A provider who knows your device count, your data locations and your downtime cost can price the real job instead of the worst case. |
| You cannot get into your own Microsoft 365 or domain account | Everything was set up under the provider's account | Ask them in writing to transfer the domain to a registrar account in your business's name and to grant you a Global Administrator account. Do this while relations are good — it is far harder afterwards. |
| You are the only person who knows how anything works | No written handover, common in solo businesses | Write one page listing every account, where the backup lives, and who to call. Store it where a trusted person can reach it. This costs an hour and is the single highest-value thing on this page. |
IT Tip Tuesday
Get a short, actionable IT tip in your inbox every week.
Nothing loads from beehiiv until you click. The form sets beehiiv's own cookies once it does.
