A specialist UK recruitment operation Orchestrix audited last year was losing 14 hours a week per worker to manual admin work. Eight people on the operations team. That is 112 hours a week. Almost three full-time equivalents of human attention spent on rekeying, chasing, copy-pasting, reformatting, and reconciling.
Nobody planned for it. Nobody hired for it. The work just accumulated.
This post is about where those hours go in the typical UK SMB, why nobody sees it building up, and what the realistic options are for getting them back.
The work that doesn't have a name
Operational admin work in a UK SMB rarely shows up on an org chart. There is no role called "the person who copies invoice numbers between Xero and the project tracker every Friday afternoon". The work just falls to whoever notices it first.
Once it falls, it stays. Six months later, that person has a mental list of fifteen recurring admin tasks they do every week, none of which appear in their job description, and all of which absolutely have to happen or something downstream breaks.
The bureau calls this "shadow admin". It is the work that exists because three systems disagree, two tools do not talk to each other, and one rule that nobody can remember the reason for. It is invisible until you measure it.
Where the hours go (six common patterns)
Across forty audits the bureau has run for UK SMBs, the same patterns show up:
1. Rekeying between systems. A salesperson closes a deal in the CRM. The same data has to land in the project tool, the accounting system, the onboarding email, and the team Slack channel. Without integration, that is four manual entries per deal. At ten deals a week, it is forty entries. At three minutes per entry, two hours a week, gone.
2. Building reports by hand. Monday morning, the operations manager pulls numbers from three systems, pastes them into a spreadsheet, formats the spreadsheet, emails it to the leadership team. Three hours a week, every week, for a report that could be auto-generated overnight.
3. Chasing for input. Approvals, sign-offs, document submissions, response forms. The work is waiting for a human to do something. Without an automated nudge, somebody (usually the same somebody) chases by email, follows up the next day, and follows up again. Two to four hours a week.
4. Reconciling disagreements between systems. The CRM says one thing about the customer. The accounts system says another. The team's spreadsheet says a third. Half a day every month spent working out which version is true, then updating the other two.
5. Document handling. Generating invoices, contracts, onboarding packs, and reports manually. Each one takes 10 to 30 minutes. At five a week, two hours minimum.
6. Compliance evidence collection. For care providers, accountancy firms, recruitment agencies, anyone in a regulated sector. Pulling DBS records, training certificates, RTW evidence into a single place for an audit or a renewal. Six hours a quarter, eight hours a renewal cycle, every cycle.
Add those up across a small team and the 15-hour-a-week figure starts to look conservative.
Why nobody sees it building up
The pattern compounds because each individual task feels small. Three minutes here, ten minutes there. Nobody schedules a meeting to discuss the three-minute task. Nobody escalates it. Nobody notices when one becomes five.
The other reason it stays invisible: the work is usually done by people who are also good at their actual jobs. The salesperson who is excellent on the phone is also the one who manually copies deals into Xero on Friday afternoons. The senior consultant who delivers the work also generates the monthly client reports by hand. The owner who knows every customer's name also reconciles the bank statement on Sunday evening.
Because the work gets done, nothing breaks. Because nothing breaks, nobody investigates. The accumulating cost is in the work that does not happen: the prospect that does not get followed up, the strategic project that does not get started, the senior consultant whose time is worth £200/hour spending three of those hours pasting numbers into a spreadsheet.
What the maths actually looks like
Take a 20-person UK SMB at £5m revenue. If half the team is doing 10+ hours a week of avoidable admin (a conservative estimate based on real audit findings), that is 100 hours a week. At a blended £40/hour fully-loaded cost, the business is spending £4,000 a week, £208,000 a year, on work that should not exist.
Most owners, when shown the maths, do not believe the number. The follow-up question is always "what would it cost to fix it?". The honest answer is "less than a year of the cost".
A focused workflow audit costs £2,500. A custom automation that addresses a specific bottleneck typically costs between £2,000 and £8,000. Even at the higher end, the payback period is usually under twelve months, and the savings continue every year afterwards while the build keeps running.
What you can do without hiring anyone
Three steps that cost nothing and identify the worst of it:
1. Run a one-week time log. Ask the people on the team who are closest to the operations to write down everything they do for one week, in 30-minute blocks. Not edited. Not curated. Just what actually happened. The bureau does this as the first step of every audit; the patterns become obvious quickly.
2. Look for the work that nobody owns. The tasks that get done by whoever notices first. The tasks that fall through every quarter and somebody has to scramble to recover. Those are the candidates.
3. Look for the spreadsheet that became load-bearing. Every UK SMB has at least one. The spreadsheet that started as a quick fix and now somehow runs the operations. If it crashed tomorrow, the business stops. That spreadsheet is a custom application waiting to be properly built.
If you do those three things and the time-log adds up to 15+ hours a week of avoidable work, you have a real operational problem and the maths to justify fixing it.
When automation is the right answer (and when it is not)
A custom build is the right answer for repetitive, rule-based, high-volume work. A hire is the right answer for varied, judgement-led work. Most SMBs end up with both: the automation handles the rules, the people handle the judgement.
The work that should go into a build is the work where the 100th instance is the same as the first. The CRM-to-accounts sync. The Monday morning report. The compliance evidence collection. The onboarding email sequence. None of those need a human; they need a script that runs whether anyone is watching or not.
The work that should stay with a person is the work where every instance is different. The first call with a new prospect. The difficult conversation about scope. The judgement about whether a tricky case is worth taking on. Automation is the wrong shape for those.
What to do next
If the 15-hour figure feels familiar, the 15-minute triage is the cheapest way to find out exactly where the time is going. The conversation maps the worst three workflows, identifies which can be fixed and which cannot, and ends with an honest fit assessment. Half the time the answer is "no, this is a hire problem, not a build problem". The other half, there is a build worth scoping.
If you want a structured way to map the work yourself before that conversation, the operational audit is the next step. It is a paid, fixed-fee engagement that produces a written punch list with fixed-cost proposals for each fix. The audit fee credits in full against any build that follows. Worst case, you walk away with a deliverable any other developer or agency can act on.
The hours are out there. Most UK SMBs are paying for them every week and not seeing the bill.